Naval History and Heritage Command, Department of the Navy

Washington, DC
Published by
Naval History and Heritage Command
805 Kidder Breese Street SE
Washington Navy Yard, DC 20374-5060
www.history.mil

Use of ISBN
This is the official U.S. Government edition of this publication and is herein identified to certify its
authenticity. Use of 978- 1-943604-24-1 is for this edition only. The section 508–compliant PDF
of this title is cataloged under ISBN 978-1-943604-25-8. The title’s e-book edition is cataloged
under ISBN 978-1-943604-45-6.

Library of Congress Cataloging-in-Publication Data

Names: Friedman, Norman, 1946– author.
Title: Winning a future war: war gaming and victory in the Pacific war/Norman Friedman.
Other titles: War gaming and victory in the Pacific war
Description: Washington, DC: Naval History and Heritage Command, Department of the Navy,
[2017] | Includes bibliographical references and index.
Identifiers: LCCN 2017050896 | ISBN 9781943604241
Subjects: LCSH: United States. Navy—Maneuvers—History—20th century. | Pacific Area—
Defenses. | World War, 1939–1945—Pacific Area. | Naval Maneuvers—United States—
History—20th century. | War games, Naval—United States—History—20th century. | Naval
War College (U.S.)—History—20th century. | Naval strategy—Study and Teaching—United
States—History—20th century.
Classification: LCC V245 .F76 2017 | DDC 359.4/80973—dc23 LC record available at https://
lccn.loc.gov/2017050896
I   C H N, USN
C

Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1. Naval Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Exercises: Full-Scale Fleet Problems and Games at Newport . . . . . . . . . . . . . . . 8
Naval Aviation as a Driver Toward Transformation . . . . . . . . . . . . . . . . . . . 11
The Inter-War Navy and Its World . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
The Strategic Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Naval Arms Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Ships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

. The Naval War College and Gaming. . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

. War Gaming and War Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

The “Applicatory System” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
War Gaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
War Gaming at the Inter-War War College . . . . . . . . . . . . . . . . . . . . . . . . 54
Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Some Limits of Gamed Reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Using War Gaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
War Gaming and War Planning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

vii
. War Gaming and Carrier Aviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Guessing What Aircraft Could Do . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Gaming and Early Carriers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Reeves and Operating Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Putting It Together—the Yorktown Class . . . . . . . . . . . . . . . . . . . . . . . . . 97
Aftermath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

5. The War College and Cruisers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109

Evaluating Alternatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Cruisers at War: Three Years of Red-Blue Warfare. . . . . . . . . . . . . . . . . . . 129
Postscript: The Fate of the Flight-Deck Cruiser . . . . . . . . . . . . . . . . . . . . . 139

6. Downfall. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142

. Conclusion: Games Versus Reality in the Pacific . . . . . . . . . . . . . . . . . . . .160

Appendixes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
A: Playing the Games . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
B: War Game Rules—Aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
The Airplanes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Carrier Air Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Bombing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Bombs Versus Carriers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Torpedo Bombing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Air-to-Air Combat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Anti-Aircraft Firepower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Aircraft Navigation and Reliability . . . . . . . . . . . . . . . . . . . . . . . . . 198

Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .201

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256

viii Winning a Future War

A

T his book began with the observation by Captain Jerry Hendrix, USN (Ret.), who was
then director of the Naval History and Heritage Command (NHHC), that there was
no history of inter-war gaming at the Naval War College in Newport comparable to the
detailed histories of the full-scale Fleet Problems (fleet exercises). I am grateful to Captain
Hendrix both for initiating this project and for impelling me to collect material covering the
full range of games played at the college; I would never have done so without the sponsor-
ship of the Naval History and Heritage Command. In turn, the gaming material became
the basis for this book. I should emphasize that it was the full range of material that led me
to the conclusions in this book, not any one particular game. I am grateful to Dr. Evelyn
Cherpak, now retired, who, as archivist at the Naval War College, made it possible for me
to review the gaming material and to copy a great deal of it. I had been visiting Dr. Cherpak
and her archive for some years, but had always shied away from the sheer mass of gaming
material in the past. The advent of the digital camera, and the impetus provided by Captain
Hendrix, changed that. After Dr. Cherpak retired, I benefitted from the assistance of inter-
im archivist Scott Reilly and his assistant (later archivist) Dara Baker. Since important ma-
terial, including gaming material, was held elsewhere, I visited several other archives. Thus,
I am also grateful to the archivists of the National Archives, both in downtown Washing-
ton, DC, and at College Park, Maryland; to the Admiralty Librarian Jenny Wraight; and to
the archivists at the British National Archives at Kew. For their very enlightening insights
and comments on the ongoing project, I would like to thank my friends Chris Carlson, Dr.
Thomas Hone, Trent Hone, David Isby, and Christopher C. Wright. At the Naval History
and Heritage Command, Dr. Ryan Peeks provided a wide variety of assistance as well as
very helpful comments. Earlier, I benefitted from the efforts of Curtis Utz, who was my
contract officer at NHHC. I would also like to thank Kristina Gianotta, who, as head of
the Histories Branch, managed this project. Far from last, I would like to thank my wife,
Rhea, for her sustained loving and patient encouragement and support.

Acknowledgments ix
Mongolia
Manchuria

Korea PACI FIC

Japan
OCEA N
China

Iwo Jima Midway

Islands
Formosa
Hawaiian
Hong Islands
Kong
Mariana
Islands
Luzon

French
South Philippines Marshall
Islands
Indochina C hina Guam

Sea
Truk
Mindanao
North Palau
Borneo Islands Caroline Islands Solomon
Islands
Makin
Sarawak
Bismarck
Borneo Archipelago Gilbert
Celebes Islands
New Guinea Solomon
Islands
Java Papua
Ellice
Guadalcanal Islands

Australia
 L u z on S tr a i t

L inga ye n
G ulf

Luzon P hi l i p p i n e
Sea
Manila
Sub ic Ma nila
Ba y Ba y
Bataan

ait
Str
Mindoro rdi
no
rna
Be
S an

Mastate Samar

Panay Leyte M an i l a
Ley t e Bay
Gu l f
Str igao

Cebu
ait
r
Su

D i n ag at
I s l an d
Panaon
Negros Bohol Island

Palawan

Sulu
Sea Mindanao

M o r o Gu l f

C e l e be s
Malaysia
Sea
I

T o win the Pacific War, the U.S. Navy had to transform itself technically, tactically,
and strategically. It had to create a fleet capable of the unprecedented feat of fighting
and winning far from home, without existing bases, in the face of an enemy with numer-
ous bases fighting in his own waters. Much of the credit for the transformation should
go to the war gaming conducted at the U.S. Naval War College. Conversely, as we face
further demands for transformation, the inter-war experience at the War College offers
valuable guidance as to what works, and why, and how.
The fruits of this transformation are so commonplace now that we may easily for-
get how radical it was. The U.S. Navy emerged from World War I as a battleship fleet
similar to other navies. The British had demonstrated that naval aircraft could be a
vital auxiliary to the battleships, but anything more was a distant prospect. The war
had demonstrated that an amphibious operation could be mounted in the face of resis-
tance, but not that it would be particularly effective. In 1943–45, carriers were the ac-
cepted core of the U.S. fleet, and amphibious operations against enemy shore defenses
were routinely conducted. Indeed, without them it would have been impossible to fight
World War II.
If it seems obvious that any naval officer aware of the march of technology would
have developed the massed carriers and the amphibious fleet, the reader might reflect
that the two other major navies failed to do so. The Japanese did create a powerful carrier
striking force, but they made no real effort to back it up with sufficient reserves to keep
it fighting. They developed very little amphibious capability useful in the face of shore
defenses: They could not, for example, have assaulted their own fortified islands, let alone
Normandy or southern France. The British built carriers, but accepted very small car-
rier air groups because, until well into World War II, they saw their carriers mainly as
support for their battle fleet. Like the Japanese, they did not develop an amphibious ca-

Introduction 1
 pability effective against serious defense. Each of the three navies was staffed by excellent
officers, often with the widest possible experience. What set the U.S. Navy apart?
War gaming at the U.S. Naval War College at Newport, Rhode Island, seems to
have been a large part of the answer.
The games played by students there were a vital form of training, but at least as
importantly, the games served as a laboratory for the U.S. Navy. It seems to have been
significant that, until 1934, the Naval War College was part of the Office of the Chief
of Naval Operations (OPNAV) rather than part of the naval school system. Gaming
experience fed back into full-scale exercises (Fleet Problems), and full-scale experience
fed back into the detailed rules of the games, which were conceived as a way of simulating
reality as closely as possible. Game data were also fed to the U.S. Navy’s war planners, all
of whom had graduated from the War College and thus had considerable game experi-
ence. The successes and failures of simulation give some guidance into what is needed in
current and future games.
The two most important new developments of the inter-war period were aviation
and those involved in amphibious assault. In the case of aviation, the technology involved
changed rapidly, so its potential could be grasped in only a limited way from full-scale
experience. Gaming revealed some important problems (one of which Naval War College
graduate Captain, later Admiral, Joseph M. Reeves solved in a way that proved decisive at
Midway, where three U.S. carriers operated as many aircraft as four Japanese). Gaming
caused another Naval War College student, then-Commander Roscoe C. MacFall, to
devise the circular formations that proved very successful during the Pacific War. Yet an-
other Naval War College student, Commander (later Fleet Admiral) Chester B. Nimitz,
who had seen the formations in action, convinced the fleet to adopt them. No full-scale
fleet battle experiment could have revealed not only the problems, but tested solutions
that took years to implement. The War College taught all its students, not just those in-
terested in aviation, how to fight an air-sea war. Its success was reflected in the capability
of officers not trained in carrier operations to direct them. Admiral Raymond F. Spru-
ance is the great, but not unique, case in point.
The games showed that naval aircraft would suffer tremendous wastage in action. Be-
fore war broke out, the U.S. Navy created the necessary programs to produce sufficient
numbers of pilots. The Japanese did not; they did not look beyond a single battle to the
longer campaign in which it would be set. Once their experienced pilots were dead, so was
their naval airpower (except for kamikazes) despite their attempts to make up carrier losses.
Games also showed that carrier flight decks would often be knocked out. When the
gaming began, it was assumed that such damage would have to be repaired in a shipyard; a
carrier once knocked out was out of action for an entire campaign. The U.S. Navy unique-
ly sought quickly repairable flight decks, which made it possible for U.S. carriers to keep
fighting after suffering battle damage. Without this innovation, the United States would
have had no operable carriers in the Pacific during vital periods in 1942–43. Japanese car-

2 Winning a Future War

riers lacked this feature, which is why the two best ones, Shōkaku and Zuikaku, damaged
at the Battle of the Coral Sea in May 1942, were not available to fight at Midway in June.
The games highlighted problems of scouting in the vastness of the Pacific, and thus
encouraged the U.S. Navy to supplement carriers with large numbers of shore-based na-
val aircraft, initially seaplanes but later land plane bombers under naval control.
Amphibious assault was a different matter. The technology changed between wars,
but the key issue was how it might fit into the larger U.S. strategy of trans-Pacific assault.
The key contribution of the War College gamers was to show that a strategy that did not
include seizing bases en route to the Far East was doomed. This insight motivated the
Marines’ shift toward seizing and defending advanced bases across the Pacific, and it led
to development of the specialized ships and craft and techniques that made amphibious
attack a commonplace during World War II. Without one crucial game played in 1933,
it is not at all clear that any of this work would have been done.
To anyone looking back at the U.S. Navy of the inter-war years, the revolution is
almost invisible. Throughout the period, the U.S. Navy continued to consider the battle-
ship the core of its fleet, just as other major navies did. Aircraft were important, but they
were an auxiliary. Gaming explains what was really happening behind this appearance.
Gaming forced participants to consider not just the decisive battle that the fleet hoped to
fight—a classic battleship-on-battleship engagement—but also the campaign leading up
to it. To get at the Japanese fleet, the U.S. Navy had to cross the Pacific, passing through
chains of islands the Japanese controlled.
Gaming presented not just what the U.S. Navy wanted to do, but also what the
Japanese would choose to do to stop its advance. It became clear in games that the U.S.
Navy would have to fight a series of battles against Japanese aircraft and submarines as
it pressed westward. These secondary battles were a fair representation of what actual-
ly happened in the Pacific in World War II, as the Japanese reserved their own battle
fleet for the decisive battle they expected to fight. Incidentally, many of the tactics later
attributed to Japanese national character, such as risky night torpedo attacks, emerged
clearly in the games as the inevitable results of the Japanese strategic and tactical situa-
tion, rather than reflections of their military culture.
Why did gaming work? It was hardly the only way a navy could look toward the fu-
ture. There were plenty of highly intelligent, experienced officers with excellent military
judgment. They often found the results of gaming rather artificial, even rigged in favor
of impractical technology. When they confronted straightforward developments of ships
and weapons they had used for years, their judgment was probably much better than what
games might suggest. However, when they confronted essentially new technologies, par-
ticularly rapidly changing aircraft, they were not nearly as well prepared to judge. They also
confronted air enthusiasts who made their own inflated claims that traditional (meaning
well-developed) naval forces were obsolete. Because it was better balanced, gaming offered
a much more acceptable (and better) way to sense what a future air-sea war would be like.

Introduction 3
 Gaming reached a high point of influence in 1930–34. For most of that period, the
Chief of Naval Operations (CNO), Admiral William V. Pratt, was a former president
of the Naval War College, who was familiar with the virtues of gaming. Contemporary
records show that there was widespread skepticism in the fleet. The fleet saw, for example,
the obvious limitations of existing aircraft, some of which were dramatically reduced in
game rules (to sense what could be done within a few years). Not surprisingly, the gamers
at the Naval War College lost most of their influence after the game-friendly CNO re-
tired and the Naval War College was moved from a position as OPNAV’s laboratory into
the Navy educational system. The building program advocated in 1939–40, which was
battleship- and not carrier-heavy, dramatizes the reaction to what the War College games
had taught. As it happened, the approach of war and also pressure by an air-minded
President Franklin D. Roosevelt changed the situation. However, the Naval War College
gamers had already won a larger victory. When the United States entered World War II,
all but one flag officer were Naval War College graduates—former war gamers who had
seen the sort of war they would now be fighting. The new techniques were already in place
and the strategy demonstrated in that 1933 game had been accepted.
Enough had been done to make it reasonable to see war gaming at Newport as a
key element of victory in World War II. That is most obvious in the Pacific War, but
in Europe amphibious landings were decisive. The games motivated development of the
necessary techniques, which were tested in full-scale Fleet Landing Exercises (FLEXes).
Thus, the games can also be thanked for victory in Europe.
This book differs radically from other accounts of the inter-war Navy. It emphasizes
the role of the Naval War College as the U.S. Navy’s primary think tank rather than as
its senior educational institution. The think tank role is evident in statements by the Col-
lege’s first post-1918 president, Admiral William S. Sims, in its pre-1914 role, in much
of the gaming material, and in the implicit relationship between the Naval War College
and the OPNAV War Plans Division, whose officers were its graduates. Gaming was the
core of the think tank role, because gaming was the Navy’s laboratory, in a way that full-
scale exercises could not be. It was extremely important that attendance at the Naval War
College was in effect a prerequisite for service in the OPNAV War Plans Division—the
games told the officers who created war plans what a future war would be like. As the
fleet’s strategic and tactical laboratory alongside the full-scale Fleet Problems, the Naval
War College had enormous impact on the way the U.S. Navy prepared for a future war.
This book also differs from previous accounts of inter-war gaming at Newport in
several ways, of which three seem particularly important. First, it examines the games
primarily as the U.S. Navy’s laboratory rather than primarily as a means of training.
Second, it is based on detailed examination of all the surviving senior course games. Typ-
ically, only a few games have been examined. Although the key game that changed Pa-
cific strategy has generally figured prominently in histories, other games have not fared
nearly so well. Third, it examines the influence of the games on specific areas of naval

4 Winning a Future War

development, particularly how the U.S. Navy learned to use naval airpower. Only an
examination of the totality of the games makes it possible to show how they influenced
the inter-war Navy and its conduct during World War II. It seems clear in retrospect that
gaming had its greatest tactical impact in showing how naval airpower would work (not
merely promoting it in general) and strategically in showing the vital need for advanced
bases en route to the Far East.
The educational aspect of this book emphasizes how the Naval War College sought
to shape the way officers thought about the problems they faced. It seems clear that its
approach differed radically from that of the senior educational institutions created by the
closest equivalent to the U.S. Navy, the Royal Navy.
The games, and the U.S. Navy as a whole, were products of a very specific time. Be-
tween the world wars, the United States avoided alliances of any kind, although, in the late
1930s, some games and other documents show an interest in them. It was not imagined
that a European war would suddenly leave the major colonies in the Far East almost de-
fenseless. Thus, the Far East wars the war gamers envisaged were fought in the presence of
powerful neutrals, particularly the British Empire, which could be expected to affect the
prosecution of any war. This mind-set is probably the best explanation for the one gap in
inter-war gaming, namely the absence of any large-scale submarine war against Japanese
trade (although some games do include small-scale submarine warfare of this type).

Introduction 5
1. N T

B etween 1919 and 1941, the U.S. Navy transformed itself from a powerful if unso-
phisticated force into the fleet that won a two-ocean war. The great puzzle of U.S.
naval history is how that was accomplished. This book argues that war gaming at the U.S.
Naval War College made an enormous, and perhaps decisive, contribution. For much of
the inter-war period, the Naval War College was the Navy’s primary think tank. War
gaming was the means the college used to test alternative strategies, tactics, and warship
types, in a way that full-scale exercises (the Fleet Problems) could not. There is consid-
erable evidence that the lessons the Naval War College drew from gaming were applied
by the fleet. For example, crucial choices in U.S. aircraft carrier design can be traced to
experience gained in war games at the college.
The think tank perspective is a new way of looking at the inter-war Naval War Col-
lege and the war games that formed the core of its curriculum. Historians have generally
treated war gaming (if they have treated it at all) as an educational tool, the question
being how effective it was.1 It seems clear that the inter-war Navy valued the college. At-
tendance was, in effect, a prerequisite for selection to the OPNAV War Plans Division.2
Of the flag officers commanding combat formations in the 1941 fleet, only one had not
attended the War College. After the war, Fleet Admiral Chester W. Nimitz, who had
commanded in the central Pacific, pointed out that at one time or another, war games
had replicated nearly everything that happened during the Pacific War. Officers who
had fought simulated battles were well prepared to fight real ones in the same places,
often in similar situations. Surviving records at the Naval War College clearly show the
changing relationship between the game scenarios and the outside world, and the in-
creasing sophistication of the game rules intended to simulate reality. They do not show,
as naval records from other official sources do, how deeply game experience influenced
overall U.S. naval development between wars. This influence is somewhat obscured by

6 Winning a Future War

frequent protestations by the Naval War College (typically by its president) that it was
in no way a war-planning center. That was true in a formal sense, but what the college
did mattered enormously.
Although the influence of both the Naval War College’s gaming and of the college
itself declined after 1933, most of the key decisions shaping the wartime U.S. Navy had
already been taken. The two most important ones were on the role of naval aviation and the
form of the U.S. war plan against Japan. Furthermore, war gaming helped instruct non-avi-
ators such as Admiral Raymond Spruance about the proper role of carrier aviation. They
successfully applied these lessons during the coming war. Similarly, the War College bears
some credit for the shift from the “through ticket to Manila” to the step-by-step strategy
the United States employed during World War II. This shift was based on gaming experi-
ence, particularly the disastrous outcome of the already noted 1933 game.
The inter-war U.S. Navy had to solve two linked problems. One was that it had
virtually no experience of naval warfare on which to draw. Its last conventional naval
war had been fought against Spain in 1898, before nearly all the weapons and ships of
1919 had even been conceived. Its World War I experience was limited almost entirely to
anti-submarine warfare, in many cases in a subordinate position to the Royal Navy, the
foremost navy of the time. A second problem was that the recently ended world war had
produced further new weapons, particularly aircraft and long-range submarines, whose
potential was difficult to estimate.
The problems of the inter-war era may seem dated, but they have very real current
equivalents. No one has fought a major war using modern weapons. No one really knows
how, for example, cyber weapons will affect such a war. We can and certainly do guess,
but it takes a simulated war against a sophisticated opponent to tease out what is likely to
happen and to stimulate us to take the appropriate measures. Only a simulated opponent
will seek to take full advantage of all available weapons in ways we may not expect.
The story of inter-war simulation—gaming—is worth following because it shows
how a particular approach to this problem worked. We can see how well simulation
served the U.S. Navy because we know its outcome: the successes and the limitations of
the simulation as applied to a real war. What we might expect from successful simulation
would be insights, which could not have been gained without it. In the case at hand, they
ranged from the realization that the fleet should and could seize intermediate bases in
the Pacific—the basis of much of what was done in 1943–44—to important insights into
the way to operate aircraft carriers (not to mention the sort of carriers worth buying) and
into what other sorts of ships the fleet needed. It is also possible to contrast the success
of the U.S. Navy with the problems suffered by the Royal Navy, a comparable sea service
that did not use gaming as a primary planning tool.

* * *

Naval Transformation 7
Exercises: Full-Scale Fleet Problems and Games at Newport
The U.S. Navy found two linked solutions to the problem of preparing for future war.
One was the conventional one: full-scale exercises, which were called Fleet Problems. The
other was war gaming at Newport. Each had complementary strengths and weaknesses,
both for prediction and for creating a modern officer corps.
It took full-scale exercises to see how ships and aircraft and some of their weap-
ons—and above all the personnel manning the fleet—would perform in reality. Nothing
but a full-scale exercise could model how well a cruiser or a division of cruisers would
perform at sea. Gamers at Newport used reports of such full-scale behavior to write the
detailed rules under which they operated. In some cases it was simply impossible to mod-
el full-scale warfare. For example, despite considerable input, the war gamers at Newport
were never able to model air-to-air engagements in any realistic way. A full-scale exercise
could get closer to reality, although even it could not include the full effects of air-to-air
combat. Neither full-scale exercises nor games could quite solve the problem. Officers
commanding during an exercise could get a fair idea of what full-scale combat might be
like, but they could not use their weapons freely and they could not risk damage to their
own ships. Thus, for example, they were not permitted to try dangerous night attacks.
Moreover, in the relatively impoverished inter-war Navy, it was impossible to run what
might now be called end-to-end tests of important weapons. The best-known case was
torpedoes. The inter-war Navy developed what it thought was a radically upgraded tor-
pedo using a magnetic exploder. It was tested only once, and that test turned out to have
been unintentionally deceptive.3
Against all their virtues, full-scale exercises could not simulate extended campaigns;
their duration was necessarily limited. They had to be scripted to the extent that, because
they were expensive training opportunities, the opposing forces had to be brought into
contact. Thus, they could not even hint at the scouting problems that proved so difficult
and so important during the Pacific War. Because the whole fleet had to participate,
full-scale exercises generally simulated fleet-on-fleet battle rather than any lesser engage-
ments the fleet might have to fight to get to that point. On a subtler level, it was difficult
for most participants to gain a sense of how the exercise as a whole proceeded, how the
opposing fleets used their elements in the required complementary way. What they saw,
they nearly always saw from a necessarily narrow perspective. Afterward, they might read
the official summary of the Fleet Problem, but that was hardly the same thing as experi-
encing it from a broad perspective.
To a considerable extent, too, a full-scale exercise was a public event. However well
the Navy might try to conceal what happened, a great deal inevitably leaked out. At the
very least, the progress of the exercise could be tracked by intercepting radio messages
used during it.4 The Japanese could and did monitor U.S. Fleet Problems and at times
managed to deduce important aspects of U.S. war plans.5 Moreover, during the inter-war
period, there was considerable political sensitivity to anything suggesting an offensive

8 Winning a Future War

The great question of interwar U.S. naval history is how the U.S. Navy transformed itself from a battle-
ship-oriented navy to the victorious carrier navy of 1943–45. When Admiral Sims reopened the War College
in 1919, the battleship was indisputably the core of the fleet. Here U.S. battleships perform Battle Practice,
1921. (NHHC NH 124139)

outlook on the part of the United States, yet the main naval scenario, a Pacific war, could
be won only by offensive action.6 That did not imply that war planners or gamers were
hotheads seeking war, only that if a war broke out, they knew they had to carry the fight
to the enemy. The terms have changed, but this type of political sensitivity, as opposed to
what might happen if war actually broke out, is certainly still relevant.
A war game provided the opposite, broad, perspective. It was the only way to simulate
a whole naval campaign, or even a lengthy part of a protracted war. The shared experience
of protracted campaign games provided many within the inter-war Navy a sense of what
a future war would be like. During the inter-war period, the Navy’s General Board was
largely responsible for the characteristics (broad design requirements) of new ships. It
held hearings to gain the views of the fleets and of the technical experts. Often, its senior
member asked the same question: How will this ship function in the Orange (Japanese)
war? The U.S. Navy’s war planners produced detailed plans for particular approaches to
the war, but they changed frequently as senior officers replaced each other and offered
different perspectives. It was the shared experience of war gaming at Newport that made
it possible for all involved in the hearings, and in other Navy activities, to understand the
character of an Orange war in a meaningful sense.

Naval Transformation 9
The transformed fleet: “Murderer’s Row” of fleet carriers at Ulithi, 8 December 1944. From the front, they are:
Wasp (CV-18), Yorktown (CV-10), Hornet (CV-12), Hancock (CV-19), and Ticonderoga (CV-14), of which the
first three were named after carriers sunk in 1942. (National Archives 80-G-294131)

Gaming offered another advantage. Those involved received practical experience in

the way in which the different elements of a fleet—its ships and its aircraft and its shore
supporting structure—worked together. This was a very different perspective from that
offered by a full-scale exercise toits participants. It was particularly important in a navy
whose officers tended to specialize for much of their careers. It took games to give a cruis-
er officer like Captain (later Admiral) Raymond F. Spruance a sense of what carriers and
their aircraft could and could not do. It took a game to show just how much (and what
kind) of scouting a war spread over much of the Pacific would demand (as well as the con-
sequences of failed scouting plans). The same games offered a valuable sense of how other
kinds of ships, such as destroyers and submarines, might operate both in support of, and
in opposition to, the U.S. fleet. In many games, one of the objectives was to familiarize
students with the full range of ships, aircraft, and naval weapons, both U.S. and enemy.
Using them, even on a game floor, offered far more insight than reading about them or
listening to lectures about them.
Perhaps most important, a game was a secure way to investigate tactics and strategy,
far from the eyes of potential enemies—and also far from civilians at home who might
be alarmed by the drastic tactics and decisions a major war demanded. Conclusions that
might be drawn from games did not have to be publicized and therefore could be drawn
out much more freely than those implied by large-scale exercises.

10 Winning a Future War

The other great influence on interwar U.S. naval transformation was the shift from an Atlantic to a Pacific per-
spective. The pre-1919 U.S. Navy had had a Pacific presence, but the primary consideration in development
was the threat of an Atlantic war. In 1919, the fleet moved to the Pacific, and so did the focus of U.S. naval
strategic thinking. Here, the battleship New Mexico (BB-40) enters Miraflores Lake in the Panama Canal, 26
July 1919. (NHHC NH 50296)

Naval Aviation as a Driver Toward Transformation

For the inter-war Navy, aircraft were not too different from such radical current technol-
ogies as cyber in the way they affect the modern Navy. Aircraft had proven their value
during World War I, and at its end they promised a good deal more. No one could predict
how they might evolve over the next decades, yet fleets had to be adapted to their needs
and potential. We look back and see a battleship navy barely acknowledging the value of
aircraft, which seem very primitive compared to those that fought World War II. How-
ever, during the inter-war period, the U.S. Navy was widely seen as the most air-minded
in the world, with a very active aircraft research and development program.7 For example,
it was the U.S. Navy that developed and emphasized the new technique of dive bombing,
a form of precision attack that proved very effective during World War II. War-gaming
experience had other vital impacts on the way in which the U.S. Navy used its aircraft
and also on the way it designed its carriers, with important consequences for key World
War II battles such as Midway and the Philippine Sea.

Naval Transformation 11
Naval aviation grew much more rapidly than might have been expected because the Washington Naval Treaty
left the United States with two huge carriers converted from cancelled battle cruisers: Lexington (CV-2) and
Saratoga (CV-3). Lexington is shown leaving San Diego on 14 October 1941 with Brewster Buffalo fighters for-
ward, Dauntless dive bombers amidships, and Devastator torpedo bombers aft on her flight deck. (National
Archives 80-G-416362)

As the inter-war era opened, the U.S. Navy was interested in aircraft carriers, which
the British had invented during World War I, but it had little idea of just what it wanted.
The Washington Naval Treaty allowed it to convert the two huge battlecruisers Lex-
ington and Saratoga into carriers, far larger than anything envisaged beforehand.8 What
could or should they be designed to do? How would comparable enemy carriers affect the
U.S. fleet? Much could be learned by using simulated carriers. In retrospect, it seems that
gaming experience prior to the completion of the two carriers motivated two of the most
important U.S. tactical developments of this period: The method of operating aircraft
that gave U.S. carriers a commanding lead in capacity, and the circular formations that
made carrier operation effective during the Pacific War.
After World War II, the Japanese wrote that through the inter-war period they had
followed the U.S. Navy step by step.9 Their belief that naval aviation might be effective
had been inspired by the U.S. Navy. That might have seemed laughable as U.S. and Allied
forces suffered badly at the hands of Japanese naval aviators through much of 1942, but it
makes more sense given the triumphant development of U.S. naval aviation during the war.

12 Winning a Future War

The two big carriers demonstrated what large numbers of naval aircraft could do, such as when the new
Saratoga raided the Panama Canal Zone in 1928 during that year’s Fleet Problem. Without these ships, the
potential of naval aviation would have been far more theoretical. The Fleet Problems and the games at New-
port were complementary. Newport showed what might be done with current and planned forces; the Fleet
Problems helped show what was practical. Saratoga is shown landing aircraft, 6 June 1935.The War College
had nothing to do with the terms of the Washington Conference, hence with her origin. However, the U.S.
practice of moving aircraft forward as they landed (as here) rather than striking them below was very much a
War College–inspired innovation and it was crucial to U.S. carrier success in World War II. No other carrier
navy tried this technique before World War II. Note the black stripe on the ship’s funnel, which distinguished
her from her sister Lexington. (National Archives 80-G-651292)

The great question for the three inter-war navies that operated substantial carri-
er forces was how to integrate them into their fleets. The U.S. Navy envisaged extend-
ed carrier-versus-carrier and carrier-versus-land-based aircraft battles not because its
leadership was more visionary than those of other navies, but because the kind of war
it was likely to face would potentially include such battles. It learned this from the cam-
paign-level war games fought out at Newport. No other form of visualization could in-
clude not only what the U.S. Navy wanted to do, but what a likely enemy would do to
frustrate it—no other form of prediction included the enemy’s vote.
Differences between what the Japanese and U.S. navies did in this area suggest that
the U.S. approach, which involved large-scale gaming, was decisively better. Both navies
sought to develop strong naval air arms. The inter-war U.S. Navy created a corps of se-

Naval Transformation 13
The crucial U.S. carrier innovation, which required arresting gear and a wire barrier, was invented on board
the small experimental prototype carrier Langley (CV-1) at the behest of the Captain (later Admiral) Joseph
M. Reeves, Commander, Battle Fleet Aircraft. Langley is shown at Cristobal, Canal Zone, on 1 March 1930
with 24 aircraft on deck, far more than she originally carried, and a credit to Reeves’ ingenuity. Reeves, in turn,
learned how important it was to speed up aircraft landings when he commanded the Blue air force in a war
game at Newport. (National Archives 80-G-185915)

nior aviators by requiring that aviation activities (including the Bureau of Aeronautics
and carriers) be commanded by aviators: Earning wings became a valuable career choice.
The Japanese did not go nearly so far, but they certainly encouraged air training. In both
navies, aviators argued that conventional surface ships, particularly battleships, were ob-
solete. We look back and see “battleship admirals” opposing the air-minded future. In the
Japanese navy, aviators went so far as to advocate elimination of battleships. They clashed
with the “fleet faction,” which poured money into the ultra-large battleships Yamato and
Musashi that Japan completed at the beginning of the Pacific War. In retrospect, the
Japanese naval aviators were over-optimistic as to what their aircraft could do. No such
faction formed in the U.S. Navy. That may be because U.S. aviators realized that the all-
or-nothing argument was unrealistic. Carriers and surface warships were complementa-
ry. They could see that demonstrated by what seemed to be simulated realistic experience
at Newport.

14 Winning a Future War

 It is striking that some U.S. officers who did not have flight training, such as Ray-
mond Spruance and Frank Jack Fletcher, nevertheless proved to be able commanders
of fleets built around carrier forces. Spruance went from being a cruiser commander to
commanding one of the two carrier task forces at Midway, and then to commanding the
Third Fleet—whose main firepower lay in its carriers—in the invasion of Saipan and the
Battle of the Philippine Sea. Somehow he and other surface officers managed to under-
stand the ongoing revolution in naval affairs. This book argues that the key difference was
extensive gaming at Newp ort, which provided all participants with an understanding of
all of the tools of naval warfare, including the new air tool.
One argument often raised against the inter-war Naval War College is that it missed
the changing balance between carriers and battleships. On the eve of war in the Pacific,
the U.S. Navy had only seven fleet carriers. The Japanese had ten. It is forgotten that the
U.S. figure had been dictated largely by the dead hand of the Washington Naval Treaty,
signed in 1922. When the United States began to rearm in the 1930s, new programs
were described as a means of building up to a modern “treaty” navy, meaning a fleet of un-
der-age (as defined by the treaty) ships matching the tonnages set in 1922. At that time,
carriers and naval aircraft were only barely (if at all) out of the experimental stage. The
rather generous tonnage set in 1922 for carriers could be attributed to the Royal Navy,
which considered them essential based on its World War I experience. The U.S. Navy
would have been satisfied with a much lower figure—but the record shows that the U.S.
Navy had almost no input into the treaty process. The War College certainly support-
ed attempts to devise air-capable ships not limited so severely, particularly the abortive
flight-deck cruiser and projected (but then abandoned) conversions of merchant ships.
These ships were employed in war games and demonstrated their value.
By the 1930s, the question within the U.S. Navy, as stated when the General Board
tried to draw up long-term building programs, was not whether carriers and naval air-
craft would supersede battleships, but when. The performance and potential of carrier
aircraft was increasing rapidly, but in 1939 it was by no means obvious that they could
sink modern battleships in the open sea. It took large numbers of aircraft from multiple
carriers to sink the huge Japanese battleships Musashi and Yamato as late as 1944 and
1945. Carriers were much softer targets. For example, it was accepted that they could
be disabled or sunk by dive bombers, but the weapons from these same aircraft would
be unable to penetrate a battleship’s protective deck (torpedoes were a different matter).
The Pacific campaign was a carrier war both because the Japanese eliminated the U.S.
battle line at Pearl Harbor and also because the Japanese refused to risk their battle force
until 1944. That was by no means inevitable. That the Naval War College and most naval
officers continued to emphasize battleships in 1941 should not be a surprise, nor should
it be considered proof of some failure by the college. On the contrary, war games from the
early 1920s on generally involved large numbers of carriers, their numbers often boosted
by assumed fast liner conversions.

Naval Transformation 15
Two important U.S. naval aviation innovators, both connected with the Naval War College, pose together
in front of aerial photography equipment at Naval Air Station North Beach, 27 December 1928: Admiral
William V. Pratt (left) and Rear Admiral Joseph M. Reeves (right). Pratt served as president of the Naval War
College; he was the only president to become Chief of Naval Operations. War games convinced him that the
Navy, bound by treaty, could never have enough carriers. He helped invent the abortive flight-deck cruiser as
a way to surmount that obstacle (another was emergency wartime conversion of existing merchant ships).
Reeves invented the crucial carrier operating technique of landing aircraft into a barrier (to place them in a
deck park forward). As a consequence, U.S. carriers operated more aircraft than their foreign equivalents, to
the point that the three U.S. carriers at Midway had much the same number of aircraft as four Japanese ships.
(NHHC NH 75876)

The Inter-War Navy and Its World

The inter-war fleet that transformed itself was different enough from the modern U.S.
Navy to require some explanation.10 Between about 1905 and 1914, the U.S. Navy moved
up and down between second and third place among the world’s navies, in each case far
behind the Royal Navy. With the collapse of the Imperial German Navy at the end
of World War I, the U.S. Navy was second only to the British, and the expected post-
war resumption of Congress’ gargantuan 1916 building program promised to bring the
American fleet close to parity.
Until 1914, U.S. naval attention was concentrated on the possibility of German aggres-
sion in the new world. With Germany gone after 1918, but Japan rapidly building up a new
navy, attention shifted to the Pacific, where it remained for most of the inter-war period.

16 Winning a Future War

 To a modern reader, the most striking difference between the present and the in-
ter-war period is probably that the inter-war United States had no allies. Indeed, it
strongly rejected the idea of peacetime alliances. That made the role of powerful neutrals,
particularly the British Empire, an important strategic issue in a way utterly unfamiliar
in the present day. U.S. strategists seem to have been entirely unaware that after 1921
the British considered the Japanese their likeliest future enemies, based on demonstrated
Japanese interest in dismembering the British Empire in Asia. Although the inter-war
U.S. Navy did not know it, the British took much the same view as the U.S. Navy of an
end game in the war they expected to fight against Japan. The Navy (and the U.S. gov-
ernment) was also unaware, at least until the late 1930s, that the British saw the United
States as their only worthwhile prospective ally. When U.S. naval officers were finally
permitted to engage in talks with the British in 1938, those talks had to be kept secret
from the U.S. public, much of which still imagined that the British had drawn the United
States into World War I for their own purposes.
The inter-war United States was, moreover, hardly the world’s dominant superpow-
er. The U.S. Navy had to seize control of the sea as a first priority. The modern U.S.
Navy enjoys effective sea control over most of the world’s oceans; its problem is how to
exploit that control to effect events on land. During the Cold War, the question was how
to defend the sea control gained at great cost during World War II, and that is the ques-
tion that may now be arising in east Asia. There is no modern equivalent to the inter-war
problem of how to seize control of the sea in the first place. The inter-war games were
concerned mainly with seizing sea control, the principal fruit of that control being the
ability to choke off an enemy’s seaborne trade.
Although the inter-war Navy could certainly bomb and shell enemy territory, such
attacks were relatively minor possible consequences of gaining sea control. Bombardment
and shelling and the seizure of islands, the techniques developed both in gaming and in
full-scale exercises in the 1930s, were a means of supporting a naval campaign, not seen
as ends in themselves. No one in the inter-war period imagined a seaborne assault like
many in the European theater, as direct strategic attack on an enemy’s home territory. It
happened that the means conceived to seize island bases for naval purposes provided vital
strategic leverage, culminating in D-Day.
The logic of a single concentrated fleet was the logic of battleship-on-battleship com-
bat. During the inter-war period, it was assumed that it would usually take a battleship to
sink another battleship.11 Although other ships could certainly help enforce sea control,
nothing but the destruction of the enemy’s battleships could ensure it. It was, moreover,
generally assumed that numbers would be decisive: A widely accepted mathematical ex-
pression was that combat power was proportional to the square of the number of bat-
tleships (Lanchester’s Square Law).12 The basic logic predated the mathematics: Rear
Admiral Alfred Thayer Mahan had famously told President Theodore Roosevelt never
to divide the fleet. The immediate consequence at that time was that U.S. battleships on

Naval Transformation 17
foreign stations, including those in the Philippines, were brought home and incorporated
in a concentrated U.S. battle fleet. After 1918, when U.S. attention shifted from the
Atlantic to the Pacific, the logic of concentration implied that the U.S. Navy could either
maintain its battle fleet in the Philippines or on the U.S. Pacific coast. Battleships de-
manded considerable industrial infrastructure to keep them operational. Without much
infrastructure in the Far East, that was no choice at all.13 The inter-war U.S. Navy main-
tained the small Asiatic Fleet based in the Philippines backed by a battle fleet based at
San Pedro (Los Angeles) and San Diego, California. It built up Pearl Harbor into a base
capable of accommodating the fleet if need be. Most strategic war games began with the
fleet at Pearl Harbor prior to a sortie en route to the western Pacific. The lack of facilities
in the Pacific, particularly to repair underwater damage, became a vital point highlighted
by war gaming in the mid-1930s.
Thus, the inter-war U.S. Navy was built around a single main fleet, which until 1940
was called the U.S. Fleet (its commander had the unfortunate acronym CinCUS). A
scouting element (actually a training fleet built around the oldest battleships) was based
in the Atlantic; in wartime it would have joined the main fleet in the Pacific. The Scouting
Force, built around modern cruisers, was integral with the fleet in the Pacific.14 Until the
outbreak of World War II, the only other fleet elements were the Asiatic Fleet (which
included the Yangtze River Patrol in China) and a small presence force in the Caribbean.
Today’s Navy consists of separate carrier and other task forces, partly because the
Navy has to be present in many different places at about the same time. That was only
beginning to be the case in 1939–40, as events in the Atlantic forced the organization of
an increasingly powerful Atlantic Squadron intended to enforce U.S. neutrality. In 1941,
it expanded into an Atlantic Fleet. The larger force remaining in the Pacific was renamed
the Pacific Fleet. The Atlantic Fleet fought an undeclared naval war against the Germans
through much of 1941. It was a convoy-escort force rather than a concentrated battle
fleet. When France fell in May 1940, one driver for a massive new U.S. naval building
program was the possible need to face an Axis battle fleet in the Atlantic, a scenario that
did not feature in inter-war games.
The modern organization is practicable because carriers achieve a considerable de-
gree of concentration by concentrating their aircraft. Several carriers can operate togeth-
er, but there seems to be a practical limit to how many can do so if they have large num-
bers of aircraft onboard. The pressure to keep the fleet concentrated is gone, and to some
extent the fleet operates more effectively if it is not fully concentrated.
Pre-war concentration of the U.S. Fleet encouraged naval officers to think about
their roles in a fleet-on-fleet engagement. As it happened, a wider view of a protracted
Pacific campaign required other kinds of operations. The only way that such a view could
be disseminated was through campaign-level gaming at the Naval War College. That
wider view is a major theme in this book. It turned out that much of the Pacific War cor-
responded to the build-up to a given decisive battle rather than to the battle itself.

18 Winning a Future War

 Naval warfare was understood very differently as well. Today, we think of power
projected from the sea by aircraft and missiles. That requires reconnaissance to support
targeting, but it generally does not require our fleet to locate another fleet in the vast-
ness of the ocean. We are far more interested in preventing an enemy from finding our
own fleet in order to attack it. However, in the inter-war years the U.S. fleet was very
much concerned with destroying enemy fleets. To do that it had to find them. Scouting
was essential, and the broad reaches of the Pacific made it difficult. Naval War College
students played games intended specifically to exercise their ability to create effective
scouting plans, both for the U.S. (Blue) Navy and for the opposing navies. Large numbers
of naval aircraft, including seaplanes, were needed to scout over the vast areas involved.
Often, tracking an opponent’s fleet would begin with submarines deployed off of its base
to pick up the enemy force as it sortied, determining its initial direction. Much attention
went into techniques for shaking off scouts, and there was interest in alternative means
of scouting such as airships and long-range aircraft. In one game (Operations V, played
in April 1934 for the Class of 1934), for example, the Japanese side lost track of the U.S.
fleet after it sortied, and then guessed wrong as to the course it might take. It lost the U.S.
fleet for some time, wasting considerable effort in a panicked search.
The inter-war Navy was much smaller than the current one, not only in the number
of ships but in the number of officers and enlisted personnel. Mainly because the fleet em-
bodied much less advanced technology, the shore establishment supporting it was consid-
erably smaller. The current shore establishment in effect includes large companies that
support the fleet. The pre-war equivalent was far smaller. Inter-war Naval War College
classes always numbered fewer than 100. In the modern Navy, anything involving fewer
than 2,000 officers over two decades would have only limited impact. Such numbers mat-
tered a great deal more in a fleet that included no more than 9,897 officers in 1936. Even
in 1940, with the Navy growing rapidly due to the emergency presented by World War
II, there were only 17,723 officers.15 Until the rapid growth beginning just before World
War II, all of the officers were Annapolis graduates, which meant that many of them
knew each other. An officer returning to the fleet from the War College passed on what
he had learned, at least within his wardroom. Given very limited rotation policies, he
became a very important window onto the larger naval world, so the experience provided
by the War College had a much wider impact.
The structure of the inter-war Navy would also be unfamiliar to us. OPNAV was
nowhere near as all-powerful as it has become. It controlled the operational fleet, but the
materiel side of the Navy was handled by independent bureaus, such as the Bureau of
Aeronautics (BuAer), the Bureau of Construction and Repair (BuC&R), and the Bureau
of Ordnance (BuOrd), which were responsible only to the Secretary of the Navy. The
situation was complicated because the bureaus’ responsibilities conflicted. For example,
BuAer bought airplanes (and chose their main features), but BuOrd provided their weap-
ons. Similarly, C&R designed ships, but BuOrd developed their guns and fire controls and

Naval Transformation 19
armor, and BuAer developed the catapults and arresting gear onboard carriers. A separate
Bureau of Engineering (BuEng) developed ships’ power plants and also, as it happened, ra-
dios and sonars (BuEng and C&R merged in 1940 into the Bureau of Ships). The General
Board, compsed of senior officers appointed by the Secretary of the Navy, attempted to
coordinate the bureaus, for example in the area of ship design. The Chief of Naval Opera-
tions might or might not decide the outline requirements for new ships (which were called
“characteristics,” and were the province of the General Board). The bureaus of the past are
now mostly system commands responsible to the Chief of Naval Operations in a far more
unified naval structure, which began to coalesce during World War II.

The Strategic Problem

The inter-war period was a time of considerable U.S. strategic naval development. The
emphasis shifted from the Atlantic to the Pacific. Before 1914, the main scenario envis-
aged by U.S. naval planners was a reaction to a German attempt (actually planned by the
Germans) to gain a foothold in the new world by seizing territory in the Caribbean. The
main question was whether the U.S. fleet based at Norfolk could get to the Caribbean in
time to establish a base of its own. When the Naval War College gamed the operation in
1914, it also raised the question of whether the advancing U.S. fleet would be particularly
welcome when it arrived. In 1919, with the Germans defeated, the bulk of the U.S. fleet
passed through the Panama Canal to the Pacific. The most likely future naval opponent
was now Japan, which was known to covet all foreign possessions in the Far East—in-
cluding the U.S.-held Philippines. Planners contemplated a surprise Japanese attack on
the Philippines or, at times, a preemptive Japanese attack on the U.S. fleet to clear the
way for the conquest of the Far East. Surprise attack seemed likely because the Japanese
had opened their last major war, in 1904 against Russia, with a surprise attack against
the Russian fleet.
Looking toward such a war, the War College always emphasized the end-game, the
action it was hoped would force the Japanese to end the war. It was well known that
Japan had few natural resources, hence it lived by trade. The favored end-game was there-
fore blockade—strangulation. Blockade, in turn, required the U.S. Navy to deal with
thousands of Japanese and other merchant ships.16 The unstated implication was that
ultimately the U.S. Navy would use large numbers of small warships for this purpose,
and that these ships would have to be based on islands near the Japanese Home Islands.
Unless it was neutralized, Japan’s powerful fleet could sweep aside the blockaders. It fol-
lowed that in any Pacific War, the U.S. fleet would have to destroy or neutralize the main
Japanese fleet as a prerequisite to strangling Japan. None of the war games offered any
acceptable alternative to this basic strategy, and it is difficult to imagine one. The game
perspective, which emphasized the end-game of blockade as the objective, could not have
been gained in a full-scale exercise.

20 Winning a Future War

Battleships remained the core of the fleet through the interwar period because it generally took other bat-
tleships to sink them. The only aircraft capable of sinking a battleship were torpedo bombers. War College
evaluations showed that such aircraft were unlikely to be effective. Early torpedo bombers had low-powered
engines, so it took a lot of wing area to lift a heavy torpedo. The aircraft were so unwieldy that for a time the
U.S. Navy seriously considered eliminating them altogether. This is a T4M-1 on the deck of the carrier Lex-
ington, 8 December 1928. (NHHC NH 51371)

Clearly, blockade required that island bases near Japan be seized, and war planners
assumed they would have to do so. However, such seizure was not studied in detail. Is-
land seizures (generally in the face of enemy defenses) became important in the late-
1930s context of a new step-by-step strategy requiring intermediate bases.
Through the inter-war period, a key assumption was that the fleet could not be sent
to the Philippines during a period of tension that might lead to war; it would be mobi-
lized and dispatched only after war broke out. It followed that the fleet had to cross the
Pacific to the Philippines after war broke out, to fight the decisive battle that would open
Japan to the end-game of blockade—to seize command of the seas in the Far East. As the
think tank of the U.S. Navy, the Naval War College in conjunction with the OPNAV
War Plans Division had to find some way of getting there in the face of Japanese attacks.
The problem was made considerably worse by the post–World War I settlement that had
mandated many of the Pacific islands to Japan. The great question was how (or whether)
the fleet could pass relatively unscathed through the mandated islands (generally termed
the Mandates) to the Far East.
Until 1934, U.S. naval war planners generally assumed that the fleet could reach
Manila, or at least somewhere in the southern Philippines, either before the Japanese did

Naval Transformation 21
The Washington Naval Treaty suspended battleship development apart from improvements to existing ships.
The treaty envisaged additional protection against bombing and underwater damage. As the first post-war
Naval War College President, Admiral Sims used his new precision war-gaming technique to compare the
U.S. and British battle lines in simulated battles. He concluded that the British gained an advantage due to
the terms of the treaty. That is, reliance on traditional methods of analysis rather than his new one had led to
disaster. Later, it turned out that Sims’s analysis grossly overestimated the performance of British guns; in fact,
the U.S. battle line had the advantage. By that time, however, Sims’s analysis had convinced the Navy to press
for increased gun elevation (for greater range) in existing ships, a change the British argued violated the treaty.
Here the British battleship Barham, followed by HMS Malaya and the carrier HMS Argus, participates in joint
exercises of the Atlantic and Mediterranean Fleets about 1930. (NHHC NH 61776)

or in the face of Japanese opposition. The war plan evolved on this basis was called the
“through ticket to Manila.” Most, but not all, Naval War College strategic games echoed
this idea. The college had U.S. Army as well as U.S. Navy students. Many of them ap-
parently told their colleagues that they could not imagine holding the Philippines long
enough for the fleet to arrive, but they had no real impact. It appears that the crucial blow
to the “through ticket” strategy was experience in a 1933 war game, Operations IV. Once
its lessons had been accepted, strategy and the war games changed dramatically. From
1935 on, trans-Pacific war games always included the seizure by one side or the other
of an advanced base on an atoll. The fight for the atoll was simulated by a committee of
students, as a way of learning how to fight an amphibious battle.

22 Winning a Future War

 During World War II, the U.S. Navy developed a partial alternative to conventional
blockade in the form of unrestricted submarine warfare. The reason this strategy found
little favor in inter-war gaming is that the United States of those years lived in what to us
was a very unfamiliar world. Having refused to join the League of Nations in 1919–20,
the United States rejected any peacetime foreign entanglements (sentiment against the
British ran particularly high). In a war against Japan, many of the merchant ships in ques-
tion would be British, because Britain then had by far the world’s largest merchant fleet.
War games did show that unrestricted submarine warfare by the United States would
inevitably involve attacks on British ships, an unacceptable problem. This sort of conclu-
sion could not, incidentally, have been reached in any way except simulation—gaming.
Nor could it have been made obvious that submarines had no potential against Japanese
merchant shipping except by executing unrestricted submarine warfare (a conclusion ob-
vious from several games).17
Once the Germans triumphed in Europe in May–June 1940, the inter-war perspec-
tive changed dramatically. The United States found itself pushed toward an unprecedent-
ed peacetime alliance with the United Kingdom, and a substantial Atlantic Fleet was
created at the expense of the Pacific. However, there was no Atlantic equivalent of the
full-war campaign envisaged for the Pacific. To the extent that the United States might
expect to fight in Europe, that was much more an Army than a Navy responsibility. The
Navy’s wartime role would be largely to ensure the build-up of materiel in Europe and to
bring the Army ashore there. By this time, too, the Naval War College’s advisory role had
dwindled. Thus, for most of this book, naval strategy means the strategy of a Pacific war.

Naval Arms Control

The great facts of U.S. naval life during much of the inter-war period were the three
arms-limitation treaties, beginning in 1922 with the Washington Naval Treaty; and, for
most of the period, congressional reluctance to buy new ships. The Warren G. Harding
administration convened the Washington Naval Conference, which wrote the treaty, in
1921 specifically to outflank mounting Congressional pressure to stop the huge expen-
ditures associated with the existing U.S. naval building program. It had a temporary
advantage in that the British and Japanese, who had their own expensive programs, were
in much worse financial positions.18 The treaty effectively froze capital ship programs.
Among the consequences evident in many of the games was that it left Britain and Japan,
but not the United States with fast capital ships (battlecruisers). The cancelled U.S. pro-
gram had included battlecruisers. War games asked how (or whether) a more powerful
but slower U.S. battle fleet could deal with those fast foreign ships, and also how the
drastic U.S. deficiency in scouts (cruisers) could or should be addressed.19
The treaty had three other important consequences. First, it allowed each signato-
ry to convert two cancelled ships into aircraft carriers. In effect, it jump-started carrier

Naval Transformation 23
The shift to the Pacific emphasized the need for various kinds of scouting; war games at Newport emphasized
the difficulty of finding and tracking an enemy fleet over vast areas. The U.S. Navy tried a wide variety of
scouting platforms, including submarines and airships such as Akron (ZRS-4), shown here over Manhattan
about 1931–32. War games showed that such craft were too vulnerable to enemy fighters to be worthwhile,
although airship advocates continued to press for them as late as 1940. The same games featured a long-range
flying boat, the German Dornier Do X, which was more successful. (NHHC NH 43900)

aviation, because the resulting carriers were far larger than any navy would have wanted
at the time. Once the two huge U.S. carriers, Lexington (CV-2) and Saratoga (CV-3),
entered service they demonstrated what carrier airpower could achieve. The Naval War
College was well aware of the potential of naval air power, and its game rules were of-
ten somewhat more optimistic than current reality. Looking back, the major flaws in
the rules were excessive confidence in anti-aircraft fire, which was maintained even after
fleet experience showed that it was misplaced; excessive confidence in level and torpedo
bombing; and perhaps somewhat reduced confidence in dive-bombing. The rules did not
capture issues in fighter control, which proved extremely important in the war to come.
The second major consequence was strategic. As part of an ultimately failed attempt
to demilitarize the Far East, a Nine-Power Treaty signed alongside the Washington Na-
val Treaty prohibited further fortification of various Pacific islands. Manila Bay in the
U.S.-held Philippines was already fortified to a considerable extent, but now Guam could

24 Winning a Future War

not be fortified, and war games generally assumed that it would fall at the outset of any
Far East war. The question was how long it would take the Japanese to conquer the whole
of the Philippines, an assumed target in nearly all the Far East games.
A third consequence of the 1922 treaty and its 1930 successor, the London Naval
Treaty, was to set a ratio between U.S. and Japanese strength, measured in the total
displacement of capital ships: Japan would have 60 percent of the U.S. limit.20 Limits on
total tonnage were eliminated in the 1936 London Naval Treaty, the last of the series,
but not qualitative limits such as maximum ship displacement. The treaties thus raised
the question of how weapons not limited by treaty, which included shore-based aircraft,
could be used by the Japanese to redress the balance.
Other arms either not limited at all, or less strictly limited, took on far greater signif-
icance. For example, it was impossible to limit shore-based aircraft. Until 1930, neither
cruisers, nor destroyers, nor submarines were limited, at least in numbers. How could
their effects on a naval battle or campaign be estimated? What characteristics should
they have? The Japanese might be condemned to inferiority in numbers of battleships,
but could other types of ships tip the balance? War gaming, as it was practiced by the
War College, offered a way to understand the true balance of forces. That is why the War
College was asked to define a “balanced Treaty navy” in 1931. Its conclusions in turn
helped OPNAV shape a building program within the limits defined by treaties.
The U.S. Navy’s documentation of the inter-war treaties, in the General Board files,
seems to show conclusively that the Navy and its think tank had no impact on the U.S.
official positions during treaty negotiation. For example, General Board proposals were
simply rejected.21 The War College and others were consulted only after a treaty had been
negotiated.
The 1936 London Naval Treaty eliminated limits on total battleship, carrier, cruiser,
destroyer, and submarine tonnage, but as of 1941 it was still the old Washington and 1930
London Treaty limits that applied to the U.S. Navy. The reason was political. When U.S.
naval rearmament began in the 1930s, most Americans did not want to spend more on
the Navy. The country was deep in the Depression, and it seemed inconceivable that it
would have to fight in a few years. Naval rearmament was justified as a means of main-
taining a “modern Treaty navy,” the treaties in question being the Washington Treaty
plus the London Naval Treaty of 1930—not the escape from treaty limits that emerged
at London in 1936. That language was deliberately adopted in legislation to avoid fights
over specific new ships, such as battleships. Having renounced the treaty system in 1934,
the Japanese were not similarly bound. Because U.S. naval rearmament was keyed to
the Washington and London treaties, it took special congressional authorizations to add
carriers, one of which, Hornet (CV-8), was authorized in 1938. That was well below what
the War College had suggested.22
U.S. strategists, including the Newport war gamers, were well aware that the U.S.
economy dwarfed Japan’s. They assumed correctly, for example, that the U.S. aircraft in-

Naval Transformation 25
The fleet needed more effective scouts, so the U.S. Navy invested in large cruisers. To an extent, they replaced
the battle cruisers that had been barred at the Washington Naval Conference. At naval arms control confer-
ences, the Navy collided with the British, who badly wanted to limit the size (hence cost) of cruisers to keep
them affordable. The War College did not originate the cruiser requirement, but it was extremely important
in setting U.S. cruiser policy under the 1930 London Naval Treaty. Here three heavy cruisers of the Scouting
Force exercise together in the early 1930s. Note the floatplanes in the foreground, which extended the cruis-
ers’ scouting range. (National Archives 80-G-462561)

dustry would grossly out-produce the Japanese. However, they were also painfully aware
that building ships would take considerable time. It seemed that destroyers and perhaps
cruisers begun when war broke out might be completed in the second or even the third
year of a war. Anything larger—a carrier or a battleship—would probably be too late to
make much of a difference. As a consequence, when the Naval War College envisaged a
future war in, say, 1938, it could not envisage the sort of carrier-heavy U.S. Navy that
fought Japan in 1944. In the pre-war context, the only source of tonnage was an ally; as
early as 1931, the Naval War College was calling for a wartime alliance with the British.
World War II turned out very differently because the United States began to mo-
bilize in June 1940 when France collapsed and the prospect of a two-ocean war loomed.
Congress approved a massive appropriation intended to provide sufficient forces to fight,
if need be, in both the Atlantic and the Pacific. Given acceleration after Pearl Harbor,

26 Winning a Future War

the ships ordered in 1940 entered the war in 1942–43. Without such a source of fresh
tonnage, the U.S. Navy would have been in considerable trouble. The 1940 appropriation
provided the mass of fleet carriers that won the air-sea battles of 1944–45.

Ships23
The ships of the inter-war Navy, whose designs were much affected by war gaming, are
relatively unfamiliar to us. We think of carriers as the ultimate capital ships, their aircraft
capable of delivering decisive weapons—until fairly recently, including nuclear weapons.
The carrier aircraft of the late inter-war Navy were intended mainly to attack ships. Their
torpedoes could sink any enemy ship, including battleships, but experiments and games
showed that it would be difficult to hit a maneuvering target. The most potent air weap-
on, the dive bomber, could seriously damage or sink anything but a modern battleship. It
was widely understood that at some point aircraft would improve to where they would be
able to sink battleships at sea, but it was not at all clear when that would happen. World
War II did indeed show that aircraft could sink a maneuvering battleship at sea, but in
retrospect the main early-war example, the British battleship Prince of Wales, fell victim
to a lucky hit.24 The torpedo hit that doomed the German battleship Bismarck in May
1941 was possible partly because she had no supporting ships. The Battle of Midway in
June 1942 showed that U.S. torpedo bombers were far less effective than had been imag-
ined, and it was not until October 1944 that U.S. carrier aircraft (in numbers unimag-
inable pre-war) sank the Japanese Musashi.
The object of naval warfare, then and now, and very much in the war games, was to
gain sea control by sweeping away the enemy’s challenge to free U.S. use of the sea. During
the inter-war period, the ultimate challenge to free use of the sea was the enemy’s battle-
ship fleet, because, if the logic explained above is correct, it usually (if not always) took
battleships to sink other battleships. On the other hand, carrier aircraft could reasonably
be expected to sink anything short of a battleship, and carrier air reconnaissance would
generally keep carriers out of the way of battleships that might otherwise sink them.25
Carriers were also the only way to project power from the sea against most targets ashore.
Such potential targets included an enemy’s naval bases and the ships lying in them.
Given the objective of defeating the Japanese fleet in the western Pacific, any U.S.
strategist would have understood that the Japanese would seek to wear down (attrite)
the U.S. fleet before it got there. It took gaming experience to see exactly how that could
be attempted, and what the U.S. Navy could do to maintain its strength. The attrition
campaign, which had to precede a full fleet-on-fleet battle, emphasized carrier aircraft
and torpedo craft. To fight this campaign, the U.S. Navy had to transform itself from its
pre-war character (battleships supported by a few destroyers) into a true all-arms fleet.
This force could function effectively even after its pre-war battle line had been destroyed
at Pearl Harbor. Gaming simulated all-arms battles. In many of them, the battle line was

Naval Transformation 27
Submarines were also valued as long-range scouts, because they could operate off an enemy’s bases in the
face of his own seapower. The first new U.S. submarine designs of the inter-war period were conceived as
long-range cruisers valuable mainly as strategic scouts. Many inter-war games featured equivalent Japanese
submarine scouts trying to detect the U.S. fleet as it sortied and also to determine its course (in one game they
failed completely, the U.S. fleet went in an unexpected direction, and the Japanese found it nearly impossible
to recover). Here the tender Holland (AS-3) is shown with four long-range submarines, from alongside to
outside: Nautilus (SS-168), Barracuda (SS-163), Narwhal (SC-1), and Bass (SF-5), off San Diego, 10 November
1932. (NHHC NH 61589)

secondary or not even present. Full-scale fleet exercises did not simulate such battles, in
which the battleships played only a minor part.
Aside from carriers, the most visible new type of warship in the inter-war U.S. fleet
was the large long-range cruiser. In 1919, the U.S. Navy had three elderly fast cruisers
and another ten light cruisers under construction; it also had large obsolete cruisers. The
General Board had repeatedly but unsuccessfully asked Congress to approve additional
modern cruisers. In a pre-aircraft era, cruisers were vital scouts for a fleet and small cruis-
ers were the best way to beat off enemy destroyer torpedo attacks. In 1920, cruisers were
the occasion of the last clear victory of the General Board over Rear Admiral William S.
Sims and other rebels who wanted to supplant it with an all-powerful OPNAV. Looking
toward the Pacific, the General Board saw a desperate need for long-range scouts, a point

28 Winning a Future War

which many games would demonstrate. They had to be large, with the most powerful
possible guns. Sims, the wartime U.S. naval commander in Europe and the first post-war
president of the Naval War College, returned from service in London commanding U.S.
naval forces in European waters with much the point of view of his British hosts. They
preferred small cruisers because the Royal Navy needed numerous affordable, hence
smaller, ships to protect merchant shipping. The British found strategic scouting less
important, partly because they had effective long-range submarines, which they thought
could operate freely off of an enemy’s bases. Moreover, wartime British code-breaking
had provided a valuable measure of ocean surveillance. Looking back nearly a century,
the cruiser fight seems less than significant, but the later cruiser controversies of the early
1930s offer a useful gauge of War College influence.
The other major surface warship type of this time, the destroyer, seems to have been
affected only indirectly by gaming. The U.S. Navy began building modern destroyers in
numbers only after the War College and gaming had lost most of their influence. Howev-
er, even in these ships the impact of gaming can be seen. It was Naval War College games
that enforced the view that to survive to fight the main Japanese fleet, the U.S. fleet would
have to beat off heavy air and submarine attacks in the Mandates. This idea came up
repeatedly when the General Board, for example, discussed what sort of destroyers the
Navy needed. It was responsible for the decision to provide U.S. destroyers with a heavy
anti-aircraft battery directed by an elaborate fire control system.26 This choice was unique
among the world’s navies. It is difficult to dismiss the indirect effect of gaming in causing
it. Some of the later inter-war U.S. destroyers were also unique in carrying the heaviest
torpedo batteries in the world.
Another important new type was the long-range submarine. During World War I,
the Germans built cruiser submarines capable of crossing the Atlantic and attacking
merchant ships in U.S. and Canadian waters. Gaming made it obvious that such sub-
marines might perform valuable service as strategic scouts off Japanese bases. During the
war, the British built fleet submarines that could operate directly with a battle fleet, an
idea that had also attracted the pre-war U.S. Navy. Scout and fleet submarines featured
in many war games. The strategic scouting mission, more than the fleet submarine con-
cept, justified U.S. development of the fast long-range submarines that proved so effective
during World War II. Because strategic games emphasized the importance of various
types of strategic scouting, often by submarines, the characteristics of the fleet subma-
rines of World War II can be seen as consequences of the gaming process.
In the late 1930s, the U.S. Navy and its Marine Corps developed another unique
capability: amphibious assault in the face of beach defenses. World War I had featured
one major operation of this type, at Gallipoli, and other navies (British and Japanese)
seem to have concluded that successful assault could be conducted only away from enemy
shore defenses. That made good sense in Europe, with its long shorelines, and in China.
The U.S. capability against defended beaches was developed after a crucial 1933 game

Naval Transformation 29
showed that the fleet could not reach its intended destination unless it had intermediate
bases. From 1935 on, games at Newport generally featured attack and defense of fortified
islands, and the Navy conducted full-scale Fleet Landing Exercises (FLEXes). The Navy
also developed the specialized craft it needed. It planned emergency conversions of mer-
chant ships. They were conceived specifically for assault operations against islands, rather
than as conventional port-to-port transports. No other navy seems to have planned this
sort of assault force.

30 Winning a Future War

. T N W C  G

B efore World War I, the Naval War College functioned more as a staff adjunct to the
General Board of the Navy—its war planners and policy advisors—than as an edu-
cational institution. The General Board itself had been created in March 1900 by Secre-
tary of the Navy John D. Long primarily to resolve planning problems demonstrated by
the U.S. Navy during the Spanish-American War.1 It was deliberately made advisory be-
cause Congress strongly resisted the creation of a U.S. general staff. Lacking a large staff,
the General Board often turned to War College instructors and attendees for input on
policy, strategy, and force structure. It created its Second Committee in 1903 specifically
to prepare war plans, and between 1906 and 1910 the committee and the War College
worked together. At this time the college ran a “short course” built around the summer
conference organized to discuss particular questions posed by the General Board. War
College archives include tallies of which officers took which positions on such General
Board questions as the best position for a fleet commander in battle and the best tactics.
Although there were lectures on various naval and legal issues, answering these General
Board questions was at the heart of the War College experience. Students and instructors
spent most of their time discussing and gaming those questions, before presenting their
“solutions” to the college and the General Board, which usually summered in Newport.
The board also made use of the War College outside the conferences. The board valued
War College gaming as a way to test new technologies and warship design, sending the
college a number of designs and force structures to test. In the decade before World War
I, for example, War College analysis played a critical role in the development of “all–big
gun” battleships in the Navy.
The “big gun” recommendation exposed a fault line within the Navy. The bureaus
that designed and built the ships tended to be conservative. They resisted the General
Board’s call for a new kind of battleship (which, incidentally, would have predated HMS

The Naval War College and Gaming 31

Admiral William S. Sims returned to the War College as president in 1919. He transformed the course and
sought to make the War College the Navy’s think tank. Gaming became the core of the curriculum. Sims is
shown after his return from Europe, where he had commanded U.S. naval forces during World War I, on 7
April 1919. (NHHC NH 2839)

32 Winning a Future War

Dreadnought, which is usually considered the progenitor of such ships). Reform-minded
officers saw bureau resistance as rejection of the legitimate views of those who would have
to fight a future naval war. In their view, the fleet rather than the technicians should make
the key decisions. The General Board and the War College were steps in that direction.
A January 1908 article in McClure’s Magazine by Henry Reuterdahl, American editor of
Jane’s Fighting Ships, summarized complaints made by the General Board and by many
officers, including Commander (later Rear Admiral) William S. Sims, who had gained
prominence as a reformer in U.S. naval gunnery.2 He was serving as naval aide to Presi-
dent Theodore Roosevelt, who was personally interested in the Navy. Roosevelt was not
pleased with the article, because it argued that his “Great White Fleet,” which was then
steaming around the world to showcase U.S. sea power, was much less effective than it
looked. The article particularly criticized the new battleship North Dakota, the first U.S.
dreadnought.3
President Roosevelt used the Naval War College summer conference of 1908 to re-
solve controversy. Because the issue of the design of the North Dakota was raised, the
conference was called the “battleship conference,” although it was more concerned with
the way in which design requirements were framed. The President attended. Not sur-
prisingly the conference endorsed the critics’ view: Henceforth, the General Board would
decide the basic requirements—characteristics—of new U.S. warships. The fleet, not the
technicians, would decide what it needed. As the General Board’s think tank, the Naval
War College would now take a central role in deciding the character of the U.S. fleet as
well as its war plans.
It was understood that the critics were attacking the wider administration of the
Navy, particularly the independence of the bureaus, which answered only to the Secre-
tary. In 1909 former Secretary of the Navy William H. Moody, now a Supreme Court
justice, chaired a commission to study naval administration. He recommended that the
Navy Department be divided into functional areas, one of which (operations) would
be presided over by a Chief of Naval Operations. As before, Congress was unwilling
to adopt a naval general staff. Another board, headed by Rear Admiral William Swift,
offered a more modest version of the Moody recommendations, which was adopted in
October 1909 by Secretary of the Navy George von Lengerke Meyer. There would be
four divisions (Fleet Operations, Material, Inspection, and Personnel), each headed by a
rear admiral styled an “aid” to the Secretary. The Aid for Operations issued operational
orders to the fleet under the signature of the Secretary of the Navy.
The Naval War College moved from the old Bureau of Navigation to the new Oper-
ations Division, which would later become the Office of the Chief of Naval Operations
(OPNAV).4 That formalized its advisory role as the think tank of the operational side of
the Navy. In 1912, a British observer likened the long-term students and faculty of the
War College to a U.S. naval war staff.5 This role is easily forgotten if the War College is
seen as a purely educational institution.

The Naval War College and Gaming 33

 As—in effect—the staff of the General Board, the War College also became deeply
involved in ship characteristics decisions. These roles left little time for any educational
function. In 1911, War College President Rear Admiral Raymond P. Rodgers asked for
more staff; without it, he could not be both educator and advisor to the Operations Di-
vision and the General Board. Instead of providing it, in October 1911, Secretary of the
Navy Meyer formally limited Naval War College participation in war planning to work
compatible with its educational mission, which meant war gaming. The General Board
was given its own staff, but not its own gaming facility. Through the 1930s, successive col-
lege presidents insisted that they were not war planners. However, at the same time, the
war gaming conducted as the core of the Naval War College course informed the formal
war planners, who probably could not have worked effectively without it.
The aid (or aide) system did not last. Woodrow Wilson’s Secretary of the Navy Jo-
sephus Daniels, who took office in 1913, showed little interest in the naval efficiency for
which the system had been designed. Daniels was interested mainly in social issues, such
as prohibiting liquor on shipboard. Naval critics satirized him by saying that he was in-
terested mainly in creating a new position, the Aid for Education. The outbreak of World
War I in Europe offered Daniels’s critics within the service an opening. They thought
it entirely possible that the United States might be drawn into the conflict; hence, that
naval efficiency was suddenly essential. Daniels’s Navy Department seemed not to care;
Fiske saw only peacetime thinking. In January 1915, behind Daniels’s back, he turned to
Representative Richard P. Hobson (D–AL), a naval hero of the Spanish-American War
and a member of the House Naval Affairs Committee.6 Fiske’s staff, consisting of officers
involved in the previous fight over naval administration, drafted language that Hobson
incorporated in the current naval appropriations bill. It called for appointment of a Chief
of Naval Operations. Daniels opposed the proposal as “Prussianism,” but all he could do
was water down the bill. The Chief of Naval Operations was made responsible (under the
Secretary) for the operation of the fleet, its readiness for war, and its general direction.
He was given a 15-man staff, and the Naval War College was included in his domain
(initially the only other naval organization in the Office of Operations was the Office of
Naval Intelligence). The CNO could not issue orders in the name of the Secretary. The
act was passed on 3 March 1915.
Daniels’s revenge against Fiske and the other naval reformers who had come out of
the Naval War College was to pass over all the flag officers to appoint Captain William
S. Benson to the new post. Benson had attended the short Naval War College course in
1906, and he placed Naval War College graduates in key positions in his office. He also
managed to convince Daniels of his value, so that in 1916 new legislation much strength-
ened his office: All orders he issued were to be taken as coming from the Secretary. His
rank and title were raised to admiral.
Freed of much of the war-planning mission by Meyer in 1911, Rodgers turned the
War College more toward education for command. He created a new “Long Course” to

34 Winning a Future War

supplement the usual summertime “Short Course,” which had been intended mainly to
assist the General Board (the last summer conference was held in 1914). The first Long
Course was held in 1911–12 for four students. It was dwarfed by the first and second
committees formed for that year’s summer conference, with 16 officers each.7
The Long Course set the pattern for the war college. That year, the only course was the
one beginning in January 1914, with 13 students, five of them captains. In 1915 the War
College graduated two classes, one beginning in July 1914 (six officers, only one of them
a captain) and one beginning in January 1915 (14 officers, of whom eight were captains).8
The star pupil of the first Long Course was Captain William S. Sims, who used that
training in his next command, the Atlantic Torpedo (Destroyer) Flotilla. Sims discov-
ered that his new command had no real tactical sea experience. There was, for example,
no understanding of how to find and attack heavy ships in the open sea—his command’s
most basic task (the Royal Navy discovered in exercises at sea that this was an almost
impossible problem). Sims compelled his new command to develop practical tactics—to
bridge the gap between what could be drawn on paper and what could be done in re-
ality. Sims had been skeptical of the value of the War College course, which included
instruction in how to frame tactical decisions and also gaming to test tactical ideas. His
experience with the destroyers changed his mind. At Newport in 1914, Sims explained
that he used a combination of repeated conferences, including game-board work, before
maneuvers with the fleet, followed by thorough discussion afterward, just as the college
handled its own tactical games.9 Sims was rewarded by being appointed president of the
Naval War College when that position opened in 1917, and then promoted to rear admi-
ral.10 He planned to make gaming the core of the curriculum.11
Sims did not have long in this post. By this time, it seemed likely that the United
States would soon enter the war. The Germans announced unrestricted submarine war-
fare, which was widely expected to trigger war, on 1 February 1917. The German foreign
ministry soon sent the Zimmermann Telegram to Mexico, offering the American south-
west in exchange for a Mexican alliance. The combination of the submarine campaign,
which would inevitably sink U.S. ships, and the outrageous German offer to Mexico (sent
via U.S. diplomatic channels made available as part of a U.S. peace initiative) brought
the United States into the war on 6 April 1917. During February and March, as war
clearly approached, Sims was often called to Washington for conferences with the Gen-
eral Board and the National Advisory Committee. At the end of March, he was called
secretly to Washington and assigned to liaison with the Royal Navy. He sailed on 31
March, just before war was declared. The Naval War College course was suspended. Sims
spent the war as commander of U.S. Naval Forces Operating in Europe. He was already
considered an anglophile, and during the war he formed close relationships with the Roy-
al Navy. For example, in the absence of the British commander at Queenstown, a vital
anti-submarine center, he was given command of all naval forces based there, including
those from the Royal Navy.

The Naval War College and Gaming 35

 In London, Sims created the Planning Section, which attempted to apply War Col-
lege methods to the naval war.12 It would handle questions of immediate interest raised
both by the force in European waters and by the Navy Department in Washington, but
it would also raise questions of importance in the longer-term execution of the war. All of
this might be considered so obvious as not to be worth mentioning. U.S. officers used to
planning and to concepts such as the “Estimate of the Situation” got a glimpse of a very
different way of thinking—what might be considered the thinking of a navy without the
U.S. type of Naval War College.13 The Royal Navy created its own plans section (later
division) as part of a general reorganization in July 1917. It is not clear to what extent this
new group was a direct response to demands by Prime Minister Lloyd George.14
The U.S. planners were surprised that the Royal Navy had no real interest in their type
of work.15 The British had a pragmatic approach to everything; they disliked formal plan-
ning of any type. They thought in very specific terms, not asking whether some alternative
operation might offer a better payoff. The pre-war British decision to create a war staff had
had very little impact on British naval thinking. In particular, the British rejected a long-
range view of the naval war. They much preferred separate and often unrelated operations,
linked (if at all) in the mind of a senior officer. The Naval War College had long advocated
a formal planning process, although more for a single operation than for a war as a whole.
Sims’s London Planning Section tried not only to deal with immediate problems, but also
to anticipate problems that might arise. Its most spectacular projection was to point out
that the Germans might decide to send some of their battlecruisers into the Atlantic to raid
Allied convoys, which had no suitable protection. Its arguments were convincing enough
that the U.S. Navy based battleships at Bantry Bay in Ireland specifically to deal with this
possibility, and special signals were arranged to warn of a possible break-out.16
A few months after Sims arrived, the British naval staff was drastically reorganized.
Among its new elements was a plans division, which to Sims’s staff should have been the
core organization of the staff. American officers later wrote that they were the only ones
who took the plans division seriously; Sims’s Planning Section took the orphaned plans
division under its wing. CNO Admiral Benson was impressed enough with the success of
Sims’s section to recall one of its members, Captain Harry E. Yarnell, to set up a parallel
organization in OPNAV.17
The existence of a planning section in OPNAV opened a major question that the
legislation of 1915–16 had not resolved. How did the General Board, which was certainly
the peacetime planning organization, fit into the evolving Navy? Wartime naturally made
the operational naval staff dominant. It alone had the ability to frame and issue orders
on a day to day basis. Wartime records show that OPNAV also took most responsibility
for the massive wartime building program, designed mainly to fight the war against the
German U-boats. This threat had had almost no impact on the pre-war General Board,
whose big 1916 building program was designed to provide a U.S. battle fleet “second to
none.” The logic of the 1916 program was that, at the time, President Woodrow Wilson

36 Winning a Future War

was trying to protect American shippers from the British blockade imposed on Germany
and her allies. In Wilson’s view, British naval superiority made it possible for the British
to levy unreasonable demands. If the U.S. Navy was as powerful as the Royal Navy, the
British might find that much more difficult. The situation changed completely once the
United States entered the war. The General Board’s sin was not to have foreseen any such
shift, at least in framing its building program.
Admiral Sims ended the war as a hero, having presided over the U.S. part in the vic-
tory over the German U-boats. Offered his choice of post-war assignments, he chose the
Naval War College, which he now saw as the core of the U.S. Navy.18 He naturally want-
ed to complete the naval reorganization begun in 1915: The Navy should have one chief,
and that should be the Chief of Naval Operations. Sims considered the General Board
an anachronism now that OPNAV existed. In response to the demands of war, OPNAV
had grown enormously; the question was who would rule afterward.19 In the wake of the
war, CNO Benson reorganized OPNAV into formal divisions. The Naval War College
remained in the OPNAV organization, but was not a formal division. The divisions were:

• Policy and Liaison (Central Division)

• War Plans
• Intelligence (Office of Naval Intelligence)
• Naval Districts
• Fleet Training
• Materiel (Fleet Maintenance)
• Ship Movements
• Inspection
• Communications
• Secretarial

Of these divisions, Fleet Training was responsible for tactical development. In-
ter-war handbooks of tactical practice were generally published as FTPs—Fleet Training
Publications (or as USFs, United States Fleet publications).
None of the divisions included any sort of war-gaming capability to act as a lab-
oratory for planners. That role was reserved to the Naval War College. Sims told the
December 1919 Naval War College class that their course work and gaming would allow
them to “develop new applications of the principles of warfare as applied to modern naval
conditions . . . . They will also explain how the tactical game demonstrates the necessity
for new or modified types of vessels, or new or modified uses for those already built.”20
At times, Sims went further, arguing that the organization should not be called a
college at all. It was, rather, “a board of practical Fleet officers brought together here
to discuss and decide the extremely important question of how we could best conduct
naval war under the various conditions that may arise. You should think of this board

The Naval War College and Gaming 37

As a student, Harris Laning suggested that the War College collect lessons learned during war games and provide
them to future students. War College President Sims made him head of the tactics department, where Laning
wrote about Captain Joseph M. Reeves’s performance as commander of Blue aircraft. Laning returned to the War
College as president. Following that assignment, he was made commander of the Battle Force. Admiral Laning is
shown (left) in Hawaii in 1935 with Admiral Harry Yarnell, who, as commander of the carrier Saratoga, carried
out the surprise attack on the Canal Zone in the 1928 Fleet Problem. At this time, Admiral Yarnell commanded
the Fourteenth Naval District, which included the critical Pearl Harbor base. (NHHC NH 85318)

as belonging to the Fleet—as being what you might call a Fleet Board on strategy and
tactics, frequently making reports to the Fleet and the service upon these vital subjects.”
Unlike an ordinary college, Newport had no fixed policy, and neither its head nor its staff
was permanent. “They [students and faculty] are Fleet Officers and are continually being

38 Winning a Future War

replaced by other officers from the Fleet. The College is in effect part of the Fleet and
it exists solely for the Fleet. The students bring to the War College their practical Fleet
knowledge and experience. They are asked to consider this practical knowledge and ex-
perience in connection with the principles of warfare [which] are nothing but deductions
from the accumulated experience of those who have gone before us . . . . When you finish
your course you will carry these principles back to the Fleet, and will, I trust, be guided by
them.” Students who were assigned to the College Staff would transmit their experience
to the new class. Many of the students had never before been involved with strategy or
with some forms of tactics.21
Sims said that “some officers complain that they do not understand the terms, the
strange terms, used by the College. I do not understand any of the strange words used by
golfers, because I have never played the game, but I understand that some such words are
necessary. They are equally necessary for the game of war. Every art must necessarily have
its own rules, principles, and methods, and these must have names if we are to talk about
them . . . . This game, like all other games, can be learned only by playing it. The mere
study of the art of war, even though very thorough, will no more make you competent in
the practice of strategy and tactics than book knowledge of golf and tennis will make you
good players. It is for this reason that the College insists not only upon the study of the
art of war, but also upon the practice of it in chart maneuvers and on the tactical game
board. You will find that by playing these practical games you will gradually acquire con-
fidence in your ability to estimate a situation correctly, reach a logical decision, and write
orders that will insure the mission being carried out successfully.” Sims added that in the
war just ended effort and even many valuable lives had been wasted because of the lack of
training needed to reach logical decisions “based upon the well-known immutable princi-
ples of war . . . the constant prayer of those who bear great responsibility in time of war is
that they may be spared the results of the decisions of the so-called practical officers who
are ignorant of the art of war and who have not been trained to think straight—that is,
who have not been trained to make a logical estimate of a situation.”
Sims also argued that the college was so important to the fleet that flag officers
should be appointed to head its two key departments, strategy and tactics.22 CNO re-
jected that idea.
To make simulation effective, Sims ordered the preparation of war game rules far
more detailed than any yet used, including evaluations of how well various ships in dif-
ferent navies could survive battle damage. While this work was proceeding, the United
States convened the Washington Naval Conference, which led to a naval arms-control
treaty. At least in theory, the treaty left the United States at parity with the Royal Navy, a
result unimaginable before World War I. The General Board recommended approval on
that basis. Sims sponsored a study using the War College’s new method of fleet compar-
ison, and he publicized its result: The treaty left the British superior (it turned out that
when more accurate data were used, the U.S. fleet was superior).23

The Naval War College and Gaming 39

 Laning publicized another innovation first tried out in games: circular formations. They were in-
vented by Roscoe MacFall, shown here as a captain in 1945. Initially, their advantage as a cruising
formation was that the fleet could respond quickly to an attack from any direction—essential for
a U.S. fleet crossing the Pacific in the face of Japanese opposition. During World War II, circular
formations turned out to be well-adapted to multi-carrier operations; they were crucial to a fleet
concentrating large numbers of carriers, as the U.S. Navy did in the western Pacific. (NHHC
NH 48033)

40 Winning a Future War

 This study, doubtless intended to gain attention for the college as the fleet’s think
tank, seems to have been the first major post-1918 example of gaming-related arguments
used to reach important real-world conclusions. It and successor modified studies be-
came important factors in the first major post–Washington Treaty naval controversy.
The Washington Treaty prohibited major modification of the turrets of existing ships.
Using the same methodology as the 1922 study of battle line performance, it was possible
to show that British battleships could outrange their U.S. counterparts by flooding their
bulges intentionally and thus increasing gun elevation. At this time, the U.S. Navy, which
was becoming interested in longer-range gunfire as a way of gaining a decisive advantage,
wanted authority to increase the elevation of its guns. The British argued that the word-
ing of the treaty prohibited any such modification. Eventually the U.S. Navy rejected
British opposition and beginning with the Nevada-class (BB-36) ships were modified.24
Not surprisingly, Sims became the center of the post-war debate about Navy or-
ganization. The pretext for his involvement was differences with Secretary of the Navy
Josephus Daniels over U.S. Navy recommendations for the award of the newly created
Distinguished Service Medal for wartime service. Sims publicly rejected his own medal,
urged others to do so, and charged that Daniels himself had failed to prepare the Navy
for war.25 Sims’s attack on the existing system, particularly the General Board, was only
partly successful. The General Board retained its central role both in determining ship
characteristics and in laying out formal U.S. naval policy. The latter responsibility offered
the boardat least in theory a central role in the inter-war arms-control negotiations.26 A
quiet struggle between OPNAV and the General Board continued through the inter-war
period, the board losing most of its remaining power during World War II.27 However, it
managed to survive through 1950.
In this war, the Naval War College was initially connected to both sides. It advised
the General Board and it was part of OPNAV. Attendance at the War College was effec-
tively a prerequisite for service in the OPNAV War Plans Division. War gaming became
a key OPNAV laboratory technique. However, only one inter-war War College presi-
dent, Admiral William V. Pratt, became CNO. OPNAV retained its control over the
Office of Naval Intelligence (ONI). The close relationship between ONI and the college
is evident in the numerous intelligence reports that survive in the Newport archive (as
Record Group 8). To some extent these reports were a necessary way of keeping the war
games realistic, but they were also needed for the think tank role.
The Naval War College of the early-to-mid 1920s was essentially that envisaged by
Sims: a board of officers developing tactics and strategy by simulating future wars. In ac-
cord with this vision, the instructors were all drawn from previous students. At this time,
the principal departments were Command, Strategy, and Tactics. Command was headed
by the College’s chief of staff. The Strategy and Tactics departments ran separate games.
As president, Admiral William V. Pratt reorganized the faculty in 1926 to emphasize
the vital role of logistics as a separate discipline. The Strategy and Tactics departments

The Naval War College and Gaming 41

were merged into a single Operations Department to emphasize that strategic and tacti-
cal considerations could not be separated; to some extent that was due to the reach of air-
craft. The Command Department was renamed “Policy and Command.” It later became
the core of a War Plans/Policy and Plans Department. Pratt added an information office,
which later became the Intelligence Department. It received information from the Office
of Naval Intelligence, on the basis of which war game rules were constructed.28
Probably the most important change, instituted by Admiral Laning as Naval War
College president after 1930, was to abandon this practice and appoint long-term in-
structors (albeit still naval officers, generally Naval War College graduates). The Class
of 1933 (1932–33) was the first to work with a new department of research, and that
year there was a separate war plans instructor. Intelligence and research merged with the
Class of 1935.
Beginning in 1924, the College added a junior class (lieutenant commanders and
lieutenants) with its own staff and, in some cases, its own exercises. As in the past, the
senior class was intended to teach graduates how to wield a fleet. The junior class was to
teach them how to command major units. This book is concerned with the senior class
and the games it played, although it includes conclusions from each year’s “big game.”
This was a month-long campaign game played by a larger team combining the two cours-
es, the juniors taking up the more junior commands.
The college also ran a correspondence course. It attempted to teach the decision-mak-
ing technique employed in the war games, based on the “Estimate of the Situation” (see
Chapter 3). Without gaming, this is unlikely to have been very effective. The rise of the
correspondence course probably explains why, in the mid-1930s, considerable attention
went into producing successive much-expanded versions of the handbook Sound Military
Decision, sometimes called the “Green Hornet” after the color of its cover. Later editions
were roundly criticized as nearly impossible to read (or use), most likely because nearly all
readers came out of the correspondence course.29

42 Winning a Future War

. W G  W P

The “Applicatory System”

The inter-war Naval War College used an “applicatory system” of instruction introduced
by War College President Rear Admiral Raymond P. Rodgers about 1910. At the time,
it was considered revolutionary: It was learning by doing rather than via instruction.
“Applicatory” meant applying principles—having them inculcated by using them. Prior
to the applicatory system, much effort went into formalizing the principles of war and
teaching them on a formal basis. The students would not apply the principles until they
had to fight. The basis of the applicatory system was that memorizing the principles was
not nearly enough: A student would not even fully understand what he was being told
until he tried to apply them to a realistic situation. That went even more strongly for the
more technical aspects of naval warfare, as they might be reflected in maneuvering a fleet
against an enemy fleet. A later War College president likened the method to the case
method which had been introduced by leading law schools.1 Students did hear lectures
on the law, but they were forced to think through how they would apply their knowledge
in a series of cases. Many science students have experienced just this reality, that it is im-
possible to fully grasp physical laws without solving problems using them. For the War
College, the war games were the way students gained experience.
As employed at the War College, the applicatory method incorporated a systematic
approach to problems, as reflected in the form used to write orders.2 At least in theory,
this approach helped students apply their knowledge to the particular problems present-
ed. Concentration on practical problems did accelerate a tendency at the War College to
discard more academic pursuits; for example, the history chair was eliminated.3
Rodgers learned the system from his relative Captain William L. Rodgers, who had
learned it at the Army War College. Ultimately, it was derived from practices developed
by the German army, who had gained enormous prestige by their victory over France

War Gaming and War Planning 43

Admiral Sims’ great innovation was to tie the gaming used in the “applicatory” system to realistic data de-
scribing U.S. and foreign fleets. The great question, which was not really raised until the War College was
in decline as a think tank, was how well foreign navies were described. That applied particularly to the likely
adversary, the Imperial Japanese Navy. Japanese counterintelligence was excellent, and we now know that U.S.
(and, for that matter, British) knowledge was quite limited. Of the four Japanese capital ships shown in this
photo, taken in the 1920s, the Nagato (in the foreground) was substantially faster than imagined (26.5 rather
than 23 knots). The next ship was the Kongo-class battlecruiser Kirishima. The U.S. Navy certainly did know
that she was fast; the question (which gaming certainly could help answer) was how much of a handicap the
United States suffered because it had no fast capital ships. The other two ships, the battleships Ise and Hyuga,
held no great surprises when this photo was taken, but U.S. intelligence was unaware that when they were
rebuilt in the 1930s, they were given considerably increased speed. There was also no attempt to credit the
Japanese—or any other foreign power—with distinctive tactical style, although war gamers were told about
the peculiar Japanese command system in which army and navy were coordinated only through an imperial
council. (NHHC NH 111609)

in 1870–71.4 They seem to have invented the idea of mission-oriented orders in place
of detailed orders to subordinates: The subordinate was left to work out how to execute
the mission. Mission orientation offered subordinates much more initiative. They had
to understand the overall objective so that they could apply their own ideas successfully.
Mission-oriented command is often associated with an army’s ability to adapt to a rapidly
changing tactical situation, with forward commanders understanding what to do without
having been formally instructed. Without such orientation, an offensive might well stall
because officers who had already reached their initial objectives would not keep going.
Exactly that happened during the British amphibious assault at Gallipoli in 1915, when
troops facing little resistance stopped at their planned objective line instead of pressing

44 Winning a Future War

ahead. It might be argued that it was also what prevented British success during the
night action after the Battle of Jutland the following year. British CinC (commander in
chief, i.e., fleet commander) Admiral Sir John Jellicoe never learned what was happening
because no one apparently thought it necessary to report to him. As a consequence, he
never knew that the German fleet, which was withdrawing, had passed through his own.
Everyone apparently assumed that someone else had reported. After the battle, Jellicoe
emphasized the need to report.5 The idea of mission-oriented orders is still a very active
topic. For example, this command technique became an important issue in efforts to re-
form the U.S. Army during the latter stages of the Cold War.
To support this command technique, the Germans developed a decision-making
method beginning with formal determination of the mission—not a trivial problem,
since the orders from above would necessarily be much more general than those a subor-
dinate would have to issue. The statement of the mission was followed by an assessment
of enemy forces and intentions, assessment of own forces, and then evaluation of alterna-
tive courses of action to reach a formal decision. The next step was to write orders using
a standard order form based on German practice. Its five parts corresponded to the steps
required once the mission had been identified. The Army War College typically evaluat-
ed the student’s solution using either a map exercise or a staff ride. The latter was a ride
to examine the terrain in which the battle would be fought to examine whether planned
dispositions and possible enemy actions were practicable.
The Naval War College equivalent of the Army’s preparatory document was the “Es-
timate of the Situation.” Its core was a list of possible enemy courses of action followed
by the corresponding list—written after working out enemy courses of action—of own
courses of action. The student had to guess what the enemy’s preferred course of action
might be. No one ever described it this way, but comparing courses of action by both sides
was not too different from the usual game theory exercise of tabulating possibilities and
then evaluating likely payoffs for both sides before choosing a course of action that mini-
mizes risk or seems to maximize payoff. This technique imposes a discipline on whoever
uses it because listing what the enemy can do limits wishful thinking.
The inter-war cruiser problem, which will be discussed in detail later, is a good il-
lustration. Asked for advice after the conclusion of the 1930 London Naval Treaty, the
Naval War College first laid out courses of action that the other major navies—the Brit-
ish and the Japanese—could and likely would take. It based its advice on those estimates.
That might seem obvious, but the British experience at the same time shows what could
happen if this process was not followed. The British objective during treaty negotiations
was an attempt to end the competition in large expensive cruisers that the Royal Navy
was finding unaffordable. The British assumed that, just as they were driven mainly by
economic necessity, other navies would welcome the chance to limit themselves to more
affordable ships. They surmised further that the cruisers had become so large only be-
cause they had to accommodate the largest allowable guns. If only smaller guns were

War Gaming and War Planning 45

The war-gaming data system included rules to calculate battle damage. They were as conservative as possible:
It was assumed that ships would succumb only slowly to damage. No allowance was made for sudden cata-
strophic damage, such as that which befell the British battle cruiser (or fast battleship) Hood in May 1941,
when she was sunk by the German Bismarck. Hood is shown in Gatun Lake, on the Panama Canal, in July1924.
(NHHC NH 60404)

permitted, surely all the world’s navies would build only cruisers comparable to the less
expensive ones the Royal Navy planned. British records show no attempt whatever to
look at the cruiser issue through others’ eyes. They were unpleasantly (and expensively)
surprised when neither the U.S. Navy nor the Imperial Japanese Navy followed suit.
Both preferred large, expensive cruisers armed with larger numbers of the smaller guns.
As a consequence, the British found themselves compelled to build far more expensive
ships than those they had previously contemplated.6
If an officer understood the mission, he could apply standard tactics. Many of those
who espoused the applicatory method saw it as the basis for introducing doctrine into the
fleet.7 For example, as commander of the Atlantic destroyers, Sims developed a menu of
tactics. He likened them to the plays of a football team; each member could execute them
simply by hearing the number of the play called out by the quarterback. Each member
was responsible for applying those plays to a mutually understood objective. Mutual un-
derstanding was crucial: It was created by gaming and conferences (reviewing the tactics

46 Winning a Future War

Because little was known about Japanese and other foreign aircraft, the War College used the details of current
U.S. naval aircraft for both the U.S. Navy and foreign navies (there were some exceptions). These Japanese
aircraft are shown on a carrier, about to attack Pearl Harbor. (National Archives 80-G-71198)

used in the games). In modern terms, it might be said that Sims was determined to pro-
mulgate commander’s intent rather than simply instructions.
World War I Royal Navy experience, particularly at the Battle of Jutland, showed
why it was vital that the Navy share a common doctrine based on pre-war exercises. The
British seem to have assumed that upon elevation to command, the fleet commander
would immediately gain (or quickly develop) the ability to work out the necessary tactics.
This belief in the genius of a single senior officer is reflected in the decision, about 1910,
not to publish a tactical handbook because that would unduly cramp fleet commanders.
Put crudely, officers were not expected to think for themselves until they reached fleet
command, at which point they had better think for the fleet as a whole—without much
real preparation for that role. At best, a future fleet commander like Jellicoe might enjoy
some experience as deputy to a fleet commander (as Jellicoe actually did).
In effect, the Battle of Jutland was a test of the British system of command. Critics of
Admiral Jellicoe’s performance have argued that his centralized style of leadership failed
to develop individual initiative. Probably the worst result was a failure both during and
after the battle to report enemy sightings and positions. That was extremely unfortunate,

War Gaming and War Planning 47

since Jellicoe relied on his subordinates to feed data into the plot he was maintaining.
He did not share the nature or importance of the plot with the commanders of his ships,
and they in turn often failed to report. The failure of individual initiative has also been
blamed for his ships’ failure to fire at the Germans when they came into view during the
night after the battle. Jellicoe might have argued in the latter case that it was better to
withhold fire than to risk fatal errors of identification at night, but this was the sort of
question that would have been raised in Sims-type conferences and in war games.
Jellicoe certainly did try to envisage all possible contingencies, often using small
model ships his flagship’s crew made for him to try different formations and tactics. He
encouraged gaming in other ships. However, ultimately his Grand Fleet battle orders
were designed to control individual initiative, and he expressed only limited interest in
educating subordinates to understand the logic of the orders so that they could function
effectively if communication broke down. Neither Jellicoe nor his subordinates had ex-
perienced simulated battle situations. For example, no one seems to have understood just
how the fleet would be affected by poor visibility. The Royal Navy did conduct many full-
scale exercises, but, unlike the U.S. Navy, they do not seem to have provided subordinate
officers with a sense of what was happening and of what was needed. Nor, apparently, did
the Royal Navy share a common menu of expected tactical moves.
Without much sense of Jellicoe’s guiding principles, it might well have seemed that
far too much responsibility had rested on him. What would have happened if he had
been killed, or had collapsed—as he did the night after the battle? How could continuity
of command been assured? The elaborate Grand Fleet battle orders offered little guid-
ance for reactions to unexpected developments. For example, no one had worked out a
response to the torpedo attack and retreat the Germans carried out.8 Prior to the battle,
Jellicoe simply counselled that the fleet should not follow the Germans into a submarine
and mine trap. He may have been inspired by German tactical instructions warning that
the British might do the same thing.
In fairness to Jellicoe, it should be noted that on taking command of the Grand
Fleet he wanted to institute decentralized divisional tactics, but did not think that he had
enough time before he went to war to train the fleet properly. The kind of decentralized
authority that Sims envisaged in his destroyer force required considerable indoctrination
and experience on the part of subordinates. When Jellicoe took command, the Royal
Navy had no such tradition, and Jellicoe could argue that he needed several years to re-
verse decades of an attitude that the fleet commander automatically knew everything and
therefore ought to be followed blindly.9
Every post-war navy studied Jutland intensely, not because such a battle had the
slightest chance of recurring, but because it offered so many examples of the way in which
the reality of war affected command and tactical performance.10 The British experience
mattered to the U.S. Navy first because the Royal Navy was the largest and most pow-
erful in the world, hence a model for a rapidly-growing U.S. service. It was also the navy

48 Winning a Future War

Gaming scenarios were based on a realistic understanding of current political issues, but there was no way to
estimate the political impact of particular operations. This did not mean that it was entirely ignored. During
the 1930s, a gamer detached a carrier to attack Tokyo. A later summary of lessons learned from gaming
mentioned the operation and added that there had been no way to evaluate it (in the game, the detachment
of the carrier carried real penalties). When a U.S. carrier force, supplied with Army B-25B bombers, actually
did attack Tokyo in April 1942, consequences included the disastrous Japanese decision to attack Midway in
hopes of drawing out the U.S. fleet. B-25Bs are shown on the flight deck of the carrier Hornet (CV-8), ready
for launch against Tokyo. (NHHC NH 53289)

that provided the U.S. Navy with its experience of modern naval warfare during World
War I. U.S. cruisers, destroyers, submarines, and sub-chasers were assigned to the an-
ti-submarine war, in many cases integrated with Royal Navy organizations. A U.S. battle
squadron was assigned to the Grand Fleet in 1917–18. U.S. officers found the Grand
Fleet battle orders (or instructions) far more sophisticated than their own much simpler
ones. Overall, exposure to the Royal Navy during World War I was a terrific shock to a
U.S. Navy that on the eve of war had considered itself very competent.
However, the unsatisfactory outcome of the Battle of Jutland showed that the Brit-
ish system was limited—that the playbook envisaged by Sims really was necessary. It
justified the Naval War College, which was the source of the common command culture
the U.S. Navy needed. Command education produced an officer capable of using the

War Gaming and War Planning 49

Commentators discussing games often pointed out that new construction would take so long that it probably
would not influence the outcome of a war. As it happened, U.S. mobilization began in 1940, and new ships
began to appear as early as the end of 1942. In effect, the entire pre-war U.S. fleet was more than replaced, and
the mobilization program was responsible for the enormous growth in the U.S. carrier and cruiser forces. All
the ships in this photo of Task Force 38, maneuvering off of Japan on 17 August 1945, were completed after
war broke out. They include six Essex-class carriers, of which Wasp is closest to the camera, four light carriers,
at least three fast battleships, and cruisers and destroyers. (National Archives 80-G-278815)

playbook—and, incidentally, capable of fitting into an existing tactical organization even

if he had not had years of experience within it. A common command culture had to be
created, in which all officers understood a common approach to thinking through tactical
and strategic issues.
In place of the prescriptive Grand Fleet battle orders, the inter-war U.S. Navy creat-
ed tactical handbooks best described as playbooks—compendiums of alternative tactics
and formations from which a fleet commander could choose.11

* * *

50 Winning a Future War

War Gaming
William L. Rodgers became War College president in 1911. He combined the applica-
tory system with war gaming.12 A student might list an enemy’s courses of action, but in
reality the enemy had a vote, which could be simulated only by gaming. A game could
pit each side’s estimates of the other’s decisions against the other. It could also tease out
surprises an enemy might spring, in a way that no direct analysis could.
The Naval War College seems to have invented the use of gaming to explore the
problems of fleet tactics. In all other navies it was apparently assumed that experienced
officers would instinctively know what to do when they assumed command. Even full-
scale tactical experimentation was a relatively new idea, introduced into the British Med-
iterranean Fleet by Admiral Sir John Fisher in 1899, and probably not known outside the
Royal Navy. The U.S. Navy badly needed the gaming technique, both because it had no
recent combat experience (the war against Spain had involved weapons now quite obso-
lete) but also because its officers had little experience of fleet operations of any kind—the
Navy had grown too rapidly for that. Given its ability to conduct naval war games, the
Naval War College could use them to develop new tactics against a background of rapid
technical change.
Through the inter-war period, the War College emphasized that it was teaching stu-
dents to outwit capable opponents, who might have equivalent weapons and ships and
aircraft. Much is often made of the racism that is said to have contaminated U.S. pre-war
thinking about Japan. It is not to be found in gaming. Game documents did include what
were supposed to be racial characteristics of the enemy, but in the case of Japan the usual
special characteristic was fanatical courage and devotion to duty. It was not stupidity or
rigidity. To emphasize that an opponent would be quite as intelligent as the U.S. Navy,
the War College rotated its students between the U.S. and enemy fleets. There was no
attempt to shape the behavior of an enemy fleet in accord with known foreign doctrine
or tactical style.
Note that the War College generally did not refer to its exercises, at least during the
inter-war period, as games. They were chart or table maneuvers, the name suggesting that
in some way they were comparable to the full-scale maneuvers played out during the Fleet
Problems (the term gaming was revived after World War II). Strategic games were chart
maneuvers; tactical ones were table maneuvers, because they were typically conducted
using a large table marked out in scaled squares. In the 1930s, a checkerboard floor was
laid down in Pringle Hall, making it possible to simulate battles over wider areas.
In 1911, the War College already used tactical games to experiment with warship
design and also with tactics. For example, the U.S. Navy’s Battle Plan No.1 was devel-
oped on the basis of gaming at Newport.13 It was a matter of considerable pride, to the
extent that a U.S. naval attaché in Italy commented about 1910 that the Italians had
only just progressed to such sophistication. In 1917, Battle Plan No. 1 must have seemed
juvenile in comparison to the voluminous British Grand Fleet battle instructions issued

War Gaming and War Planning 51

In contrast to ships, the War College assumed correctly that the large U.S. aircraft industry would quickly
out-produce the Japanese. Perhaps more importantly, war gaming showed that pilots would suffer a very
high wastage rate. The wartime U.S. naval air establishment turned out vast numbers of pilots. Neither of
the other two carrier navies seems to have understood the need for pilots in any numbers. That suggests that
without gaming, the high wastage rate would not have been imagined. Very rapid production of pilots made
it possible to keep large fast carrier forces in action on a sustained basis, pilots and airplanes being replaced
while the ships remained in a forward area. This type of replenishment was as important as the more evident
development of a fleet train to provide fuel, ammunition, and stores on a sustained basis. The prototype escort
carrier Long Island (CVE-1) is shown as an aircraft transport, off of the California coast, 10 June 1944. She is
carrying 21 Hellcat fighters, 20 Dauntless dive bombers (SBDs), and two J2F utility airplanes on her flight
deck, plus other aircraft stowed below; clearly her flight deck is far too packed for her to operate any of them.
(National Archives 80-G-236393)

to U.S. battleships joining the Grand Fleet. Battle Plan No. 1 and its relative Battle Plan
No. 2 were descriptions of how to fight under a particular idealized set of circumstances.
The Grand Fleet battle instructions were designed to respond to the widest likely variety
of circumstances. Moreover, they were plans for a combined- arms fleet, in which guns
and torpedoes should be used together in complementary fashion. The U.S. battle plans
described a gunnery fleet, its destroyers a wholly separate arm. Encountering the Grand
Fleet battle instructions was a great shock for U.S. officers attached to the Grand Fleet
during World War I.
However, the Battle Plans held the germ of an advantage. They were a corporate U.S.
Navy creation, not the fruit of a single tactical genius, if that is what Jellicoe was. They
were not identified with the man who happened to be fleet commander when they were
adopted. The current fleet commander and other senior officers certainly had a great deal

52 Winning a Future War

to do with the way in which the battle plan was shaped, but no single officer was respon-
sible for it. It followed that the gap between them and something like the Grand Fleet
instructions could be filled by a corporate U.S. naval effort. That effort, in turn, could
draw on experimentation, both full-scale and in gaming.
The inter-war U.S. Navy used gaming as the laboratory to test tactical ideas. The
full-scale Fleet Problems taught the fleet how to work together, and to some extent they
tested the assumptions built into the games. Only the games could simulate full naval
campaigns as well as situations, such as massive night battles, which could not safely be
exercised by ships and aircraft. Conversely, the game rules were translated into the um-
pire rules used to evaluate the outcome of the Fleet Problems.
The U.S. Army had its own Army War College. At times, it and the Naval War Col-
lege conducted joint exercises involving overseas landings. The Army War College seems
to have adopted the applicatory system. However, it did not adopt gaming (the Army
may have used gaming at its tactical-level school, the Command and Staff College). Thus,
its contribution to a joint exercise was an elaborate set of logistical requirements and
complementary orders, but not a simulated land campaign after troops went ashore. The
Army certainly did become involved in national war planning through its membership on
the Joint Board, which ultimately approved U.S. war plans, but its war college seems to
have had only limited interest in the strategic level of war that the Naval War College so
often examined. That may have been a natural consequence of the difference between the
two services. An army ties itself to a particular land area; its job is to seize it or to defend
it. Withdrawal is painful and dangerous. A navy is not tied to any geographical area; it
can approach or leave as necessary. Particular places are generally expendable, to be revis-
ited later as necessary. This difference was reflected in the later profound disagreement as
to the appropriate end-game in the Pacific War.
To the Army, as reflected in its official history, the issue in the Pacific was Japanese
seizure of U.S. possessions, particularly the Philippines. The main goal of U.S. strategy was
to eject the Japanese from these places. This interpretation may have been affected deeply
by the personal identification that the senior Army officer in the Pacific, General Douglas
MacArthur, felt with the Philippines, but it seems to have run more widely within the U.S.
Army. The Navy’s view was that, having attacked the United States, Japan had to be defeat-
ed. Once that happened, whatever Japan seized would have to be disgorged. This view was,
incidentally, much more compatible with a coalition war than was the Army’s. In 1944, the
conflict between the two views was reflected in the Navy’s preference for seizure of bases
on Taiwan rather than a direct assault on the Philippines. At the gaming level, it is obvious
from the inter-war games that the objective was generally the Japanese fleet rather than
some particular place in the Far East. The unspoken but universal Navy assumption was
that once the Japanese fleet was gone, Japan would fall because only the fleet stood between
Japan and starvation. Places were valuable to the extent that they made blockade possible or
to the extent that their seizure would force the Japanese to expose their fleet to destruction.

War Gaming and War Planning 53

 Gaming could not and did not make the Japanese fleet an objective. That was fun-
damental. However, campaign-level gaming could and did give students a sense of what
the end-game of blockade and starvation required on the part of the United States. It also
avoided making any particular place important beyond its strategic significance.

War Gaming at the Inter-War War College

As Sims intended, war gaming—learning by doing—was the core of the inter-war War
College curriculum. The command course did not conduct games; instead its version of
learning by doing was to have its students write national policies corresponding to var-
ious historical events, particularly the Civil War and World War I. By doing so, they
could gain insight into why the various combatants had chosen the policies they actually
favored. The games and policy work were supplemented by extensive lectures; non-Navy
lecturers were particularly frequent after 1934.14
The core curriculum of the War College began with a course in constructing a valid
“Estimate of the Situation,” which described the mission and then listed courses of action
for one’s own and opposing forces. Deducing the mission from more general statements
was an important part of the exercise. It was not a trivial one, because the statement often
could be interpreted in terms of multiple possible missions. If, for example, the fleet was
covering an island assault while an enemy fleet was approaching, was the primary mission
to cover the assault or to destroy the enemy’s fleet in the interests of wider strategic needs?
Given the situation, the writer made a formal decision and explained its rationale. The
next step was order writing, because poorly written orders could sabotage any operation.
A game began with students receiving papers setting out a strategic or tactical sit-
uation. Students turned in their own estimates, which were critiqued by the staff. Since
only one estimate on each side could become the basis of war orders and hence of a game,
the students of both sides were issued with accepted estimates (incorporating strategic
choices) prepared by the staff. Many of these assigned estimates have survived; they are
much of the basis of the accounts of particular games in this book.
Once students had learned to write estimates and orders, they were taught to play
the war games the War College had developed. Students typically played at least eight
strategic and tactical games during their time at the War College after watching one or
two demonstrative games to learn how to play. Most of these exercises lasted a few days
at most, so they could not simulate the effect of a protracted campaign, in which both
sides’ ideas of how to fight would probably change due to experience. Each class also
played a month-long “big game” representing a protracted campaign (usually, but not
always, trans-Pacific). In the 1920s, the Naval War College worked with the Army War
College to conduct some joint games involving landings, generally unopposed. Typically,
the Army part of the exercise emphasized the logistics of a long-range Army operation,
including massive data describing exactly what the fleet had to move.

54 Winning a Future War

 A history of maneuver was kept for each game, and some have survived. Post-game re-
views (“wash-ups”) were often written up. They often included analysis by students compar-
ing reality as seen by the fleet with the college’s rules.15 As head of the tactics department in
1923 and 1924, Harris Laning was particularly concerned with tactical lessons the games
might teach; his analyses were particularly detailed. That applied particularly to the “big
game” played annually by both senior and junior classes. During the period the research de-
partment was in operation, it provided a formal post-game analysis, including lessons learned.
The full-scale Fleet Problems did not generally correspond to war games played at
Newport, but there were at least two exceptions. One game played by the Class of 1922
was intended to test the scenario for the upcoming Fleet Problem. For that reason, stu-
dents were asked not to discuss the game with anyone outside the college. It was an un-
usual scenario, a “bolt from the blue,” in which a Japanese fleet visiting the west coast of
South America proceeded suddenly to attack the Panama Canal and prevent the At-
lantic-based U.S. battle fleet from intervening in the Pacific. The other exception was a
partial one. Operations III of 1927 culminated in a battle in Narragansett Bay that was
plotted at the War College as it was fought. This also seems to have been the only War
College exercise to have received any publicity at all.16
To make games realistic, the enemy navy was typically either Red (British) or Or-
ange (Japanese). Each was credited with the ships it actually had, described as accurately
as the college’s intelligence department could. Its data were the best the Office of Naval
Intelligence could provide. Descriptions included ship capabilities and estimated vulner-
abilities. Many games simulated parts of the evolving U.S. war plan to fight Orange. It
was well known that the U.S. Navy considered Orange the likely future naval opponent;
Orange was certainly the object of most naval war planning.

Simulation
After 1919, the War College worked hard to make its simulation more and more realistic.
To do that, Sims ordered the preparation of war game rules far more detailed than any yet
used, including evaluations of how well various particular ships in different navies could
survive battle damage. It was apparently assumed that information about foreign ships
and guns was good enough to use directly, but it was understood that information about
aircraft was far less reliable than information about ships. The War College generally
simply adopted current U.S. aircraft data for both sides. Overall, the rules combined the
best estimates by the Naval War College Intelligence Department with actual experience
gained by the U.S. Navy in exercises, including the Fleet Problems. War College records
(Record Group 8 of the Intelligence and Research departments) include numerous in-
ter-war letters to and from college personnel conducting the annual revision of the rules.
The attempt to improve simulation made for far more complex war game rules than
those used before 1917.17 The new rules required a special staff, substantial documenta-

War Gaming and War Planning 55

tion, and, to some extent, special facilities. They were very much a War College function.
It was no longer possible for the fleet to play the games; they required the staff and the
facilities of the Naval War College. Rule books grew from perhaps 20 small pages before
1914 to about a 100 to 150 larger ones from the 1920s on.18 The rule books embraced so
wide a variety of topics that in retrospect they seem to be a good way to learn about the
way inter-war technology affected naval warfare. Scenarios included weather, and the rules
included estimates of how far away ships and aircraft could be seen at various times and
under various conditions. Communications sections included estimates of the time lost
in transmitting, coding, and decoding radio messages. Students relied on written ship-
to-ship messages, just as in real life they would depend on decoded radio messages rather
than voice radio. Students’ battle plans generally included frequency plans showing how a
fleet’s radio resources would be used. From the early 1930s on, the new technology of so-
nar was included, both for submarine detection and as a means of secure communication.
The effect of imposing increasingly complex rules was to stretch game time well be-
yond real time, as a great deal of measurement and calculation had to be done after a move
that might represent a few minutes in real time. Gaming equipment included calibrated
range sticks, so that the ranges between ships could be measured and fed into various
fire control and damage calculations. Supplementing the game rules were handbooks de-
scribing different fleets and different ships’ resistance to various kinds of damage. During
simulated combat, ships’ remaining combat capabilities were tabulated after each move.
Special efforts were made to keep students from seeing more than they would have from
the bridges of real ships. For example, the College ran special “quick-decision” games in
which students positioned their ships but “fog” simulated by screens kept them from see-
ing their opponents. The game opened by lifting the screens and the students were timed
to see how quickly they could decide what to do.
Generally, the prefaces to student orders listed the objectives of any particular game.
The most important were to develop the student’s ability to understand situations and to
frame decisions and to produce effective written orders. The orders, formalized as deduc-
tions from an estimate of the situation, were critically important to the gaming process.
Critiques of war games often mention poorly written orders that would have caused huge
problems in combat. They cautioned that while students might have a few days to formu-
late orders for gaming, they might have hours or less to do so in reality, and they would
be formulating orders under considerable stress.
Another objective was to familiarize the student with both U.S. and foreign fleets.
When they reached the War College, officers were very aware of the capabilities of their
own ships and usually of the units, such as destroyer flotillas, in which they operated.
They generally did not have much of a sense of how the fleet could or should operate as
a combined-arms whole. The War College often provided the only route for developing
that feel before being thrust into higher command.
A typical additional objective was to familiarize the student with the waters in which

56 Winning a Future War

the game was set. Again, a student who had spent much of his career operating in U.S.
home waters would not generally have a feeling for the geography of potential theaters
of war such as the western Pacific. Fighting notional battles or campaigns in a particular
region thus served as a familiarization exercise. Game data included local factors such as
normal weather during the period in which the game was set.
The rules were backed by elaborate fire-effect tables giving the effect of various weap-
ons on actual ships—or, at least, on ships whose details were given. They were first con-
structed in 1922 and revised in 1931.19 The 1922 publication was the first to attempt to
provide a mathematical basis for constructing tables from which the average fire of actual
guns could be evaluated in terms of damage to actual target ships in three-minute inter-
vals (representing single moves on the game board). In the past, hypothetical ships equal
in attack power and in survivability had been used; now actual navies were being mod-
elled. To enable gamers to compare damage inflicted by different weapons, a penetrating
14-inch hit was taken as the unit of damage and lifetime. Other weapons were assigned
capability as proportions of 14-inch hits. The original version of the tables did not take
long-range penetration of deck armor into account, because it was believed that no navy
had perfected the necessary fuse. In 1924, BuOrd contributed deck armor penetration
data, albeit with considerable caveats. The strength of the fire-effects method was that it
took into account both offensive and defensive characteristics. It summarized them in a
particularly convenient way. A battleship, for example, might have a complex system of
protection, but the War College could show how it compared to that of other ships under
a particular kind of attack.
The 14-inch gun standard for ship “lifetime” under fire made it possible to compare
different fleets and different weapons. The “Estimate of the Situation” generally included
tables of standardized firepower and ship lifetime. The idea was widely known in the
fleet. When future Rear Admiral Bruce McCandless wrote about his experience as a
young officer onboard the heavy cruiser San Francisco (CA-38) during the night battles off
of Guadalcanal in November 1942, he mentioned his ship’s nominal lifetime in 14-inch
hits, because she was being shelled by a Japanese battleship armed with just such guns.20
Despite the tables, San Francisco survived many more hits than the War College allowed,
partly because they were inflicted at short range, well above the waterline, a scenario not
accounted for in the rulebook.
Overall, the fire-effect system was deliberately conservative. The rules credited ships
with maximum survivability. Damage was assumed to be cumulative. The rules did not
allow for the kind of spectacular results single hits in lucky places sometimes achieved:
the explosions that destroyed HMS Hood and Arizona (BB-39), for example. In the
war-game tables of fleet data, HMS Hood was what the Royal Navy thought she was: a
well-protected fast battleship comparable to her battleship contemporaries. As evidence
of that reality, when Hood was sunk, the Admiralty warned the commanders of contem-
porary British battleships that their ships were no better protected.21 Under the same

War Gaming and War Planning 57

rules, single dive-bomb hits could not have sunk four Japanese carriers at Midway, al-
though they certainly would have put them out of action by destroying their flight decks.
In his memoirs, War College President Rear Admiral Harris Laning wrote that the War
College generally described the U.S. Navy as conservatively as possible, but that it al-
lowed considerable latitude to foreign navies.22 In that way, it prevented over-confidence
on the part of students playing the U.S. side.
The War College carefully avoided assigning superiority based on qualities of per-
sonnel or leadership to navies. Thus, it assumed that all navies had equally well-trained
personnel and comparable guns developing the same rate of fire. Estimates of firing er-
rors, such as dispersion, were based on U.S. performance. It is often assumed that the in-
ter-war U.S. Navy grossly underrated the Japanese on racist grounds. That was certainly
the case with the British, who discounted the strength of the Japanese battle fleet on that
basis when they estimated what margin of strength they needed in the Pacific. Individual
U.S. officers may well have had similar views. For example, one explanation of Admiral
Kimmel’s failure to provide against a Japanese carrier air threat to Pearl Harbor is said to
have been his belief that “the little yellow men” would not dare to mount such an attack. It
is therefore all the more striking that no such assumptions appear anywhere in the theory
or practice of war gaming at Newport.
The inter-war U.S. Navy had little or no knowledge of foreign tactical practices, be-
cause U.S. efforts to collect foreign intelligence were very limited. The best source of
tactical information would have been signals intelligence taken from observations of Jap-
anese naval exercises. Although the U.S. Navy did collect such information, its dissem-
ination was very limited for security reasons. It had no discernable impact on the Naval
War College. Students were free to choose tactics in simulated battles. The fact that the
Japanese players used tactics in games resembling those they actually used during the Pa-
cific War demonstrates that circumstances rather than some form of perceived Japanese
military culture determined much of what they did.
The method of analysis developed for war gaming had much wider significance. It
made direct comparison of navies possible, based on a combination of offensive and de-
fensive strength. For example, this means of comparison was employed by OPNAV when
it advised the Secretary of the Navy in the early 1930s on the appropriate composition of
a “balanced Treaty navy.”
While the new rules and the fire-effect tables were being developed, the United
States convened the Washington Naval Conference, which led to the 1922 Washington
naval arms-control treaty. At least in theory, the treaty left the United States at parity
with the Royal Navy, a result unimaginable before World War I. The General Board rec-
ommended approval on that basis. As touched upon previously, Sims sponsored a study
using the War College’s new method of fleet comparison, and he publicized its result: The
treaty left the British superior.23
This study, doubtless intended to gain attention for the college as the fleet’s think

58 Winning a Future War

tank, seems to have been the first major post-1918 example of gaming-related arguments
used to reach important real-world conclusions. Sims claimed that he had made two in-
novations. First, he took into account “the actual types of ships—their actual rates of
fire, percentages of hits, and the relative damaging effects of the different calibers [of
guns] when fired against the various thicknesses of armor at the various ranges at various
angles of impact. In other words the method now in use makes a comparison of fighting
strength between ships by taking into consideration their actual structure, armor, and
armament. The underlying principles of the present method of comparison are the same
as those of the past—that is, if the hitting power of the guns is known and the value of the
ships in ‘lives’ has been determined, then the Fighting Strength is equal to the Aggregate
Hitting Power times the Aggregate Life.” Second, Sims considered an actual engagement
in which factors other than gun power and armor would count. Only the British had bat-
tlecruisers. With their superior speed they could concentrate their fire on the head of the
U.S. line. The War College also pointed to an overall British one- knot speed advantage.
In the War College’s analysis, with the two fleets engaged on parallel courses at 15,000
yards, the British speed advantage would allow them to choose the range.
For all the precision Sims claimed, the British never felt particularly superior. They
considered their ships far too vulnerable to long-range shellfire, and argued that U.S.
ships were better protected. They had nothing remotely like Sims’s analysis on which to
fall back; it is not at all clear whether they were right. Sims’s analysis gives some idea of
the advantages the U.S. Navy enjoyed as a result of the War College’s new way and far
more precise manner of looking at fleets.
At least, that appeared to be the case. However, the apparent precision of the War
College technique hid a major problem. Errors in the data could tip the analysis. In this
case, the key error was over-rating the most important British gun, the 15-inch/42-caliber
which armed ten battleships and three battlecruisers. In Sims’s analysis, it was credited
with muzzle velocity so high that it outperformed the best U.S. gun, the 16-inch/45-cal-
iber. With a more realistic muzzle velocity, the British gun was midway between the
16-inch/45-caliber and the 14-inch/50-caliber that armed other U.S. battleships. On
this basis, overall the British battle line was inferior to the U.S.
Much depended, too, on assumptions about the armor of foreign ships and even
about how their shells were fused (because that determined how well shells could pene-
trate deck armor). The U.S. Navy of this era calculated zones of immunity for ships: the
range beyond which side armor could not be penetrated and inside which deck armor
could not be penetrated. The fire-effect tables included such data. Bands could be (and
were) calculated, inside which one ship could attack another without being endangered.
In war games, it was natural for fleet commanders to rely on such calculations as they po-
sitioned their ships. In 1937, a Bureau of Ordnance officer pointed out that all of this was
far too shaky.24 The U.S. Navy knew little more about foreign navies than what might
be read in open-source publications like Jane’s Fighting Ships. Some navies, but not the

War Gaming and War Planning 59

most likely enemy, Japan, did allow gun data like shell weights and muzzle velocities to be
published. Even that data was insufficient to estimate how shells would behave once they
struck, or how their fuses would act. Most side armor was visible and its thickness (on
the outside of a hull) could be estimated, but that certainly did not apply to deck armor.
Cruisers might have side armor integral with their hull structure, making observation
entirely irrelevant. Worse, by the late 1930s, armor quality was improving dramatically,
but the calculations did not take that into account.
The critic missed an important point. Having a way of evaluating and comparing
fleets—however flawed—forced War College students to think systematically and com-
paratively. Moreover, the range-band idea focused officers on seeking combat at greater
ranges, with an important if unintended consequence. To fight beyond 20,000 yards,
ships had to rely on spotting (to correct aim) by airplanes. Control of the air above a
battle line became an important theme well before the striking power of aircraft seemed
likely to be decisive. It opened the way for later development of naval air power in War
College thinking.
Quite aside from its impact on U.S. tactics, the War College analysis had import-
ant policy implications for the Navy. In 1924 the War College reported to the Navy
Department that the post–Washington Treaty U.S. fleet would definitely be inferior to
its British counterpart. That conclusion fed into an international controversy. The U.S.
Navy wanted to overcome British superiority by modifying turrets to increase gun ele-
vation, hence range. However, the Washington Treaty prohibited major modification of
the turrets of existing ships. The British argued that increasing elevation was a prohibited
improvement. Using the same methodology as the 1922 study of battle-line performance,
it was possible to show that British battleships could outrange their U.S. counterparts by
flooding their anti-torpedo bulges intentionally and thus increasing gun elevation. Even-
tually, the U.S. Navy rejected British opposition and beginning with the Nevada (BB-36)
class ships were modified accordingly.25
In 1930, the General Board asked the War College to revisit the question, comparing
the U.S. and British fleets after further reductions had been made under the new London
Naval Treaty. Now, the War College concluded that the U.S. fleet was clearly superior.
Even then, some of the data were not credible. For example, the War College estimated
that the British 15-inch/42-caliber guns would penetrate decks the much heavier U.S.
16-inch/45-caliber type could not. Rear Admiral Joseph M. Reeves reviewed and savaged
the War College report. He had served as head of the tactics department and he had
gone on to revolutionize U.S naval aviation (see Chapter 4); his opinions could not easily
be discounted. Reeves found that data for the British gun, which armed nearly all their
capital ships, had been handled inconsistently and hence arbitrarily. He was not amused.
How important was all of this? In a naval world apparently ruled by big-gun ships, it
certainly mattered politically. In 1922, many U.S. naval officers considered the Washing-
ton Treaty a sell-out by politicians because the agreement cancelled a building program

60 Winning a Future War

that would have given the United States superiority, at least in the most modern ships.
Within the Navy, Sims was probably seen as a hero for showing just how bad the treaty
had been. Outside the Navy, no one much cared; the pressure for naval arms control was
widespread and far too strong to overcome. War College analysis did help justify the
aggressive U.S. program of battleship modernization begun after the treaty was signed.
How important that ultimately was depends on the reader’s view of how important rela-
tive U.S. battle-line strength was in 1941.
In 1930, former War College President Admiral William V. Pratt had helped make
the London Naval Conference a success. He was soon made CNO. In that job he was
clearly determined to sell the treaty. It would have been surprising had analysis at an in-
stitution inside OPNAV not shown that the London Naval Treaty was a good deal for the
U.S. Navy. As it happened, that was true. The U.S. Navy had aggressively modernized
its battleships. The British had not, although they had built two very powerful new ships,
and they still had the advantage represented by faster ships (the three battlecruisers and
five Queen Elizabeth–class battleships).
As it happened, U.S. information on the likely future enemy, Japan, was poor. Just
how poor is evident in the surviving ONI/naval attaché reports in the U.S. National
Archives. Like the U.S. Navy, the Imperial Japanese Navy modernized its battleships. In
the 1930s, the Japanese spent very large sums to increase their speed; the Navy became
aware of the higher speed of some, but not most, of the Japanese ships only in 1936.26 The
Japanese speed advantage exceeded the known speed advantage enjoyed by the British. At
least in theory, it would have enabled the Japanese to choose their fighting range to suit
their level of protection. Also apparently unknown to the U.S. Navy, the Japanese were
very interested in fighting at maximum range. The U.S. Navy was certainly aware of the
advantage higher speed gave the Japanese battle line, and in the 1930s it decided to build
fast battleships specifically to deny them that advantage in future.
Assigning detailed capabilities to ships made war games seem realistic. It provided
students with a sense of the differences between their fleet and that of their possible
opponents. That would have been impossible without assigning details to the various
U.S. fleets. It is impossible to say whether the critiques of the fire-effects data mattered
in reality, because the United States never fought a World War II battle in which range
bands mattered, or in which the balance between the two fleets was close enough that
details of relative firepower and protection would have affected tactical decisions. That
was true even of the bitter cruiser-on-cruiser battles in the Solomons. The chief tactical
issues there had to do with the problems of night combat, which the fire-effects data did
not touch.
War gaming had a much more practical impact on U.S. war plans, fleet composition,
and tactics outside the battle line.

* * *

War Gaming and War Planning 61

Some Limits of Gamed Reality
The game rules missed important facets of naval combat, many of them connected with
the subtleties of command/control and air tactics. For example, although many games
featured night combat, it is evident in hindsight that the problems presented by night
battle were not properly simulated. Thus it is clear from the games and from tactical
handbooks that the inter-war navy thought it understood how to fight at night. The prob-
lem seems to have been that the inter-war U.S. Navy generally did not exercise at night,
for fear of fatal accidents. The Japanese and the British did practice to prepare for night
combat in the inter-war years. Their efforts paid off in successful night engagements such
as Savo Island and Cape Matapan.
Game designers admitted that air tactics were poorly simulated; they wrote as much
in the war game rules. That applied to both air-to-air combat and to attacks on ships. The
main implication of air-to-air rules in actual games was that fighter defense was unlikely
to be effective. Combat air patrols were generally ineffective in the games (as indeed they
would be until the advent of radar-based fighter patrols). Some games did hint at the need
to reserve fighters to deal with multiple waves of attackers, or with attackers approach-
ing simultaneously at different altitudes, in the sense that multiple attacks tended to
overwhelm the defenders. Overall, the games offered no hint of the need for, or the com-
plexity of, fighter control, only pessimism as to fighters’ ability to beat off attacks, which
was realistic in a pre-radar world. Air combat could really only be simulated in full-scale
engagements in Fleet Problems, not on a game board.
The exceptions to the conservative approach were attempts to test new weapons and
new types of ships. This meant aircraft carriers in the years before the U.S. Navy had
experience with them; an abortive new type of cruiser called a flight-deck cruiser; fleet
and cruiser submarines before these types had been developed; and the radio-controlled
Hammond torpedo in 1920. Games also employed drifting mines laid during naval bat-
tles and gas in various forms. Both were in various navies’ weaponry through the in-
ter-war period, but neither figured in World War II naval combat.
Reality could never be fully simulated, and it is important to keep the limitations
of gaming in mind. The number of players was always limited. Players notionally com-
manded units such as destroyer squadrons, but not individual ships. War gaming was
unsuited to modeling the lowest levels of war, and could not reflect the operational reality
of individual pilots or smaller warships. To some extent, the full-scale Fleet Problems
acted as a corrective.
The games also missed features of foreign practice that might be significant tactical-
ly. The inter-war U.S. Navy observed major Japanese fleet exercises, largely by intercept-
ing and decoding radio signals. No tactical insights gained in this way seem to have been
provided to the intelligence department at Newport. The entire code-breaking operation
was presumably far too secret to be risked in this way. Unfortunately, the records of the
signals intelligence collection operation are too vague to show whether it might have been

62 Winning a Future War

applied to the Orange side in the war games. However, it seems notable that there was
apparently no attempt to apply specific British practices, either, even though the Royal
Navy was far easier to observe. That may be because the games against Red (the British)
were intended mainly to test U.S. tactics against a fully equal enemy—as the Japanese
would be if they succeeded in destroying U.S. ships early in a war.
Although students at Newport spent considerable time evaluating national policy,
the games were never political, in the sense that they did not evaluate the political side
of any future war. They did involve U.S. politics, but only to the extent that, for exam-
ple, students might be told that the scenario was placed well after the outbreak of war,
and that the U.S. population was impatient for offensive action. There was certainly no
attempt to evaluate Japanese reactions to various U.S. national strategies. It was simply
assumed, not only by Newport, but also generally by U.S. war planners that a Japanese
government would come to rational terms if it faced strangulation by blockade. There is
no evidence that anyone imagined unconditional Japanese surrender as a possible out-
come of a Pacific War. This war was never envisaged as the fight to the death that U.S.
war planners came to foresee after bloody island battles beginning with the conquest of
Saipan in June 1944. Nor, for that matter, did the gamers ever envisage an invasion of
Japan. U.S. war planners assumed that the United States could never create a mass army
to fight in the Pacific. The only alternative they imagined to strangulation was a bomb-
ing campaign against Japanese cities, which it was assumed were flimsy and vulnerable.
Because the islands needed to support the bombing campaign were the same as those
needed to support a blockade, this possibility did not affect the logic of the Pacific War
as fought in games at Newport.

Using War Gaming

Toward the end of Sims’s tenure, one of the War College students, Commander Harris
Laning, pointed out that the war games were yielding valuable lessons.27 Surely it would
make sense to collate them as a source for future work. Laning was planning to spend
the next year at the Army War College; such cross-education was not uncommon. Sims
convinced him to stay on at the War College as head of the tactics department, and to im-
plement his idea. When Laning returned as president in 1930, he formalized the system,
creating a Research Department headed by Captain Wilbur van Auken.28 The research
department function of maintaining files of lessons learned seems to have lapsed when
van Auken left the War College in 1936.
Van Auken was outspoken in his War College analyses, which were embodied in
reports and in letters for the president of the War College—presumably for onward
transmission to CNO and the War Plans Division. These analyses survive mainly in
the files of the OPNAV War Plans Division rather than those of the War College. Van
Auken was well aware of the artificialities of the games, since he was also responsible for

War Gaming and War Planning 63

keeping the game rules up to date (they were revised annually). In 1931, he concluded
that the United States had neither the ships nor the bases (with repair facilities) it would
need for a war against Japan. In a remarkable postscript to a 1931 Blue-Red (i.e., against
the British) game (Operations III for the Class of 1932), he wrote that it was essential
to game and plan a cooperative war by both countries against Japan and its allies. Van
Auken referred both to a shortage of cruisers and to the impossibility of replacing losses
during a war. Van Auken repeated his comment the following year in commentary on a
Blue-Orange game (Operations II for the Class of 1933).
It says a great deal about the college’s ability to think freely and to avoid publicity that
van Auken was able to draw the conclusion he did in a document that circulated freely
through the college, and also through the War Plans Division of OPNAV, without any ill
effect. It is difficult to exaggerate how radical his proposal was. Popular U.S. sentiment at
the time, and for years afterwards, was squarely opposed to any sort of peacetime alliance
with the British. The belief was widespread that the United States had been lured into
World War I by the British in combination with U.S. East Coast bankers, who had lent
them vast sums and did not want to lose their investment. This explanation of recent his-
tory had real impact; it led Congress to pass a neutrality act that precluded future loans
and required that any arms buyer pay in cash (“cash and carry”). This act badly hobbled
British and French arms purchases in 1939–40. Moreover, van Auken was writing two
years after the big triumph of American arms control diplomacy (at the London Naval
Conference, which wrote the 1930 London Naval Treaty) had been to force the British to
accept cuts in cruiser numbers so that they would no longer be overwhelmingly superior
to the U.S. Navy. Given Van Auken’s analysis, this triumph had been disastrous: The
more cruisers the British had, the better they could join the United States as an ally
against Japan. Van Auken himself had written the college’s analysis of what sort of ships
the U.S. Navy should build under the London treaty. One of his assumptions had been
that U.S. politics would demand some sort of parity with the British in cruiser strength.
The unspoken assumption as the United States entered treaty negotiations was that the
goal was to cut overall British cruiser strength to what the United States could match.
Another point van Auken made struck home. The fleet had to have a mobile support
force far beyond what already existed. To some extent the desired auxiliary ships were
included in the building programs of the late 1930s. To a much greater extent, they were
built by the Maritime Commission created in 1938 specifically to provide both new U.S.
merchant ships and potential wartime auxiliaries. It seems fair to argue that it was the
war games that showed the need for many of those auxiliaries, particularly the ones that
constituted a mobile base and repair organization.29
In several cases, the War College staff applied gaming experience to vital current
questions. The cases developed in detail in this book are aircraft carrier and cruiser char-
acteristics and associated recommendations for the U.S. Navy’s building program. There
is also evidence that War College experience was reflected in the battleship program.30

64 Winning a Future War

On the strategic side, the OPNAV War Plans Division, which developed War Plan Or-
ange used against Japan during World War II, received and used Naval War College
gaming material: The surviving files include substantial amounts. Exactly how it was
used is less clear than in the case of the building program, however.
The historical record of lessons learned and transmitted is somewhat obscured for
the 1920s because both OPNAV and the General Board had direct informal access to
War College experience: The Naval War College was an OPNAV division and its presi-
dent was an ex officio member of the General Board. Only after 1930, when the president
was removed from the General Board, did the War College have to provide written com-
ments on various questions.31

War Gaming and War Planning

The Naval War College maintained a connection with the OPNAV War Plans Division,
but in the absence of correspondence files on either side it is difficult to say exactly how
close. The War Plans Division files include considerable Naval War College war game
material, much of it not duplicated in Newport.32 The large quantity strongly suggests
that the college was the think tank for the War Plans Division. Conversely, each year the
Naval War College played a month-long “big game.” At least the 1927 version was de-
scribed explicitly as a simulation of the war plan as then understood. In this connection,
it seems significant that attendance at the college, which meant experience in its gaming
technique, was considered a prerequisite for selection to the War Plans Division. At the
very least, that made the war planners sensitive to whatever surprises games simulating
their plans might reveal.
The “war plan” means the war plan against Japan. That Japan was considered the
most likely enemy throughout the inter-war period is evident in General Board hearings
on new types of warships, in which the question “how will this ship operate in the Or-
ange war” almost always arises. However, some war games, including some “big games,”
examined a Red-Blue scenario: war against the United Kingdom. A reader gains the im-
pression that no one seriously believed that such a war could break out.33 The usual stated
cause for such a war was economic rivalry, which in the 1930s was widely believed to have
triggered World War I. By way of contrast, Blue-Orange wars were generally triggered
by U.S. resistance (via the “Open Door” policy in China) to a Japanese attempt to gain
control of the Far East.
It seems likely that war games against Red were intended to teach students how to
fight a defensive war against a peer competitor. They would also familiarize students
with areas outside the usual U.S. operating sphere, the Pacific. For example, a defensive
scenario might be set in the Caribbean. It would give students insight into the strate-
gic significance of various islands and potential bases, very valuable experience once the
Atlantic sea war broke out in 1939. Most of the Red-Blue games were fought in U.S. or

War Gaming and War Planning 65

Canadian waters. The two typical scenarios were either U.S. resistance to a British expe-
dition to deal with a U.S. threat to Nova Scotia or U.S. resistance to a British attempt to
relieve bases in the Caribbean. In effect, the first reversed the usual U.S. problem of re-
lieving the Philippines in the event of a Japanese attack. The Caribbean scenario offered a
sense of which British Caribbean bases were worth trying to acquire, as the United States
actually did in 1940 in the destroyers-for-bases deal.
For three consecutive years, a very different Red-Blue scenario was played out. Like
Japan, Britain lived largely by her seaborne trade. The three “big games” focused on a
U.S. attempt to defeat Britain by choking it off. They included a U.S. attempt to estab-
lish a base off of the West African coast as a preliminary to trying to block the Straits
of Gibraltar. The games showed how difficult it would be to set up and support such a
base without any chain of intermediate bases. By analogy, it would be difficult to do so in
the Far East, if for some reason the fleet could not go directly to an existing base in the
Philippines.
This game showed how rapidly Blue’s cruiser force would be expended. That raised
the more general question of how wartime losses of major warships could be made up.
Even the huge U.S. industrial base would need time to build ships. The more complex the
ships, the longer it would take to build them. A cruiser ordered at the moment war was
declared might not be completed for two years. Battleships and carriers would take even
longer. Looking at the sheer scale of losses in the big trans-Atlantic trade war game, van
Auken wrote that the only way to gain sufficient cruisers to fight a sustained war against
Japan would be to ally with a country with a large cruiser force. The only candidate was
the United Kingdom.34
The United States did fight the Pacific War in alliance with the British, but the
Royal Navy was so fully occupied in European waters that it could not provide the U.S.
Navy with additional cruisers. What saved the U.S. Navy was that the European war,
and particularly the fall of France, inspired a massive building program begun in 1940
(under the “Two-Ocean Navy Act”). The 18-month head start, coupled with much more
intense working conditions once the war began, provided numerous new ships far earlier
in the war than pre-war analysts had imagined.35
The Blue-Orange war was of far more direct significance, and it accounted for most
of the “big games.” Students correctly saw the “big game” as a version of the current
Orange War Plan. Introductory material for the 1927 “big game” shows that this was
exactly what was intended.36 Until the mid-1930s, the game generally reflected the pre-
vailing concept that the fleet would begin the war by heading directly from its war base
in Hawaii to the Philippines. The latter would be the forward base from which the fleet
could fight and destroy the Japanese fleet, and thus open Japan to blockade. Under the
Washington Treaty, even the forward base in the Philippines was strictly limited. Fur-
ther fortification was banned, although existing forts were not destroyed. These already
controlled the entrance to Manila Bay. Manila’s repair facilities were limited to the float-

66 Winning a Future War

ing dry dock Dewey, which had been designed to handle battleships in service in 1907.
The ships in service during the inter-war period were about twice as large. Congress was
unwilling to buy a replacement dry dock or a permanent graving dock at Manila. Thus,
even if the fleet reached Manila, it had no way of making good underwater damage to its
capital ships. The battle-damage issue was exactly the sort of problem gaming empha-
sized, because games envisaged the entire campaign leading up to the entry of the fleet
into a Philippine port.
It was generally assumed that the Japanese would open the war with an attack on the
Philippines. Just like the Americans, they would realize that they were subject to stran-
gulation by blockade. Because the War College’s method of analysis including examining
a situation from the enemy’s point of view, assessments generally began with Japan’s need
to assure control of vital overseas sources of supply, mainly in East Asia but also farther
afield. The Japanese would have to secure the sea routes to the south. Because the Phil-
ippines lay astride such vital routes, it seemed obvious that any initial moves by Japan
would include the islands’ capture as well as the seizure of relatively nearby undefended
Guam. Conversely, unless a Japanese position in the Philippines was either overthrown
or neutralized, the desired blockade of Japan would be difficult.
The significance of the Philippines exemplified the difference between the strategic
view developed at the Naval War College and in the War Plans Division of OPNAV and
the Army view that the issue in a Pacific War was the recovery of a U.S. possession (the
Philippines) which the Japanese might seize. The strategic view was that whatever the
Japanese seized would inevitably be disgorged once Japan surrendered or came to terms.
The Philippines became an important strategic objective not because it had to be recov-
ered from Japan, but because it was an obvious part of the only end game that could be
envisaged.37 It seems arguable that this kind of strategic analysis was prompted by war
gaming and by the thinking it encouraged.
The Japanese actually had an alternative supply route. If they could make their peace
with the Russians, they could rely instead on the short shipping route across the Sea of
Japan from Siberia; goods could come from Europe via the Trans-Siberian Railway. The
(simulated) Japanese tried this in a few games, and it was not at all clear how the United
States could deal with this possibility. This possibility did not arise in reality because
even though Japan signed a non-aggression treaty with the Soviet Union, when war broke
out in the Pacific, access to goods was cut off by the war on the Eastern Front in Europe.
The question was how long the Philippines, or at least Manila Bay, could hold out
against a Japanese attack. In theory, the fleet could convoy sufficient troops to beat back
a Japanese invasion force—if it arrived in time. The war plan therefore envisaged the
quickest possible run from Hawaii to the Philippines—the “through ticket to Manila.”
To get to Manila, the fleet had to pass through the Mandates. It was widely sus-
pected that the Japanese had violated mandate agreements forbidding fortification of the
islands. If anything, this suspicion was reinforced by the Japanese themselves, who suc-

War Gaming and War Planning 67

cessfully prevented U.S. visits to the islands.38 Even without fortifying them, the Japanese
could base submarines, surface torpedo craft, and seaplanes there. This possibility trans-
formed the War College simulations. Instead of concentrating simply on the hoped-for
decisive battle against the Japanese battle fleet, the War College had to simulate the fight
between the U.S. fleet (or a massive convoy) and Japanese forces based in the Mandates.
The college correctly credited the Japanese with a strategy of attrition to balance
the odds for the decisive fight. The Washington and London naval treaties had left the
United States with a 5:3 superiority over the Japanese in numbers of battleships and in
total carrier tonnage. It was understood that in the ultimate battleship fight, the balance
of forces would be proportional to the square of the numbers. The Japanese could use air-
craft, which were not limited, to improve their situation. Thus, they could use unlimited
aircraft based in the Mandates to make up for limited numbers of carriers.
A subtle consequence of the assumption of a fight in the Mandates was that the carri-
er-based aircraft defending a U.S. fleet had to be able to counter the highest-performance
land-based aircraft. Had it not been for the Mandate problem, the U.S. Navy might
have accepted instead that its aircraft only had to be good enough to deal with other
carrier-based aircraft. Other navies, particularly the Royal Navy, accepted the idea that
carrier operation in itself limited aircraft performance. Through much of the inter-war
period, the British accepted that limitation on the theory that they would be fighting only
carrier-based aircraft. That left them in considerable difficulty when they faced land-
based Axis aircraft in the North Sea and the Mediterranean during World War II. They
seem to have been rather fortunate in that they had access to U.S. naval aircraft whose
designers had assumed from the outset that their fighters would face land-based attack-
ers. It also turned out that Japanese carrier-based aircraft were quite competitive with
land-based types.39
It took war gaming to force the fleet and the larger navy to understand the impor-
tance of the highest possible aircraft performance. The U.S. Navy was also fortunate
that the particular scenario its war planners envisaged required the fleet to fight its way
against large numbers of enemy land-based aircraft.
It was not universally agreed that the “through ticket” strategy would work, but the
“big games” reflecting expectations in the Far East generally involved something like it.40
It was by no means clear that Manila Bay and its defending positions on the Bataan pen-
insula and at Corregidor could hold out, but it seemed that it would take the Japanese
time to occupy the whole of the islands. The assumed destination of the fleet was shifted
well to the southern part of the archipelago, to Tawi Tawi or Malampaya Sound or Du-
manquilas Bay.41
The 1933 “big game” (Operations IV) tested this idea. As usual, the Japanese mount-
ed attrition attacks as the fleet steamed toward its southern Philippine haven. The south-
ern Philippine harbors were entirely unprepared, with no facilities for repairs (Manila
was not too much better, given the limited capacity of its one floating dry dock). The U.S.

68 Winning a Future War

fleet emerged from a battle still superior to the Japanese, but with most of its battleships
damaged by torpedoes. These ships were certainly designed to withstand such damage
without sinking, but they had to keep pumping to stay afloat. That would not have been a
problem had they arrived at a fully equipped base with dry docks, hence with the ability
to repair underwater damage. Van Auken pointed out what had to happen: As soon as
the ships stopped pumping, they would settle onto the bottom. They could not be re-
stored until something like a first-class base could be improvised.
Worse, the battle had not been conclusive. Japanese battleships also still remained
afloat, albeit damaged. Unlike the U.S. force, the Japanese fleet had fully equipped ship-
yards within steaming distance. Unlike the U.S. fleet that limped into Dumanquilas, it
could be restored to full capability. Neither side could complete new battleships in time
to affect the balance of sea power. Japan would emerge superior in the Far East, as though
the Orange fleet had won the big battle outright.
This conclusion was so important that War College President Rear Admiral Luke
McNamee, Laning’s successor, sent van Auken’s analysis and a summary of the game to
the Chief of Naval Operations. As far as the records of the War Plans Division show, this
was the only war game material ever sent to the CNO. The entire package survives in the
files of the War Plans Division.42 The game carried two implications, both of them fatal
to the “thrusting” concept. First, it showed that underwater damage would likely trump
whatever the fleet could do in the western Pacific at the outset of a war. Ships would al-
most inevitably be torpedoed, and they would have to go somewhere other than the Phil-
ippines for repairs. If enough of them had to be repaired at a rear base, the fleet would no
longer be superior to Orange’s. Given the agreements barring fortification of Far Eastern
bases, only Pearl Harbor could repair U.S. capital ships. Sending them back to Hawaii
would forfeit the Far East to the Japanese. Yet the U.S. fleet had to retain superiority to
win the engagement making the blockade—the end game against Japan—possible.
Second, in the Mandates the Japanese had exactly the resources—aircraft and sub-
marines—they would need to inflict the type of damage that would send the U.S. fleet
back to Pearl Harbor for repairs. Gaming showed that the Japanese would almost cer-
tainly be able to inflict sufficient underwater damage on enough ships. The U.S. war plan
changed. The westward advance would have to be step by step. The United States would
need an advanced base west of Hawaii with substantial repair facilities. Ultimately, that
meant development of floating dry docks and other auxiliaries that could turn an empty
lagoon into a major fleet base. Initially, it meant rapidly developing one or more perma-
nent bases.
A step-by-step advance could also change the terms of the air threat to the fleet. Once
an initial base had been seized, aircraft operating from it could strike Japanese air power
deeper in the Mandates, reducing the air threat the fleet would face as it seized further
islands. In effect this would reverse the expected Japanese plan to shuttle aircraft between
islands so that they could be concentrated against the U.S. fleet wherever it penetrated

War Gaming and War Planning 69

the area Japan controlled. To a lesser extent, long-range U.S. aircraft based on one island
could also, in theory, attack other Japanese assets in the Mandates, such as submarines.
From 1935 on, the War College’s “big games” always entailed the seizure of island
bases by one side or the other. Gaming material always included a special module involv-
ing such an operation. Specially assigned Marine Corps and Army officers conducted
this phase of the exercise. In 1935, when this was first done, War College President Rear
Admiral Edward C. Kalbfus told students that this time the “big game” did not reflect
the war plan. That was probably to protect the security of the radical change, from the
“through ticket to Manila” to a more gradual advance requiring the capture of islands.
That year, Army Chief of Staff General MacArthur signed a new version of the joint war
plan in which the relief of the Philippines was no longer contemplated.43
The 1935 strategic shift was seismic, with visible consequences. The Marines adopt-
ed amphibious assault as their primary future fleet role, and they sponsored development
of specialized landing craft (ultimately the “Higgins boat” and the amphibious tractor).
The Navy began to conduct Fleet Landing Exercises. Among other things, these taught
the fleet how to conduct bombardments in direct support of landing Marines. It is diffi-
cult to avoid associating the changes with Operations IV of 1933, particularly since the
OPNAV War Plans Division was using gaming as a means of testing elements of its plan.
It does not appear to have had any other means of simulation.44
All of the Orange games showed how vital scouting would be. Game solutions always
included scouting plans, generally employing long-range aircraft. The penalties of failed
scouting were demonstrated. In the most spectacular case, the Japanese failed to find
the U.S. fleet before it had reached the Far East, and the game showed their panicked
reaction. Various alternative long-range scouts were tried. For example, in some games
the fleet was provided with long-endurance dirigibles, sometimes equipped (as two U.S.
dirigibles actually were) with fighters for self-defense. Long-range flying boats, notionally
represented by the German Dornier Do-X, were also tried, and they were more satisfac-
tory. By 1933, the Navy could buy a U.S. equivalent, which became the PBY Catalina.
Unlike carrier aircraft, such long-range scouts were not limited by the arms control trea-
ties. They could move across the Pacific, supported by tenders that could operate from
unimproved lagoons.
It seems significant that by 1937 the fleet’s patrol aircraft had been reorganized. Be-
fore that, they were assigned to the Base Force at Pearl Harbor. Mobility meant that they
could move with the fleet to the Philippines, where they would be the fleet’s scouts pro-
viding support prior to the big battle against the main Japanese fleet. In 1937, the patrol
aircraft were made part of the Scouting Force. The building programs of the late 1930s
included new seaplane tenders as well as ships described as seaplane carriers. There was
also a program of small seaplane tenders (AVPs) designed specifically with draft shallow
enough to enable them to enter small lagoons.45 The big tenders made the seaplanes far
more mobile. Given the step-by-step strategy, they could move with the fleet. That would

70 Winning a Future War

not have been possible in the context of the “through ticket,” because the big seaplanes
could not be handled or launched or landed in the open sea. They needed temporary bas-
es in sheltered water from which they could hopscotch with the fleet in the step-by-step
advance across the Pacific. No one seems to have imagined that the U.S. Navy would gain
the ability to construct air bases ashore as rapidly as it did during World War II. Games
did not include any sort of rapid base construction.
Gaming explicitly avoided political considerations, except for an assumption that in
some cases U.S. public opinion would demand offensive action after any lengthy period
of consolidation. Nor did the Japanese players in any of the “big games” develop a Japa-
nese end game that might cause the United States to abandon a war. They concentrated
on resisting the U.S. strategic attack. This was realistic. Once Japan had seized the Phil-
ippines and gained a free hand on the Asian mainland, the best it could do would be an
offensive-defensive, the goal of which was destruction of a U.S. fleet steaming west. The
Japanese had to hope that the United States would quit at this point. The gaming system
never took into account the possibility that the United States, its population enraged,
would refuse to quit while creating a new fleet, as happened in World War II. Converse-
ly, because it strenuously avoided special assumptions about the enemy, the war-gaming
system never raised the possibility that the Japanese would refuse to come to terms when
they saw that they had been defeated—as happened in 1945.
It is difficult to see how the U.S. naval war planners would have reached the conclu-
sions they did without using simulation—gaming—to tease out the implications of their
work. It was easy enough, for example, to say that the Japanese would try something to
wear down the fleet as it passed through the Mandates, but gaming offered insight into
just what that probably would be. The Navy’s Fleet Problems were necessarily concen-
trated on the fleet-on-fleet battle that was expected to be fought after the U.S. fleet man-
aged to reach the Far East. It was gaming that offered insights into the rest of the war, or
at least into possible events prior to that big battle.
The alternative to gaming was to apply what would later be called military judgment.
It was a composite of wartime and exercise experience. As simulation, gaming often of-
fered results contrary to military judgment. They were not always at all acceptable within
the wider Navy, and some of the War College’s judgments were rejected outright in the
late 1930s. Perhaps the fairest evaluation of military judgment versus gaming would be
that in areas in which considerable full-scale experience had been accumulated, military
judgment was much more likely to be accurate. Gaming offered insight into wars that had
not yet been fought, involving new weapons—particularly aircraft. As we look back, the
shift in War Plan Orange—a shift which proved extremely beneficial—was one of those
areas.
Strategic gaming offered a kind of evidence about such warfare. It was particularly
important as the U.S. Navy shifted its preferred war plan after 1933. Whatever the war
planners might have said, it seems unlikely that their shift would have been accepted

War Gaming and War Planning 71

without the weight of that 1933 “big game” behind it. This was a major change, with large
political implications. If the fleet did not rush immediately to the Philippines, their loss
and the loss of their garrison would have to be accepted. The Philippines was the Army’s
important stake in the world outside the United States. The service could not easily ac-
cept loss of the islands, even in the context of a strategy designed to defeat Japan and thus
force that country to disgorge the islands later on. Junior Army officers in the Philippines
might be aware that the islands could not be held, but what counted in the mid-1930s was
the view from Washington. That was where it mattered that the president of the War
College provided OPNAV war planners with a solid forecast of what would happen if the
fleet tried to cash in the “through ticket.”

72 Winning a Future War

. W G  C A

C arrier aviation was both the greatest triumph and the greatest test of inter-war gam-
ing. It was the most radical technology that had come out of World War I, the one
for which previous experience was least relevant. War experience had certainly shown
that aircraft were important and had great potential, but they had not been involved in
battles. No one knew how they would perform in large-scale naval combat.1 The triumph
of gaming was both to show the U.S. Navy a way forward and to educate non-aviators
in what aircraft could and should accomplish. During World War II, U.S. non-aviators
such as Admiral Raymond F. Spruance and Admiral Frank Jack Fletcher (both War
College graduates) successfully wielded carrier forces. Non-aviators in other navies do
not seem to have done nearly so well.
In the immediate aftermath of the war, the closest the U.S. Navy could come to
simulating future air-sea warfare was the gaming facility at Newport. The Navy was con-
verting the slow collier Langley into an experimental prototype carrier (CV-1), but she
could hardly be considered a viable fleet unit. No airplane landed onboard her at sea
until October 1922, and she did not officially join the fleet until the end of 1924. She
was classified as an experimental ship until 1926, and her initial air complement was a
paltry eight aircraft. In 1919–21, there was active U.S. Navy interest in building a fast
fleet carrier. What actually provided the U.S. Navy with such ships was an accident:
The terms of the Washington Naval Treaty signed in 1922. As noted above, it allowed
each signatory to convert up to two capital ships, which would otherwise be scrapped,
into carriers. As it happened, the U.S. candidates were huge incomplete battlecruisers,
Lexington and Saratoga. As a result, these first U.S. fleet carriers were large and capacious.
Before they were completed, many thought them far too large, with too many aircraft
to operate effectively. The Japanese converted two more or less equivalent ships (Akagi
and Kaga); the British converted smaller and flimsier “large light cruisers” built during

War Gaming and Carrier Aviation 73

The carrier Ranger (CV-4), shown here in April 1938, was conceived before the War College could formulate
any concept of carrier characteristics based on war-game experience. When she was designed, it seemed clear
that the main requirement was for the carrier force to generate maximum immediate striking power, which
meant providing the maximum number of strike aircraft on deck for a first strike. Calculation showed that
the best compromise between flight deck area (which set the number of aircraft that could warm up at the
same time) and seakeeping within the limit set by the Washington Treaty was a relatively small ship with a
flush flight deck. By the time Ranger was being completed, trials of the two big carriers Lexington and Saratoga
had shown that it was best to provide an island for ship control; nothing could be done about the awkward
arrangement of tilting funnels (shown here in the “up” position). Limited displacement allowed no protection
and also favored a simple lightweight flight deck separate from the hull (the British and the Japanese used
flight decks integral with their carriers’ hulls). Ranger is shown in April 1938. (NHHC)

World War I (Courageous and Glorious). Until well after the two big U.S. ships entered
service late in 1927, no one in the U.S. Navy had anything but War College games as a
basis for deciding how they might be used. Of the foreign navies only the Royal Navy had
any operational carrier force at all, and the U.S. Navy had no access to its experience.
Americans were thus forced to debate the challenges and future of naval aviation without
much empirical evidence.2
* * *

74 Winning a Future War

Guessing What Aircraft Could Do
When he returned to Newport as War College president in 1919, Admiral Sims was well
aware of the need to understand aircraft and almost immediately ordered the preparation
of new war game rules to reflect their potential.3 By 1923, the aircraft rules were quite
elaborate.4 They indicate what the War College (and the fleet) thought that carrier air-
craft would be able to do once the carriers entered service. As such, they were a basis for
simulating naval war in that near-term future. Like other sections of the game rules, the
rules for aircraft were constantly revised on the basis of feedback from the fleet and from
other Navy experts. Asked to comment on the current rules in 1935, Chief of the Bureau
of Aeronautics (and later CNO) Rear Admiral E. J. King pointed out that aircraft were
the only naval weapon not yet really tested in war, so to some extent the rules reflected
opinion as well as experience.5
The 1923 rules distinguished between torpedo bombers (VT), fighters (VF), Scouts
(VS), patrol planes (VP), two- and three-seat observation planes (VO), and a project-
ed submarine-borne one-man scout floatplane (VSS). Now that the United States was
building carriers, most types could be produced as either land or seaplanes. If it could hit
its target, a torpedo bomber could deliver a significant blow using either a 1,650-pound
torpedo or a 1,500-pound bomb (or three 520-pound bombs). Fighters could carry
25-pound bombs, which were significant because they could wreck a carrier’s flight deck.
Aircraft speeds were considerably more impressive than in 1919: 95 knots for a torpedo
bomber, 135 knots for a carrier fighter. It was assumed that carriers (and other ships)
would stow their aircraft disassembled—it would take two hours to assemble a fighter
and six to assemble a torpedo bomber (a carrier would have six assembly crews). From
1926 on, the rules embodied the actual characteristics of current U.S. naval aircraft,
attributing the same performance to equivalent aircraft in other navies unless there was
no U.S. equivalent.6
The game rules envisioned four different classes of carrier: a large offensive carrier
(CVT), a small offensive carrier (CVO), a second-line carrier (OCV), and an auxiliary
carrier (XOCV). Lexington and Saratoga were CVTs; Langley was an OCV. A CVT was
an armored ship credited with a lifetime of twelve 14-inch hits, however they might be de-
livered. She might easily have her flight deck wrecked, but she would be difficult to sink.
The CVO displaced about 23,000 tons (which happened to be one third of the tonnage
available to the U.S. Navy after the two Lexingtons were completed), with a speed of 27
knots and a life of ten 14-inch hits. An OCV was a converted merchant ship with a life of
four and a half 14-inch hits. Finally, an XOCV was a large passenger steamer temporarily
converted into a carrier.7
At this point the CVT was credited with 18 VF and 6 VT on her flight deck plus
18 VF and 30 VT in her hangar and 18 of each stowed in her hold as disassembled
replacements. Other ships had proportionately weaker air complements. However, nei-
ther OCV nor XOCV carried any torpedoes. The rules assumed that half of the torpe-

War Gaming and Carrier Aviation 75

The War College played a very different role when a new carrier was planned in 1930. War College arguments,
based on game experience, led the Navy to reject the idea of building a small carrier specifically to work with
the battleships. Instead, a heavy fleet carrier was built. War-game experience also showed that carriers often
lost their aviation capacity to flight deck damage. The new carrier Yorktown (CV-5), shown on trials in May
1937, had several special features to keep her in action despite attacks. The light wooden flight deck (on a steel
supporting structure) conceived to save weight in Ranger was retained because, unlike the usual heavy steel
structure, it could be repaired rapidly by the ship’s force. Arrester gear was installed at both ends of the flight
deck so that the ship could recover aircraft even if one end of the deck was disabled. The roller blinds shown
in the nearly down position over the forward part of the ship’s hangar covered one end of an athwartships
catapult, which could launch aircraft even if the flight deck was entirely destroyed. Flight deck survivability
features kept Yorktown’s sister ship Enterprise (CV-6) in action well enough that she fought in every carrier
battle of World War II. (NHHC NH 42341)

does aircraft dropped would run properly. The courses of those judged to run properly
were plotted to see whether hits would be made. By this time, real warships were be-
ing attacked with bombs on an experimental basis, but it was difficult to say what the
chance was that a bomb or torpedo would hit a moving, maneuvering warship. Estimated
bomb-hitting probability declined as experience showed how difficult it was to target a
maneuvering ship.8
Through the inter-war period, the most salient reality of bomb damage was that it
would take very little to neutralize a carrier by wrecking her flight deck. According to
the 1923 rules, two 100-pound bombs would destroy a carrier flight deck. Ten 25-pound
bombs were credited with the same effect. The effectiveness of light bombs was important,
because from 1926 onward the U.S. Navy practiced dive-bombing, initially using fighters
limited to light bombs. This technique offered vastly better accuracy than conventional

76 Winning a Future War

The standard fleet carrier of World War II, the Essex class, was essentially an enlarged Yorktown. Essex (CV-9)
herself is shown soon after completion, (National Archives 80-G-68097)

(level) bombing. No single bomb was expected to sink a carrier (as actually happened at
Midway), but a carrier whose flight deck had been wrecked—a much easier goal—would
be completely neutralized.
In combat, much would depend on how quickly aircraft could take off and land. The
rules stated that a fighter could take off every 15 seconds, a torpedo plane or scout or ob-
servation plane every minute. The assumed elevator cycle was two minutes. An elevator
could bring two fighters or one torpedo bomber to the flight deck at the same time. To re-
cover aircraft, a carrier had to steam into the wind. She might need six minutes to steady
on that course. Airplanes could land every two minutes (every four minutes at night).
That was dangerous: The chance of wrecking an airplane while it landed was 5 percent
(10 percent at night). Flying was dangerous in general: There was a 4 percent chance that
an airplane would not make it back to the carrier at all.
The rules included considerably more detail, such as ranges at which aircraft could
see other aircraft and ships, and rules for catapult aircraft and seaplanes. Rules were also
given for air-to-air combat.

War Gaming and Carrier Aviation 77

 Looking back, the central questions for carriers were how many aircraft they could
operate and how efficiently; how much of a threat those aircraft represented both to other
aircraft and to enemy warships; and how vulnerable their carriers were to air attack (all
other forms of attack were easily taken care of because there was considerable experience
of them). Before 1928 and the completion of the two big carriers, all of this was guess-
work. There was very little experience on which to base rules.
By 1925, the experimental carrier Langley was in commission and the designs of the
two big fleet carriers were relatively mature. Operations with Langley provided initial
experience in bringing airplanes up to the flight deck and launching them. The interval
between arrival on the flight deck and take-off (to manhandle the airplane into place,
start and test its engine, and make last-minute adjustments) might be four minutes for a
fighter and five for something larger such as a VT. Once the airplanes on deck were ready,
a fighter could fly off every 15 seconds and a torpedo bomber every two minutes. With
proper preparations, a limited number of aircraft could be flown off twice as quickly.
The maximum size of such a quick launch (under the 1925 rules) was nine fighters or
seven VT. These figures explain why the U.S. Navy thought in terms of deck-load strikes,
bringing a mass of aircraft onto the deck, starting them up together, and then launching
them as quickly as possible. By this time, aircraft were considered sturdier, so the rules
offered a better chance that a landing airplane would not be wrecked.9
When carrier characteristics were finally included in the pamphlets describing the
fleets fighting in the games in 1926, they were standardized. Each carrier was assumed
to have the five-squadron organization then planned for the two big U.S. carriers: two
fighter squadrons (36 operating aircraft), two torpedo-scout squadrons (32 operating air-
craft), and one observation squadron (12 operating aircraft), with further knocked-down
aircraft in reserve (18, 16, and 6 respectively). A ship would also carry spares equivalent
to another 25 percent of the total number of aircraft (120).

Gaming and Early Carriers

The first major impact of war gaming on U.S. naval aviation thinking came in 1924. The
Washington Treaty allowed the U.S. Navy 69,000 tons of carriers beyond the two big
carriers, assuming that the experimental Langley was scrapped or converted to non-car-
rier use. How should that tonnage be used? Led by Rear Admiral William A. Moffet,
the Bureau of Aeronautics argued that the key issue was how to take the largest possible
total number of airplanes to sea. It would take a carrier to accommodate the larger naval
aircraft: the spotters, scouts, and torpedo bombers. However, although it would take a
carrier to recover them, the smaller fighters could take off from short platforms onboard
other types of ships. During World War I, the British had placed take-off platforms
onboard capital ships (on turret tops) and cruisers. Moffett pointed out that similar plat-
forms could be built onboard tankers and even destroyers. The new Washington Treaty

78 Winning a Future War

The special features of the U.S. fleet carriers were deduced from gaming—from simulated war— experience.
Other navies took very different approaches. The Royal Navy adopted armored hangars (the armor did not
cover the whole flight deck) on the theory that, when faced with air attack, carrier aircraft should shelter in the
hangar (U.S. carriers were expected to mount an active defense using fighters). Flight deck armor impressed U.S.
naval officers who inspected HMS Illustrious when she was repaired at Norfolk, but they apparently did not real-
ize that she was spending a lengthy period in the U.S. yard because any bomb which actually did penetrate that
armor would put the ship out of action for many months. U.S. carriers survived serious damage because they
had considerable deck armor below their flight decks. On the other hand, British armored decks were certainly
enough to keep out kamikazes, which badly damaged U.S. carriers by causing massive fires in their hangars.
HMS Victorious is shown about 1941, an Albacore torpedo bomber landing on board. The Albacore, with its lim-
ited performance, seems to have been accepted because the Royal Navy did not expect to face high-performance
enemy fighters. Again, war gaming forced the U.S. Navy to see things very differently. (NHHC NH 73690)

imposed no limits on the numbers of cruisers, destroyers, and tankers. Later, the tur-
ret-top flying-off platforms would give way to catapults, and they would typically launch
scouts and spotters, but that was not yet the case in the early 1920s.
Gaming offered a way of envisaging a war involving carriers. In January–February
1924 the War College played the “Battle of Siargao” (actually the Surigao Strait) as Tac-
tical Problem III.10 It was a convoy problem. Blue had secured most of the Philippines
and also held Guam as a forward base. The fleet in the Philippines badly needed supplies,
including oil, carried by a convoy from Guam.
Each side had a carrier (with some fighters onboard) plus fighters on turret-top plat-
forms onboard capital ships. Each of Blue’s 19 tankers had fighters onboard: The Blue
fighter force included 50 from the tankers and 12 from the six battleships escorting the
convoy in addition to fighters onboard the carrier. Blue’s cruisers carried only observa-

War Gaming and Carrier Aviation 79

tion planes. Blue chose to place its carrier in the center of his formation for protection.
This ship had to steam away at high speed when it launched aircraft. By way of contrast,
Orange chose to keep its carrier well away from his main force after it launched aircraft.
As a consequence, it was not seen by Blue scouts who spotted the rest of the Orange fleet.
Blue used carrier aircraft as scouts. Initially, Orange used scouts that were launched by
light cruisers.
Once they found the Blue fleet, the Orange scouts were to maneuver near the Blue
carrier in hopes of drawing off the defending fighters. Orange could therefore hope that
its bombers would reach the Blue carrier unopposed. Initially, the Blue fighters did chase
the Orange scouts, but they noticed that they were unarmed. Recognizing that they were
decoys, most Blue fighters did not chase them.11 The Orange bombers had no trouble
finding the convoy, as the weather was clear. However, the convoy was defended by nu-
merous fighters, including those onboard the tankers and the capital ships.
The Orange attempt to keep its carrier from being seen was frustrated because Blue
scouts saw the high-altitude fighter barrier that had been set up specifically to shield the
Orange fleet and carrier. Worse, although the barrier protected the Orange fleet to some
extent, Blue attackers could get at the carrier by flying around it.12 The barrier was too far
from the carrier to shield it, and it offered the rest of Orange’s fleet only limited protection.
As usual in a game, the element of chance was represented. In this case, Orange’s
strike approached the convoy just as Blue’s carrier headed into the wind to recover air-
craft to refuel them. The convoy still had its own fighter cover provided by airplanes from
the tankers. When Orange’s bombers headed for the carrier, fighters waiting to land (to
refuel) joined the defense. They shot down eight of 12 Orange attackers, and the others
made no hits. Meanwhile, Orange launched a second attack (6 bombers covered by 18
fighters). It met 29 (of the original 32) Blue fighters, which were chasing the surviving 4
bombers of the first strike back to their carrier. That forced the Orange bombers back,
but as soon as the Blue fighters flew back to their carrier the Orange strike re-formed.
Meanwhile, Blue was unable to find the Orange carrier. Faced with a possible air
attack, Blue chose to orbit a ready strike force over its carrier while more airplanes were
assembled on deck.13 When an Orange cruiser group caused Blue too much trouble, eight
of those bombers were sent out to attack. They were level bombers, so the targets largely
frustrated their attacks by zig-zagging. Even so, two hits sank one cruiser and another
inflicted 50 percent damage on a second. Blue submarines working with the convoy sank
another light cruiser.
The re-formed Orange strike was spotted by outlying Blue ships, giving the force
time to launch fighters. When the attack arrived, the carrier launched 18 fighters, which
joined 22 more waiting to land. The tankers contributed another ten and the Blue battle-
ships another four. This large fighter force drove off the attackers, who did not manage to
damage the Blue carrier. Two of the five Orange bombers that got near the Blue carrier
were shot down.

80 Winning a Future War

War gaming showed again and again that the U.S. Navy could never have enough carriers to fight a Pacific
War. Through the mid-1930s, the solution was to convert fast liners into carriers. The projected conversions
were elaborate, and they seemed to require much more time than would be available. They may also have been
dropped as plans shifted toward a step-by-step advance through the Mandated Islands held by the Japanese.
Such an advance would require numerous amphibious assaults and troop transport capacity would become
much more valuable. Once war broke out, the need for numerous carriers re-asserted itself. The main solution
was a series of simple conversions of existing merchant or auxiliary hulls (later some escort carriers were built
as such from the keel up, but their hulls were essentially those of merchant ships). The escort carrier Chenango
(CVE-28), shown in 1944, was built on a tanker hull. (NHHC NH 95703)

Blue finally found the Orange carrier the next day.14 Blue had already launched two
strikes against the Orange cruisers. Before a third could be assembled to attack the Or-
ange carrier, Orange launched a third strike against the Blue carrier. It sighted the Blue
force about the same time that a Blue battleship sighted Orange. Blue had 41 fighters
overhead, including aircraft waiting to land to refuel. They sufficed to drive off the Or-
ange attack, shooting down two bombers and protecting the carrier from any hits. How-
ever, many of the fighters were low on fuel. They could not land back onboard the Blue
carrier quickly enough. Seventeen of them had to ditch (their pilots were saved).
During the day, the Orange commander realized that he could not prevent the Blue
convoy from reaching Surigao. He was reduced to bombing it and also trying to drive off
Blue bombers holding down his submarines. He had to give up any hope of attaining air
superiority by attacking the Blue carrier. The attack on the convoy was more successful,

War Gaming and Carrier Aviation 81

To meet a perceived emergency need for fast fleet carriers, nine light cruiser hulls were converted on much the
same lines as the escort carriers (a proposal for a more elaborate conversion was rejected). The first of the class,
Independence (CV-22), is shown in San Francisco Bay, 15 July 1943. Comparison with Essex gives some idea
of just how small her flight deck was. This was very much not a concept developed by the War College, but
it can be traced back to the War College’s conclusion from gaming: Carriers were essential, in large numbers.
(National Archives 80-G-74433)

because the convoy had very little antiaircraft protection, and fighters assigned to protect
the convoy had gone off chasing Orange scouts.
Blue now launched an attack on the Orange carrier. It was not seen until it was nearly
overhead. Orange fighters already in the air attacked, but were not successful; Blue planes
managed to drop 66 bombs, of which eight hit. They demolished the ship’s flight deck.
Under the rules then in force, such damage could not be repaired outside a shipyard. Blue
now had air superiority. Moreover, Orange aircraft in the air when the carrier was struck
had to ditch.
Much of Blue’s success was attributed to the care the Blue air commander took to
conserve his airplanes by arranging to refuel them constantly. Unfortunately, only the
carrier could refuel these aircraft. Because it took so long for them to land on the carrier,
Blue lost a significant number of fighters, which had to ditch. Thus, the game emphasized
the need to be able to recover aircraft very quickly.

82 Winning a Future War

 In his analysis, Captain Laning, the head of the Tactics Department, pointed out
that aircraft could not easily be replaced during a campaign in the western Pacific. Heavy
losses were unacceptable. Conversely, to be viable late in a campaign, a fleet steaming west
from Hawaii would need to take very large numbers of aircraft with it.
In retrospect this was a remarkable game. It was not conceived as a test of carrier
aviation. At the outset, Orange’s main tactical plan was to mount a night torpedo attack
on the convoy. The players were not aviation fanatics; most likely none of them was even
a pilot. The players’ reasoning led them to see that air superiority was key to anything
else they might want to do. Both commanders in the game sought it; the carriers became
primary targets. Only in a game could the large numbers of aircraft that fought the bat-
tle have existed, at least at the time. Captain Laning was a surface officer without any
aviation experience. The Blue air commander, whose efforts he considered excellent, was
almost certainly Captain Joseph M. Reeves Jr.15 He would soon graduate and be chosen
as Laning’s successor as head of Tactics.
Surviving game material shows that, as head of the Tactics Department, Reeves re-
ceived another lesson in carrier operations while at the Naval War College. In Tactical
Problem II of the Class of 1925, two large U.S. carriers faced six much smaller British
ones, total air strength on both sides being about equal.16 Both of the big U.S. carriers
had their flight decks destroyed while only four of the six British carriers had their flight
operations similarly put out of action. The other two would have enforced British control
of the air (although that proved more difficult than many had imagined), with devastat-
ing consequences for the U.S. fleet.
In game after game, the lesson was that carriers had to achieve as much as possible
with their initial strikes, because they might be unable to launch further ones. At the very
least, the fleet with fewer carriers had to strike all of its enemy’s carriers. To do that, it
needed as many aircraft as possible. Much the same could be said of any attempt by the
carrier’s own fighters to defend her. Overall, games were ambiguous as to whether fighter
defense was likely to be effective.
At this time, the U.S. Navy had a single experimental carrier, Langley. Lexington and
Saratoga were still being completed (they would be commissioned late in 1927). Assum-
ing that the Langley would be replaced, the United States could build three 23,000-ton
carriers or five 13,800-tonners, or some combination of these ships, none of them displac-
ing more than 27,000 tons.17
There was no limit on carriers displacing less than 10,000 tons. The Bureau of Con-
struction and Repair prepared sketch designs of all possible carriers. It pointed out that
although larger ships would find it easier to handle more aircraft, a smaller carrier had
about 15 percent more deck area per airplane. The total number of aircraft a carrier
could accommodate would depend on deck area (hangar and flight deck). A larger num-
ber of smaller carriers would offer more total deck area. The bureau estimated that five
13,800-tonners could accommodate 6 percent more airplanes than three 23,000-ton-

War Gaming and Carrier Aviation 83

ners. It rejected the small carrier as useless for anything larger than a fighter, hence not
worth pursuing.
In 1925, Secretary of the Navy Curtis Wilbur asked the General Board whether it
would make sense to use the hulls of some of the recently authorized 10,000-ton cruisers
to build small carriers not limited by the Washington Treaty instead. The Naval War
College provided its own analysis, based on tactical games.18 The games showed that car-
riers were generally damaged by bombs dropped on their flight decks; it took only a few
hits to wreck a flight deck and eliminate a ship as a carrier. This type of damage could
be repaired only by a shipyard. Control of the air “has nearly always been obtained by
destroying the flying-on decks of enemy carriers.” This was more or less obvious, but the
lesson of the games was not: It took no greater effect to disable a large carrier than a small
one. Numbers were vital.
The War College President quoted the result of Tactical Problem II. Blue did about
twice as much offensive bombing as Red, so it knocked out twice as many carriers. Un-
fortunately, that was only four of the six Red carriers—and Red knocked out both Blue
carriers. Once Red had control of the air, its spotting aircraft represented a crucial advan-
tage. Using spotting from the air, Red guns could hit Blue at ranges Blue could not match
without deploying its own air spotters. When carriers were dispersed singly, attacking
each one required the same search effort, whether the target was large or small. It also
seemed that each carrier, large or small, could launch aircraft at about the same rate. A
larger number of smaller carriers could place more aircraft in the air more quickly. That
was likely to be a considerable tactical advantage. Against that, for the greater number of
aircraft to concentrate, the carriers launching them had to operate close together. In that
case a successful enemy search would locate all of them, and they could all be attacked by
the same enemy force.
Which was better, then—concentration or dispersion? Concentration increased the
number of aircraft that could work together. Dispersion made it more difficult for an
enemy to find all the carriers. This question was important through the whole inter-war
period and into World War II. At the Battle of Midway in 1942, the U.S. carriers were
dispersed to some extent. The Japanese carriers were concentrated to focus their striking
power. To some extent that concentration was their downfall, as the dispersed U.S. car-
riers managed to strike three of the four Japanese carriers in very quick succession.
In 1925, it seemed that the most important aircraft role in a major battle would be
to enhance the gunnery of the battleships inflicting serious damage on the enemy’s bat-
tleships. That meant seizing and using control of the air above the battleship fight. The
side that controlled the air could spot for its battleships, gaining them maximum range. It
seemed that small carriers could provide both the fighters maintaining air control and the
spotters exploiting it. It also appeared that small carriers could operate with the battle
line while the two new carriers Lexington and Saratoga mounted long-range strikes, op-
erating far from the battleships. Smaller carriers operating directly with the battleships

84 Winning a Future War

War gaming certainly affected the U.S. Navy’s idea of what it needed in its carrier aircraft. War games depicted
campaigns in which carrier-based U.S. fighters would face land-based Japanese bombers; the fighters had to
have the highest possible performance despite the limits a carrier might impose. The Vought Corsair, which
won the 1938 fighter competition held by the Navy’s Bureau of Aeronautics, had the highest performance in
the world (in terms of speed) when it flew in 1940. An F4U-1A flown by the Navy’s then highest-scoring ace,
LTJG Ira Kepford of VF-17, is shown over the Solomons in March 1944. (National Archives 80-G-217819)

would be easier for an enemy to find and destroy, but no such success would deny the U.S.
fleet all its airpower, because the two big carriers would still be available.
The War College was not, of course, equipped to say whether the small carrier was
technically feasible. For the time being it was dead because the ship designers of the Bu-
reau of Construction and Repair rejected the concept altogether. When the General
Board submitted a five-year building program in March 1926, it favored the 23,000-ton-
ner.19 The board subsequently reversed its position based on an April 1927 Naval War
College submission emphasizing the need for a maximum number of aircraft.20
The CNO asked the War College to comment on the General Board plan, based
on game experience. In reviewing past experience, War College President Rear Admiral

War Gaming and Carrier Aviation 85

William V. Pratt decided to include the results of the just-conducted Operations Problem
II He also took into account a new study of how deck handling would affect the outcome
of a carrier battle. Carrier deck handling was simulated using model flight and hangar
decks and cardboard airplanes. The detailed analysis was carried out by then-Lieutenant
Forrest Sherman, a future CNO. Pratt wrote that “in view of the present quiescent status
of the carrier question in the United States Navy,” the delay involved in completing this
detailed study seemed warranted.
How the U.S. Navy should allocate its carrier tonnage was controversial. Of five
officers at the War College who were queried, two favored a large number of small car-
riers, one a small number of large ones, and two a medium size (20,000 to 23,000 tons).
Clearly the large carrier (like the Lexington) offered convenience and coordination of air
operations, but no such ships could be built under the treaty rules. The intermediate size
offered similar advantages. Sherman wrote that it “seems to be a crystallization of British
experience.” Pratt observed, however, that the smaller carrier offered important advan-
tages: greater scouting area (each carrier could scout a given area, so more smaller ones
could scout a larger area); better overall security (due to more ships, as had been argued
before); mutual support increased by operating in pairs; and, not least, more available
aircraft. The last point could not be emphasized too strongly. Pratt wrote that
…one of the outstanding lessons of the overseas problems played each year
is that to advance into a hostile zone the fleet must carry with it an air
force that will insure beyond a doubt [emphasis in the original] command
of the air…. [T]he only way to secure it is by large numbers of carriers
and planes. This means not only superiority to enemy fleet aircraft, but
also to his fleet and shore-based aircraft combined. This is a large order,
and the only way to secure it is by a large number of carriers and planes.
If small carriers will put more planes in the air than the same tonnage of
large ones, we must have the smaller carriers, regardless of the small extra
cost per plane put in the air. This means quantity production, to which a
smaller type is much better adapted than a large one. Our shipyards (such
remnants as we have left) and our national engineering bent favor quantity
production…. Too, the smaller the carrier the quicker it could be built: we
may need to build carriers in a hurry.
Pratt added that seaplanes would be extremely useful in a Pacific campaign, a recom-
mendation that the Navy certainly heeded later.
Pratt reported that the games showed that no allotment of carrier tonnage within
the available tonnage would suffice to gain and retain air control. It would be unwise
to experiment with that tonnage, so he supposed that three 23,000-tonners might be
best. He hoped that numbers could be made up out of the unlimited 10,000-ton (or
smaller) class.

86 Winning a Future War

War gamers struggled to evaluate the dive bomber, the great attack development of the inter-war period. It
offered extraordinary accuracy; the question was how that would be affected by anti-aircraft fire and other
combat factors. The other important issue was lethality. The war game rules accepted that dive bombers could
severely damage or sink anything short of a battleship, but that typical battleship deck armor would defeat its
weapon. Thus, the carrier’s best weapon could affect the outcome of a surface battle—which it was assumed
would be decisive—by stripping the enemy’s battle line of all auxiliary ships, such as destroyers and cruisers.
Since torpedoes delivered by destroyers certainly could sink any battleship, this sort of action could well be
decisive in itself. War experience showed that this was a reasonable evaluation of dive bombing; it took torpe-
does to sink large battleships. Here, SBD-3 Dauntless dive bombers of VB-6 prepare for takeoff from Enter-
prise for the Wake Island raid of 24 February 1942. Bombs are not very visible because they were semi-recessed
under the fuselage. (National Archives 80-G-66037)

The games showed that speed was essential for a carrier. Whether a carrier could be
as fast as a cruiser depended on delicate design issues. Games did show that even when
operating with the fleet, carriers soon became detached as they turned repeatedly into
the wind to launch and recover aircraft. In this case they needed high speed to regain the
protection offered by the battle line. A carrier should be able to maintain air patrols that
would enable her to evade surface attack, and to send up fighters to counter a developing
air attack.
Gaming experience already showed the shape of a Pacific war in which the U.S. fleet
protected a necessarily massive convoy steaming west. Pratt wrote that

War Gaming and Carrier Aviation 87

 …we were more often forced to expose a vital bombing focus than was the
enemy and it was usually easier to locate a vulnerable bombing area in our
floating forces than it was in those of the enemy, for he had the power of a
wide initiative and was less circumscribed in his movements. Further due
to the great sea efforts we were forced to undertake, the quantity of targets
vulnerable to bombing attacks was much greater in our case than was the
case of the enemy. By proper strategical and tactical dispositions the ene-
my was enabled to dispose his forces so that a bombing attack delivered by
us need not usually fall upon a force massed, a contingency we could not
avoid. Also it was usually found that since the enemy was initially inferior
in gun power he was forced to develop his bombing tactics in order to
neutralize this gun preponderance as much as possible; that is, that the
bombing attacks assumed much of the character of the [torpedo] attacks
delivered by destroyers, namely, as an attempt to equalize the fighting
strength (measured in gun fire) of the two fleets, and to destroy that very
vulnerable portion of the fleet, the convoy.
Therefore, the enemy air objectives were usually first the U.S. carriers, then the bat-
tleships, and finally the convoy. Given limited U.S. air resources, the main U.S. concerns
were to protect the carriers’ facilities and U.S. fighting strength (which included the air-
craft needed to control gunfire); protection of the convoy was relegated to third place. If
these requirements could be met, the U.S. force could counterattack. If U.S. information
was good enough, the attack could be made first. However, the enemy was usually able to
take the initiative and to avoid great vulnerability to air attack.
All of these considerations led to concern to protect the fleet during the air battle
that usually preceded a gun battle, conserving those aircraft that would be needed during
the latter. Hence, the fleet tended to hold bombing power in reserve until the enemy
massed and made itself vulnerable, thus augmenting the gun superiority of the U.S. fleet.
Not all games went this way, but Pratt thought that a review of campaign games would
show that he was right.
Pratt concluded that his light mass-production carriers should be equipped main-
ly with fighters, the larger carriers accommodating mainly bombers. Observation and
patrol aircraft should be limited to non-carriers (cruisers and battleships). If possible,
destroyers and submarines should be equipped with observation planes (this included
the gunfire-spotting role).
A key point was that a carrier’s arresting gear could handle one airplane about every
two minutes, which was already faster than had previously been possible.21 If airplanes
landed one after the other, and they had a two-hour endurance, a carrier could accommo-
date and operate 60. However, in most cases an airplane’s wings had to be folded before
it could be struck below to clear the deck for the next landing. More time was needed to

88 Winning a Future War

In war gaming, the chances of torpedo bombers were problematic. In 1942, U.S. aerial torpedoes had to be
released near their targets at relatively low altitude and at low speed, making the bombers very vulnerable and
allowing the targets a good chance of evasion—as happened at Midway. The situation changed radically later
in the war, as torpedoes were modified so that they could be dropped at high speed from a much greater alti-
tude. This TBD-1 Devastator from Enterprise is shown dropping its torpedo on 26 October 1941. War gaming
showed the consequences of the reality that the U.S. fleet had to deal not only with the Japanese fleet, but also
with Japanese aircraft (and, probably, submarines and surface torpedo craft) based on the Mandated Islands.
The fleet therefore needed heavy bombers as well as anti-ship aircraft, and its torpedo bombers had a dual role.
Devastators performed well as bombers during the early carrier raids against the Mandated Islands, but they
were wiped out at Midway. (National Archives 80-G-19229)

maneuver an airplane after it had been stopped on deck. Some of that crucial time could be
saved if the airplane did not have to go below to be serviced and rearmed. That implied that
the flight deck had to be large enough for airplanes to be parked while the ship launched
and recovered aircraft. Gaming also showed the need, which Sherman pointed out, for a
new type of flight deck that could quickly be repaired after damage from light bombs.22
Summarizing carrier-design considerations for Admiral Pratt, Sherman looked to
British experience. The British had reconverted their 10,000-ton carrier Vindictive to a
cruiser. Sherman saw this as proof that 10,000 tons was too small. Sherman interpreted
the British choice to convert the 20,000-ton Courageous and Glorious to carriers as evi-
dence that this was about the right size for a future U.S. carrier.23 Apparently, he did not

War Gaming and Carrier Aviation 89

For a time in the 1930s, it seemed that big flying boats like this PBY-4 could be effective bombers, with per-
formance similar to that of land-based types. Many war games featured mass attacks by such aircraft, which
were attractive because they could move forward with the fleet, based on mobile tenders. The PBY was given
its bomber (B) designation when it was fitted with the Norden precision bombsight in recognition of its high
performance (including extraordinary endurance). By World War II, the PBY was no longer competitive with
land-based bombers, but higher-performance seaplanes were about to enter service. This PBY-4 is shown
dropping a heavy bomb, probably before the war. (National Archives 80-G-10550)

realize that these ships were the only viable British candidates for conversion under the
Washington Treaty. The British had no equivalents of the big but incomplete U.S. and
Japanese capital ships to convert. Sherman argued that a small number of larger ships
enjoyed important advantages. They would absorb fewer escorts, they would not require
so much communication to mount a large air attack, and, most importantly, they could
more easily coordinate their air efforts. He recognized the argument for dispersion, but
thought that three 23,000-tonners plus the two huge carriers offered enough.
For Pratt, the key argument favoring the 23,000-tonner was that it could be sup-
plemented by a large number of 10,000-ton carriers. However, by this time the Bureau
of Construction and Repair had decided that no such ship was practicable. In this light,
Pratt’s advocacy of numbers translated into a preference for the 13,800-tonner. It seems to
have been particularly significant that such a ship would not carry too many fewer aircraft

90 Winning a Future War

than a 23,000-tonner—aircraft capacity was proportional to the areas of the flight and
hangar decks, which would not decrease very rapidly with displacement.24 This seems to
have been the basis for a reversal of policy by the General Board. In September 1927, its
building program included five 13,800-tonners rather than three 23,000-tonners.
In its submission to the Secretary of the Navy, the General Board echoed the War
College’s 1925 reasoning: slightly greater aircraft capacity, but much greater air effective-
ness. Each carrier could launch the same number of aircraft at one time, so more carriers
meant putting many more aircraft into the air more rapidly. The larger number of smaller
carriers could also recover their aircraft much more quickly, and thus could sustain an
air effort for longer. Too, the loss of any one carrier would be far less devastating to the
overall air effort.25
The new carrier became Ranger (CV-4). She embodied Pratt’s thinking: Aircraft
numbers mattered more than anything else. The ship carried the maximum number of
aircraft at minimum cost. As an indication that aircraft numbers trumped other consid-
erations, the designers traded off power (hence speed) for more aircraft. Instead of the 33
knots of the converted battlecruisers, Ranger was rated at only 29.6 knots. Moreover, in
order to squeeze in more aircraft, Ranger was not armored, although she did have torpe-
do protection. The logic was simple: In games, and therefore probably in reality, carriers
would be able to evade surface attackers. Their main enemies would be fast airplanes and
invisible submarines. It was impossible to provide enough armor to protect the flight deck
against bombing; more airplanes were a better bargain. To provide maximum stowage
space, the constructors followed Sherman’s advice, and used nearly the whole of the up-
per deck as a hangar. Above it they erected a light superstructure topped by a flight deck.
Accounts of the design typically justified the lightness of this structure as a way to min-
imize top weight. However, but it also satisfied Sherman’s requirement that it be quick
and easy to repair. This type of structure was repeated in all U.S. carriers that fought in
World War II. Neither of the other carrier navies, the British or the Japanese, adopted
this type of construction. Neither seems to have contemplated the sort of drawn-out air-
sea war that made it desirable.
Just after the General Board made its recommendation, Captain F. J. Horne (later
to become the first CNO in World War II), acting director of the War Plans Division,
summarized the points that seemed most important to him. Although he did not cite war
gaming, his organization used war gaming as its laboratory.26 Horne’s memo reiterated
points Pratt had made, but in more striking form. It argued that a carrier, no matter
what her size, could advantageously operate only a limited number of aircraft. Although
a larger carrier could accommodate more aircraft, she could not operate them very effi-
ciently. Efficient operation demanded the maximum number of smaller carriers. They
could launch more aircraft in a given time, and they could also recover more of them. On
the other hand, it was easier to build a large fast carrier than a small fast carrier, because
machinery would take up less of the available space in the larger ship. War Plans pointed

War Gaming and Carrier Aviation 91

Seaplanes were attractive as part of the island-by-island strategy developed after the War College showed (by
gaming) that the earlier straight advance across the Pacific would not succeed. They could relocate quickly,
their tenders moving with or after the fleet. Once World War II began, the U.S. Navy supplemented its spe-
cially designed tenders with converted merchant ships. Here tender Currituck (AV-7) lies alongside the con-
verted merchant ship Tangier (AV-8) at Morotai in October 1944 to handle seaplanes supporting the Leyte
Gulf operation. The aircraft are small OS2U Kingfishers, but both ships had enough space aft to take on board
full-size seaplanes. (National Archives 80-G-1022364)

to the importance of the other air-capable ships in the fleet, those with catapults. The key
role of spotters in a battle-line engagement demanded that a carrier be included in or near
the battle line specifically to refuel spotters.
The episode is remarkable. Ranger was not laid down until 1931, but she was con-
ceived and designed well before Lexington and Saratoga had been completed.27 The rea-
soning leading to the design was based on war gaming—on the only way of visualizing
the entire campaign the U.S. Navy contemplated. It took war gaming to convince Admi-
ral Pratt that the United States could never have enough carriers to fight an Orange war,
hence that something new was needed. That might be converted liners or small mass-pro-
duced fleet carriers—or a better way of operating carriers themselves.
Above all, the campaign orientation offered by gaming showed that any Pacific war
had to involve huge numbers of aircraft. Wastage would be rapid. Ultimately that meant

92 Winning a Future War

that the U.S. Navy needed large reserves of aircraft and pilots. The two other carrier
navies seem not to have reached any such conclusion. In the case of the British, it might
be claimed that competition with the Royal Air Force for scarce resources made it im-
possible to build up reserves.28 This argument does not, however, apply to the Imperial
Japanese Navy, which did value its air arm—but never built up the reserves needed to
fight a protracted war.
After leaving the college, Admiral Pratt headed the U.S. delegation to the 1929 Lon-
don Naval Conference. Pratt was well aware that naval arms control was grossly unpop-
ular in the Navy. When he returned from London, he argued that he had extracted an
important concession. It was a clause that allowed the construction of a limited number
of cruisers with flight decks, which Pratt saw as a way of adding more fleet aircraft.29
Flight-deck cruisers became a fixture in war games and were retained in the building
program as a possible new type of ship until Pratt was relieved as CNO (see Chapter 5).
The other possibility, the converted liner, also became a fixture in U.S. war planning.
Such ships had figured in war gaming as early as 1923, but it seems likely that in their
earliest incarnation they were inspired by the British Argus, converted in 1916–18 from
an incomplete Italian liner. By 1929, the converted carrier (XCV) figured in lists of ships
to be taken up from the U.S. merchant fleet upon mobilization. The Bureau of Construc-
tion and Repair produced conversion drawings, and ships were earmarked. The concept
died because there were too few suitable fast ships in the U.S. merchant fleet and also be-
cause it seemed that conversion would take far too long. These ideas prefigured the escort
carriers of World War II, but those were far more austere and conversion was far quicker.

Reeves and Operating Practices

While all this was going on, another gaming-related development was affecting the
way in which U.S. carriers would operate. Captain Reeves, who had done so well in
the 1924 game, left the War College in the summer of 1925. His reactions to the two
games described here are not recorded, but what he soon did suggests that he digested
their lessons. At the War College, Reeves had shown no special interest in aviation; his
student thesis concerned battle-line gun tactics.30 He then went to Pensacola to take
the aviation observers course, which qualified him to command an aviation unit. Given
that credential, in September 1925 he was appointed Commander, Aircraft Squadrons,
Battle Force, a billet from which he could put his Naval War College experience into
practice. In later years, many U.S. naval officers with non-aviation backgrounds took
the course at Pensacola specifically because it opened important naval air commands
to them. That Reeves took the course before there was any such rule suggests that his
college experience—essentially gaming—had convinced him of the future importance of
naval aviation. When Reeves arrived in the battle force, his command amounted mainly
to the aircraft onboard the new Langley. His main assignment was to develop the tactics

War Gaming and Carrier Aviation 93

The inter-war Naval War College missed the potential for quick development of air and other bases on seized
islands. The wartime Navy included construction battalions (SeaBees) who could rapidly create substantial
airfields. As a consequence, the fleet was able to adopt high-performance land-based patrol bombers like these
Privateers (PB4Y-2s) from VP-23, shown over Miami in July 1949. The Navy retained seaplanes until 1965; it
valued their inherent mobility, gained by using tenders. (National Archives 80-G-440193)

they should employ. At the least, that meant air tactics. Reeves saw something more. He
would be responsible for the tactics of his force, meaning the way it was launched, the
way it was recovered, and the way it was used against an enemy. He had, in effect, a dead-
line. His new tactics had to be ready by the time really large numbers of aircraft reached
the fleet onboard the two huge carriers then being completed.
For Reeves, the lesson of the war games was obvious: Langley had to be able to take
aircraft onboard much more rapidly.31 That would also allow her to operate many more
aircraft. The two-minute landing interval enshrined in the war game rules was unaccept-
able. Reeves was anxious to get the most out of his new command, so he formulated a ques-
tionnaire for his pilots: the “Thousand and One Questions.” Among many other things,
he wanted to know how many aircraft his ship could support beyond the eight she was
assigned. Reeves’s war college experience showed that much larger numbers were needed.
Quite aside from the need for numbers to make carrier aircraft tactically meaning-
ful, Reeves needed them to devise the tactics and strategy the very large numbers planned

94 Winning a Future War

for the big carriers would need. Early in November 1925, he assembled his pilots and
crew in an auditorium at the naval air station at North Island, San Diego.32 He told them
that he planned to increase the air complement of their ship from eight to fourteen and
then to more. To do that, he had to increase the tempo of takeoffs and landings. Reeves’s
chief of staff later wrote that the pilots had learned that it was dangerous enough to land
on the small flight deck. They resented having a non-pilot tell them to take even worse
risks. Making takeoffs more rapid was not difficult, but cutting landing intervals was.
When Reeves took over, Langley typically recovered an airplane and then struck it
below before recovering another. The great innovation that made it possible to shrink
the landing interval was to push the airplane forward into a deck park instead of taking
it below into the hangar. Aircraft parked forward were protected by a wire barrier, a key
innovation.33 Arresting gear stopped a landing aircraft short of the barrier. It took far less
time to stop an airplane and roll it forward than to winch it down into the carrier’s prim-
itive hangar. Once all the aircraft were recovered, they could be rolled back to the after
end of the flight deck and set up for another flight. The new technique was not necessarily
popular with pilots; it required them to land into a space closed off by the barrier, rather
than onto an open runway-deck. It required a much greater degree of control of landings,
hence the institution of landing signal officers and, ultimately, a very fast tempo of land-
ings and takeoffs. All of these seem so natural now that it is easy to forget that during the
inter-war period they were unique to the U.S. Navy.
To an extent the pilots were right: The new technique was much more dangerous
than the past one. Anyone watching movies of U.S. carrier operations well into the jet
age will remember numerous clips of real accidents, many of them fatal. By way of com-
parison, the Royal Navy had no landing signal officers; pilots landed using their own
judgment, just as they did ashore. An account of inter-war British carrier aviation was
titled It’s Really Quite Safe. No account of U.S. inter-war carrier operations would have
had such a title. It took someone like Reeves to understand that the increased risk was
worth the gain in combat power. It took gaming to inspire him in the first place.
In December 1925, it took Langley 35 minutes to recover and stow ten airplanes
without a barrier—and that was considered an excellent performance. Langley had a
barrier by June 1926. In August 1926, recovery time per airplane was 90 seconds, and
Reeves was sure it could be improved much further.34 He recommended that Langley be
reclassified as a full combatant rather than an experimental carrier, and he expected to
double her air group to 28 aircraft. Reeves’s figures were presumably based primarily on
airplane endurance versus recovery time—at most a carrier could operate only as many
aircraft as she could recover before they ran out of fuel. The idea of using the flight deck
for servicing aircraft (as in Lieutenant Forrest Sherman’s December 1926 memo on the
new carrier) was a natural consequence of the new landing technique.
Reeves’s barrier made it possible to stow aircraft permanently on the flight deck,
not just on the hangar deck, because aircraft could land while the former remained in

War Gaming and Carrier Aviation 95

place. That must have been particularly important for the small Langley, which had a
tiny hangar and very slow elevators. For the future, it offered the U.S. Navy many more
aircraft per ship than foreign navies, which counted hangar capacity as ship capacity.35
Reeves’s barrier and the consequent innovation of the deck park were why three U.S.
carriers at Midway had onboard about as many aircraft as four Japanese carriers. In this
battle, the U.S. fleet was outnumbered in the number of carriers, but not in the number
of its aircraft.
Reeves’s innovation—born of game experience at the Naval War College—had a
subtler consequence. For most of the inter-war period, the Royal Navy had more carrier
tonnage than the U.S. Navy. However, the U.S. Navy of the inter-war period operated
many more carrier aircraft. That made for a much larger and livelier market in naval air-
craft. In Britain, the market for naval aircraft was tiny compared to the market for land-
based aircraft for the Royal Air Force. The Royal Navy could not demand the investment
that would have been needed to produce really high-performance carrier aircraft. The
much larger U.S. market encouraged manufacturers to innovate in order to meet de-
manding naval requirements. The effect of that larger market is apparent in the high per-
formance achieved by operational U.S. naval aircraft before and during World War II.
It was already well understood that very large numbers of aircraft were needed. The
lesson of the 1924 war game was that an air effort could not be sustained unless aircraft
could be recovered and serviced quickly. Aircraft could already be launched in quick suc-
cession. It was no surprise that a carrier could already launch a substantial strike. Doing
so more than once demanded quick landing, too. That was also important for a carrier’s
security. A carrier had to steam steadily into the wind to recover her aircraft. The longer
that took, the better the chance that she would be caught by a submarine. If the carrier
did operate with the battleships, the longer she steamed into the wind, if that took her
away from the battleships, the farther she got from their protection.
Reeves’s innovations were reflected in the 1929 rules. At least a quarter of a carrier’s
planes would normally ride the flight deck, and half could be accommodated there “in
such a manner that either flying off or flying on, but not both simultaneously, can be
carried on.” A fighter could be launched every ten rather than every 15 seconds, and a
torpedo bomber every 15 seconds rather than every two minutes, a vast difference. Cru-
cially, it was no longer necessary that the flight deck be cleared before an airplane could
be allowed to land. After they landed airplanes were moved forward. The 1933 rules cut
the day landing interval to half a minute.
The rules changed in 1936 to allow for launch and recovery with aircraft on deck. All
aircraft could be parked on deck, but in that case there was only enough space for aircraft
to take off, but not to land. With three quarters of the aircraft on deck, airplanes could
take off (if the aircraft were all aft) or land (with all of them forward), but not both simul-
taneously. However, with half the airplanes on deck, presumably all amidships, aircraft
could take off and land simultaneously. No more than half the aircraft could be in the

96 Winning a Future War

hangar at any one time. Such practices differed completely from those in other carrier na-
vies. Launch interval was given as ten seconds, and landing interval as 30 (intervals would
double in darkness). These figures corresponded to current U.S. practice.

Putting It Together—the Yorktown Class

Ranger having been authorized in the FY29 program, the question for 1930 was whether
she should be repeated. It was assumed that an additional carrier would soon be built.
The U.S. Navy still had 55,200 tons left in the carrier category, and the new London Na-
val Treaty did not change that. How should that be allocated? There was still very little
operational experience to inform any such decision; gaming still offered the best guid-
ance. Lexington and Saratoga became fully operational only in the fall of 1928, just before
Fleet Problem IX. That game, in which Saratoga successfully mounted a surprise attack
against the Panama Canal despite the numerical superiority of the defending fighters,
was the first demonstration of what a fast carrier operating more or less independently
could do.36 Ranger had sacrificed speed and protection for aircraft capacity. The spectac-
ular Saratoga operation suggested that this had been the wrong choice. If a carrier needed
the highest possible speed, comparable to that of the cruisers working with her, then cut-
ting size would also entail a considerable sacrifice in aircraft capacity. However, it could
be argued that the Saratoga spectacular was only a single exercise. The War College had
been running carrier battles for several years.
It was generally accepted, as during Admiral Pratt’s tenure at the War College, that
the U.S. Navy would need every seaborne airplane it could get. Both enemies in the
games (Japan and the United Kingdom) had large numbers of land-based aircraft unlim-
ited by treaty. Game after game showed not only how important aircraft were, but also
how rapidly they were shot down. Current War College President Rear Admiral Harris
Laning made this point, for example, when he provided the General Board with the War
College’s version of the composition of a balanced treaty navy in December 1931.
Ranger was not so much an ideal carrier as the best that could be done to maxi-
mize the number of aircraft at sea. Admiral Pratt brought a partial solution back from
the London Naval Conference. A flight-deck cruiser could accommodate a squadron of
aircraft; a full-size carrier could handle four squadrons. The eight flight-deck cruisers
allowable under the new treaty could be considered equivalent to two carriers. Games
played in 1931 showed that the flight-deck cruiser could be extremely valuable in scouting
and screening. In reporting this success, the War College admitted that it could not judge
the platform’s feasibility. Given the prospect of aircraft onboard flight-deck cruisers, it
could be argued that the Navy should build larger but less numerous carriers, perhaps
three somewhat larger ones instead of four more Rangers. As it happened, no flight-deck
cruisers were built, but by the time that project was dead, the vital decisions shaping the
new carriers had already been taken.37

War Gaming and Carrier Aviation 97

 By this time, no one liked the small Ranger, which had not yet been completed. Avail-
able alternatives in 1931 were four more Rangers; or two heavy general-purpose carriers
(20,000 or 20,700 tons each) and one smaller carrier; or three smaller general-purpose
carriers (18,400 tons each). Each of the alternatives offered about the same total number
of aircraft and the same number of flight decks.
This rethinking, in which gaming played an important role, turned out to be very
important. The U.S. Navy built two more fast heavy carriers, Yorktown (CV-5) and En-
terprise (CV-6). Not only were both very successful in the Pacific War, they were in effect
the prototypes of the Essex class fleet carriers that served so effectively in the Pacific
during World War II. The argument favoring two distinct types was that a ship some-
what larger than Ranger might yet prove attractive. CNO Admiral Pratt liked the com-
bination of two large and one smaller carrier, and the General Board agreed.
There was real interest in what was then called a battle-line carrier. Such a ship
would operate near the battleships, and she would keep their crucial spotters in the air
both by providing fighter protection and by servicing them. The War College’s fire-effect
rules showed how valuable long-range fire could be. It required the support of spotting
aircraft. In 1931, Commander of Aircraft, Battle Force considered a battle-line carrier as
the fleet’s most pressing need. He thought the new Ranger, which was being built, could
fill it. There was a real possibility that the next carrier would be a relief battle-line ship.
Naval War College President Rear Admiral Harris Laning used war-game experi-
ence to argue persuasively against a second battle-line carrier. He pointed out that both
full-scale experience and gaming showed that carriers would and should often operate
more or less independently, subject mainly to attack by enemy cruisers. They needed
high speed and a degree of hull protection, both of which would cost tonnage. When the
General Board revisited the carrier characteristics question in mid-1931, it demanded
both. Hence, the choice to build two heavy carriers was made. A third would be built
using the tonnage left over.
On 20 July 1931, as a preliminary to developing characteristics and setting carrier pol-
icy and a building program, the General Board asked CNO to conduct a study of the uses
of individual carriers and of U.S. carrier-based aircraft that were permitted under the Lon-
don Naval Treaty. Asking CNO meant asking his think tank, the Naval War College.38
The college study has not survived, but it was almost certainly the 5 August 1931 letter
cited and heavily quoted in a 30 July 1932 letter to the General Board, which has survived.39
The Naval War College observed at the outset that the more realistic the games be-
came, thanks to inputs based on fleet experience, the more important it was to increase
the number of fleet aircraft. The two great unanswered questions, then and through the
inter-war period, were the effect of antiaircraft gunfire and the effect of air-to-air combat.
Laning wrote that not until both questions had been resolved, “can we be sure of what we
ourselves will be able to accomplish with aircraft or what an enemy may accomplish with
his aircraft.”

98 Winning a Future War

 Even with those questions unanswered, Laning could be confident that “in all forms
of naval warfare, aircraft will exert a decisive influence, not necessarily by their direct
action and hitting power, but rather by the cumulative effect of initial advantage they
often gain for a fleet.” During the 1931–32 year in the War College’s problems and games,
“aircraft have continued to exert a decisive influence.” In Blue-Orange problems,
the information obtained by aircraft to both Blue and Orange, and the
effect of aircraft bombs and torpedoes, was invaluable to each fleet. The
games again emphasized the necessity of the United States’ possessing an
adequate number of carriers of various types to neutralize the advantages
of Japan’s carriers and shore based aircraft in well established bases near
the probable theater of war operations (i.e., in the Mandate Islands, is-
lands of the Far East, and through the Philippines). In these games, Or-
ange demonstrated how serious a blow she could strike against Blue in a
strategy of “attrition” using her aircraft, light and submarine forces, either
independently or united. And, against a Blue trans-Pacific slow-speed ex-
pedition, it appears that an overwhelming Orange air force in the vicinity
of the Marshall, Caroline, and Marianas Islands operating from carriers
and bases can exert a great influence against the Blue expedition and add
greatly to the damage already inflicted by previous submarine attacks. In
such a Blue-Orange Campaign, before the possibility of a fleet engage-
ment, the games show that the United States must have a great prepon-
derance of carriers and aircraft strength. It is vital that we have as many
carriers, ‘flying deck cruisers,’ and converted auxiliary merchant carriers
as possible, and all the aircraft practicable aboard each vessel, to use not
only for all purposes prior to a main battle line action, but to ensure that
we maintain air superiority during the approach and the action itself [em-
phasis in original].
Because it was so difficult to locate enemy air forces once they were aloft, it was gen-
erally impossible to gain “command of the air” by air combat alone. It could be attained
only by destroying or rendering inoperable the sources of enemy air operations—hitting
the enemy’s carriers as early as possible. Moreover, U.S. carrier flight decks had to be pro-
tected in all possible ways. Game rules and game experience emphasized how vulnerable
flight decks were to air attack, and how much could be lost when they were destroyed.
At this time, it was assumed that it would take many days for a ship’s company to effect
temporary flight deck repairs. All of the 1931–32 games emphasized the vulnerability of
carrier flight decks and flight-deck cruisers and the lack of suitable repair facilities in the
western Pacific.
The need for enough carriers and aircraft for both the attrition phase of the war and
the decisive battle affected the choice of the types and numbers of ships carrying aircraft,

War Gaming and Carrier Aviation 99

the types of aircraft, their armament, and the plans for the use of the material and per-
sonnel of this combined force. The treaty allowance of carriers was hardly sufficient; the
War College was looking for end-runs like the flight-deck cruiser and converted mer-
chant ships. The other point emphasized by the games was that everything had to exist
on the outbreak of war “with our most probable enemy, Japan.”40
The War College warned against trying to categorize carriers for particular duties.
This applied particularly to pressure being exerted at the time to build battle-line carriers
on the theory that the fleet already had two scouting carriers in the form of the two Lex-
ingtons.41 Unfortunately, battle-line carriers would probably be the main objective of any
enemy air attack prior to a battle line engagement. The 1931–32 games offered only a sin-
gle case of the use of battle-line carriers. In this case, the Red fleet had three 31-knot car-
riers and two slower ones. The latter were assigned to the Red battle line. They exchanged
their bombers for scouts and spotters. During the game, all five Red carriers were sur-
prised near their battle line by Blue scouting forces, which damaged all five Red flight
decks before the action began. “In general, no carrier which has remained within sight of
their own battle line has come through without destruction of her flight deck even though
not assigned to strictly battle line duties [emphasis in original].” All that the college could
suggest was that such a carrier should be small enough to hide, that she should have the
maximum possible flight-deck protection, maximum antiaircraft protection, stowage for
at least some of the battleship spotting aircraft, and possibly few or no main battery
(antiship) guns. The War College planned to assign carriers to the battle line during the
1932–33 games. Among the students were the former captains of the three existing U.S.
carriers, who were expected to provide insights into the use of carriers and aircraft.
Carrier antiaircraft guns offered little protection against bombers when the ship’s
aircraft were away. Battleship officers were certainly acutely aware of their need for air
services. During the wash-up after Fleet Problem XX (1939), they accused Rear Admiral
King, who was commanding the carriers on their side, of fighting a “private war” against
the enemy’s carriers. King replied that unless he dealt with the enemy carriers at the
outset, they would get no air services at all. Other Fleet Problems showed that a carrier’s
best hope of survival lay in evasive movements at high speed. A ship tied down to a slow
formation was likely to be hit. The War College noted that carriers had found high speed
very valuable in scouting, in evading submarines, and in dodging torpedoes and shifting
position during operations.
All carriers could and should operate all types of aircraft. All of their flight decks
were vital, particularly given their vulnerability. Each carrier or flight-deck cruiser should
carry the maximum number of aircraft practicable, compatible with maintaining all its
military characteristics.
Comments by the Bureau of Aeronautics dated July and August 1931 survive in the
General Board file on the carrier issue.42 They appear to show the influence of war game
experience. BuAer quoted former War College President (and then-current CNO) Pratt:

100 Winning a Future War

“[O]ne of the outstanding problems played each year is that to advance into a hostile zone
the fleet must carry with it an air force that will assure, beyond a doubt, command of the
air. That means not only superiority to enemy fleet aircraft, but also to his fleet and shore-
based aircraft combined. This is a large order, and the only way to secure it is by large
numbers of carriers and planes.” BuAer pointed out that given treaty limits, the only way
to add much to the fighting value of the fleet was by adding aircraft. Even large numbers
of additional cruisers (which were the main type the Navy was then considering building)
would add little to U.S. ability to destroy enemy capital ships or bases.
In November 1931, BuAer pointed out that its request for an estimate of how many
carriers and flight decks would be required had never been answered. Its own studies, as-
sisted by individual officers in the War Plans Division, indicated that at least 14 would be
needed. Moreover, the war games showed that aircraft attrition would also be enormous.
Initially it would be impossible to grow the carrier air force since, for the first nine months
of the war, the entire U.S. aeronautical industry would be making up attrition losses.43
BuAer cautioned that the six fleet carriers—all the United States could have, under the
treaties—and the eight flight-deck cruisers it envisaged “by no means represent our final
requirements in any war with a first-class power.” BuAer warned, moreover, that carriers
would inevitably be lost during the early stages of a war. Only merchant ship conversions
might make up some of the difference—but unfortunately there were few suitable U.S.
merchant ships, and it turned out that the conversions envisaged would have taken far too
long. Perhaps in hopes of keeping the carrier program popular in a battleship-oriented
naval establishment, BuAer accepted the idea of the battle-line carrier. Admiral Moffett
must have been aware of the danger of tying a carrier to battleships.
In 1931, the General Board took the War College’s advice. It could not do anything
about Ranger, so it characterized her as a battle-line carrier and argued that she could ben-
efit from the cover offered by the battleships. The Naval War College had already pointed
out that such protection was delusionary. For the future, the War College bought two
fast protected carriers. The third, whose tonnage was limited by the Washington trea-
ty, would be a slightly inferior general-purpose carrier. Unlike Ranger, this ship—Wasp
(CV-7)—was ultimately used in the Pacific.44 The General Board finessed the question
of whether she was worth building by saying that she was a point in the carrier spectrum
that could be compared with the larger Yorktowns. Relative value was indicated by the
order in which the ships were built: The two bigger and more useful carriers came first.
The Naval War College emphasized the need to make the flight deck survivable. In
games, carriers were most often put out of action by bombing, not by attacks by surface
ships or submarines. In its July 1931 memo, the college proposed that various flight-deck
facilities be duplicated to make them more difficult to destroy. This idea did not appear
in the formal characteristics reported by the General Board, but it was embodied in the
ships as built. The previous month, BuAer had already called for arresting gear at both
ends of the flight deck, so that aircraft could land on at either end. The hangar side was

War Gaming and Carrier Aviation 101

to have an open section forward from which aircraft could be launched using a fixed
athwartships catapult. This idea had been raised by Admiral Reeves as early as December
1927.45 He saw it as a way of maintaining carrier capability despite flight-deck damage.
Reeves also preferred the open hangar (which was introduced in Ranger) because it made
for maximum aircraft capacity, offering the maximum proportion of hangar space to the
upper deck. He wrote at the time that “it is evident that the fleet which takes to the scene
of battle and is able to launch into the air the greatest number of aircraft should win the
air battle, other things being equal.” Airplanes in an open hangar could be gassed and
warmed up before they were brought up to the flight deck, hence could be launched far
more quickly. All of these features were included in the new carriers and in the follow-on
Essex class.46
In describing improvements embodied in the new carrier, the Bureau of Construc-
tion and Repair responded to a specific request from Air Squadrons Battle Force for on-
board facilities to repair flight-deck damage.47 The last item in the C&R letter describing
features of the new design stated that “it is assumed that the flight deck damage...refers to
local damage to planking and plating occasioned primarily by bombing. It is believed that
the only practicable method of making quick repairs to such minor damage on the part of
the Forces Afloat will be by spiking steel plating over the holes in the decks. Such plates
can be taken from the regular stock carried by the vessel. A repair party station is being
provided just under the flight deck accessible to the galleries on the Ranger and CV 5
[Yorktown].” This was a rather casual reference to something very important—something
the War College had seen in its games. It harked back to the Ranger design and to Forrest
Sherman’s call for a light flight deck that could be repaired quickly after it was bombed.
The Ranger flight deck certainly need not have been repeated. Other navies did not
follow suit. British and Japanese carriers had flight decks integral with their hulls, their
hangars closed in at the sides to form structural supports. Ranger and her light flight
deck and open hangar had been a distinct departure from the steel flight decks and ful-
ly-enclosed hangar of the Lexingtons. Prodded to think about how the flight decks atop
open-sided hangars could be repaired, C&R noticed something important: The light
wood over steel structures could be fixed surprisingly quickly.48 The rather casual way
the issue was addressed suggests that it had not been raised before; it was certainly not
included in the formal characteristics of the next carrier, Wasp. By far, the bulk of cor-
respondence concerning the new designs concentrated on whether (and by how much)
the flight deck could be extended to increase aircraft capacity—another important issue
raised by war games.
Once Yorktown was about to enter service in 1937, the war game rules were rewritten
to allow for relatively quick repairs by a ship’s force. The time required to repair flight-
deck damage was still very conservative, but now that the ship’s own crew could make the
repairs, the ship could remain in action despite damage. The reality of such repairs was
demonstrated again and again during the Pacific War. That was particularly important

102 Winning a Future War

early in the war, when the U.S. Navy had few carriers available in the Pacific. The reason
Enterprise was able to fight in nearly every Pacific battle was that her crew could patch her
immediately after she was hit. Without the ability to make quick repairs and thus keep
its few carriers in action, the U.S. Navy would have been far less effective in 1942–43,
before the mass of new carriers appeared.
This type of construction was an alternative to flight-deck armor. The choice was
raised when U.S. naval officers inspected the British armored-deck carrier Illustrious as
she was being repaired at Norfolk Naval Shipyard in the fall of 1941.49 They were very
impressed, and said so at a General Board hearing. The response was that the U.S. Navy
had opted for more than twice as many aircraft and for the ability to make quick repairs
after battle damage. The officers were told that U.S. practice had been designed for a
Pacific campaign rather than the Mediterranean or European operations for which the
British had conceived their carriers. The use of the word “campaign” reflects an under-
standing of the whole of a Pacific war gained by gaming. Illustrious could certainly resist
some bombs, but she was being repaired because her deck had been penetrated by much
heavier ordnance dropped by dive bombers. As in the war game rules, she needed a first-
class shipyard to repair that kind of damage. Enterprise and her sisters were able to patch
their own flight decks and keep fighting. Even so, they were just short of fully exploiting
the light flight deck. That was left to the Essex class, unlimited by treaty and hence con-
siderably larger.50
The preliminary design of the new 20,000-ton carrier was formally submitted to
the Secretary of the Navy on 28 December 1931.51 In submitting initial characteristics
in May 1931, the General Board had listed roles in order of importance support of the
battle line; support of battle forces (e.g., contact scouting, both tactical and strategic, and
attacks on the enemy air force); support of fleet operations and movements, including
security of train and convoy; and operations with task forces or independently. It is a
measure of the War College’s success that the first priority changed dramatically in favor
of independent operations of various kinds. Despite some experience with the big Lex-
ingtons, the General Board stated in a covering letter to the October 1931 version of the
characteristics that “the present broad fields of usefulness of carriers have not yet been
revealed in a practical way.” Gaming provided a lot of the needed experience.
The perceived urgency of carrier construction shows in the recommendation to build
two such ships in the FY33 budget and the third in FY34: The General Board clearly
understood that carriers and their aircraft were essential.52

Aftermath
After 1934, the Naval War College never again cited war-game evidence in any advice it
offered. The terms of the 1930 London Naval Treaty closed off further carrier construc-
tion, leaving the U.S. Navy with two options to provide the aircraft it needed. One was

War Gaming and Carrier Aviation 103

merchant ship conversion, but, as noted, there were too few suitable ships and conversion,
as then envisaged, would take far too long. The other was to seek a source of aircraft not
limited by treaty: seaplanes. Particularly after the war plan strategy changed to a step-by-
step advance, the atolls of the Pacific offered numerous potential seaplane bases. In war
games, patrol planes—seaplanes not limited by treaty—often added to naval air power
not only for scouting but also for attack. Beginning in 1934, the U.S. Navy started work
on seaplane tenders that could expand the fleet’s air reach and its numbers. In a 1937
reorganization, the fleet’s seaplanes were moved from the Base Force at Pearl Harbor to
a new Patrol Wings organization as part of the Scouting Force. The implication was that
they would move forward with the fleet.
The evidence of interest in patrol planes as attackers, outside the war games, is the
decision to designate the Consolidated Catalina as a patrol bomber (PBY) rather than
in the pure patrol series of the past. Previous flying boats had been capable of carrying
heavy bombs, but the PBY added the Navy-developed Norden precision bombsight. The
attack role of the seaplanes is obscured because, by 1941, they could no longer compete
effectively with modern fighters (although PBYs did turn out to be effective night bomb-
ers, at least for a time). However, the conceptual effort was not wasted; the wartime Navy
effectively deployed large numbers of land-based patrol bombers in the Pacific. These
aircraft never figured in war games, because until 1942 the U.S. Navy was prohibited
from operating them under an agreement between CNO Pratt and Army Chief of Staff
Douglas MacArthur.
In 1936, the new London Naval Treaty abandoned limits on total tonnage in all its
categories. Everything said and written in 1931–32 would have pointed to a large U.S.
carrier program in parallel with the substantial battleship program that then began. In-
stead, only one carrier was authorized in FY38 (Hornet, effectively a repeat Yorktown),
another (Essex) not following until FY41. The design of the latter ship was greatly en-
larged once World War II began and all remaining treaty limits were abandoned. There
seems to have been a sense that seven or eight fleet carriers were quite enough. The kind
of campaign reasoning which demanded many more was not brought up. The absence
of carriers in the U.S. building programs of this time can be taken as evidence that the
war-gaming point of view was no longer of interest, at least outside the OPNAV War
Plans Division. The emphasis on battleships suggests that attention was concentrated
on the (hopefully) decisive battle against the Orange battle line rather than on the entire
campaign against Orange. That is somewhat ironic, given that the “through ticket to Ma-
nila” had been abandoned.
On the other hand, war games had also shown that carriers were terribly vulnerable
to air attack. It was widely accepted that, once enemy carriers were nearby, there was no
security at all until they had been neutralized. The same applied to an enemy’s view of U.S.
carriers. It was entirely possible that all carriers on both sides would be put out of action
early in a war. The greater combat survivability bought by the new type of flight deck seems

104 Winning a Future War

not to have been widely understood. Carriers did not really gain survivability until the
development of radar-based fighter direction, which was only dimly in prospect in 1940.
Any interpretation of what happened is muddied by the character of the legislation
under which new ships were authorized. In 1934, Congress passed the Vinson-Trammell
Act, under which sufficient tonnage was authorized to build up to a “modern Treaty
navy.” In practice, that meant replacement of over-age ships. When the total tonnage
limits were abandoned, it meant that over-age ships no longer had to be discarded. Since
there were no over-age carriers in the U.S. fleet, the question of new carrier construction
did not arise. All of the battleships were far older, hence were automatically candidates
for replacement. In 1938, a second Vinson-Trammell Act added new tonnage, and the
question of new carriers certainly did arise. The new act simply enlarged each tonnage
category by a set percentage. Thus, the authorization continued to reflect the balance
between different types of ships reflected in the Washington and London treaties. In the
case of carriers, it added 40,000 tons, enough for two ships. The use of a set percentage
avoided controversy; even in 1938, naval rearmament was a very contentious issue. The
FY38 carrier, Hornet, was a special authorization outside the tonnage increases in the
1938 act.53 Thus, it was a direct reaction to the opportunity opened by the 1936 treaty.
No carriers were included in the FY39 and FY40 programs. When the Secretary of
the Navy asked the General Board in July 1938 for relative priorities for design work for
the FY39 and FY40 programs, a new aircraft carrier came last out of five design projects.
The first priority was the new 45,000-ton battleship, which became the Iowa class.54 Even
in this case, the situation is muddied, since the new carrier might be a repeat Yorktown,
in which case no design work would be needed, and no characteristics for an entirely new
carrier design had yet been laid down. In 1938, the General Board expected to include
two 20,000-ton carriers in the FY41 program. Formal General Board policy set out in
October 1934 was not to reduce the number of carriers (seven) “in view of the need for
aircraft with the Fleet and of the large wastage factor which, due to the vulnerability of
their decks, may be expected of carriers in an active campaign.” In a 1940 memo, the
General Board observed that the combination of expected wastage and the “increasing
relative performance of land based planes to carrier based planes” had resulted in a “prob-
lematical appraisal of carrier value. Before this question gets in the open, and due to the
present popularity of aviation in Congress, it is believed wise to include one, if not two,
carriers in the 1941 program. Unless a war occurs in which the role or value of carriers
is definitely determined, it is believed that before carriers are scrapped, improvements
in carrier technique and landing and take-off facilities, will enable them to handle max-
imum performance planes of the fighting and pursuit type, which would warrant their
retention in the fleet”55
Presumably, this language reflected fleet skepticism as to the potential of carriers,
which the General Board warned would be taken poorly by the public and by President
Roosevelt, who was a strong advocate of air power. On the other hand, the General Board

War Gaming and Carrier Aviation 105

should have been aware that better carrier aircraft were coming. At this time, BuAer was
testing the F4U Corsair, which it thought was the fastest fighter in the world—and which
was intended for carrier operation.
At about the same time, the bureaus proposed a ten-year (FY39–48) program in-
cluding six carriers (two each in FY41, 45, and 46) and 16 battleships.56 The carriers
were the two new ones Congress had provided plus replacements to be laid down when
existing carriers became over-age. Not even BuAer seems to have advocated a massive
carrier-building program. None of the General Board policy papers of this time refers
back to the drastic need for aircraft in an Orange war, as war gaming had demonstrated.
Since 1931, Pacific War strategy had changed, and it is possible that the new seaplane
bombers were considered an effective counter to Japanese aircraft based in the Mandates.
The theory would have been that these bombers would provide effective air cover to the
fleet by destroying Japanese land-based air power at source. The fleet would need much
less air cover as it steamed west, at least until it encountered Japanese carriers and their
aircraft. In that case, it might be reasonable to relate needed U.S. carrier strength to Jap-
anese carrier strength rather than to the much larger total Japanese air threat including
land-based aircraft.
On the other hand, the step-by-step strategy demanded the seizure of island after
island in the face of Japanese resistance, including aircraft. Carriers were needed for these
seizures; analyses conducted in 1938 made it clear that long-range aircraft from other
islands could not effectively support a landing. Every such seizure, which tied the carri-
ers to a particular area, left them vulnerable to submarine attack. A rational peacetime
program should have allowed for such losses. There was some interest in providing the
Marines with floatplane bombers that could be launched by a seaplane tender, but even
if such aircraft were effective, they could not operate until U.S. forces had seized control
of sheltered areas of water. It is not at all clear to what extent the OPNAV War Plans
Division estimated how many carriers would be needed for this phase of a war.
Even if the big seaplanes were envisaged as alternatives to carrier attack aircraft, that
turned out to be a poor choice. The gap between seaplane and land-plane performance
opened rather than diminished as war approached, although there was some hope for
improvement in seaplane performance as engines became more powerful. The U.S. Navy
found itself operating versions of Army Air Force land-based bombers from the islands
it conquered. Even their ability to project air power forward with the fleet was limited. It
was not at all clear that such aircraft, based on the initial islands seized, could neutralize
Japanese air bases deeper in the Mandates. In the games, neutralization was achieved by
periodically gassing bypassed airfields after they had been struck. In reality, gas warfare
was a political issue, and it was never permitted. Among other things, that meant that
airfields could never be permanently neutralized. None of this would have been obvious
before the outbreak of war, but it does seem useful to point out that the carrier role
changed substantially.

106 Winning a Future War

 The General Board record suggests no wide-ranging reappraisal of overall U.S. re-
quirements in view of a Pacific (or other) strategy. The question of when carriers would
largely replace battleships was raised, but it seems to have been assumed that would lie
in the distant future. Nothing in the General Board record suggests that the new kind
of repairable flight deck had made an appreciable difference in thinking. There was no
allowance for possible war losses because there was no estimate of how many carriers the
strategy would require.
Changing U.S. interest in carriers was illustrated by the FY41 program, which ul-
timately added 11 of them. When this program was first considered in June 1939, it
included a single new carrier, CV-9, which was conceived as a slightly modified Yorktown
displacing at most 20,400 tons.57 In October, war having broken out, OPNAV was con-
sidering emergency expansion of the U.S. Navy on the basis of a 25 percent increase in
authorized tonnage. This time, it seems that tonnage of carriers, cruisers, destroyers, and
submarines could be swapped. The program showed four new carriers. A fifth new carri-
er (Hornet) was already under construction, to make an ultimate total of 11. In addition,
of 14 new light cruisers and ten new heavy cruisers, at least four were to have flight decks
(as flight-deck cruisers). Additional seaplane tenders (and seaplanes) were included. This
program was based on a “Two-Ocean Navy” program that had been developed inside
OPNAV. It was based largely on an inventory of shipyard capacity. That approach was
reasonable: Gaming had shown that the U.S. Navy would need vastly more ships. All
that could be done was to make maximum use of existing resources.58 As of late May
1940 the projected FY41 program was headed by two battleships and four carriers. An-
other eight battleships had recently been authorized, plus a single carrier.
The building program was totally revamped after France fell. Until that point, the
United States might plan to assist the allies in the Atlantic, but the bulk of the fleet could
still be concentrated in the Pacific. The U.S. Navy could fight a one-ocean war while
participating, if need be, in the other ocean. After France fell, it was no longer so certain
that Britain would survive. The United States might find itself fighting full-scale naval
wars in both oceans, in both cases against opponents with powerful battle fleets. That is
not obvious now, because the worst case scenario in the Atlantic never came to pass. The
British never surrendered, and the Germans never gained control of the French fleet.
The U.S. reaction to the fall of France included the massive new Two-Ocean Navy
Act calling for a 70 percent increase in authorized tonnage. A BuShips estimate dated 27
July 1940 showed seven more battleships and seven more carriers, to make a total of 11
FY41 carriers. The carrier figure seems to have been based on nothing more than a ratio
between the original plan for a 25 percent increase (4 carriers) and the new 70 percent in-
crease. The carrier program was by no means universally supported. Priorities swung to
carriers only in 1942, when the Navy was forced to make choices due to a steel shortage.
New battleships and large battleship-like cruisers were cancelled, but the carriers were
not. The way those decisions were taken is outside the scope of this study. What is strik-

War Gaming and Carrier Aviation 107

ing is the shift from, in effect, “we can’t have enough carriers” on the basis of simulation
by gaming to an emphasis on more or less meaningless ratios of numbers or tonnages of
carriers divorced altogether from any sort of campaign analysis.
Gaming shaped the carriers that proved so effective in the Pacific, because it forced
their designers to think in terms of the entire campaign they would fight. Gaming forced
the Navy to think about the campaign that had to be fought prior to the expected deci-
sive battle. That it would have been easy to disregard the campaign prior to the battle is
suggested by the analysis—or lack of analysis—of the late 1930s, when the United States
was rearming. The force ratios involved made sense, if at all, only in the context of a single
battle. The reality was that the Pacific War resembled the pre-battle campaign. In this
campaign, the force ratio was not carrier to carrier, but U.S. carriers to full Japanese naval
air forces.

108 Winning a Future War

5. T W C  C

C ruisers, which may not seem very important in retrospect, were the largest surface
warships that could be built during the “battleship-building holiday” imposed by
the Washington Treaty. They were also the class in which the United States was weak-
est, Congress having rejected new cruiser construction after approving three small scout
cruisers before World War I (the big 1916 building program included ten more scouts).
Cruisers became a major concern for the inter-war War College and a major application
of its gaming technique.
For inter-war navies, cruisers had three distinct roles. One was as scouts, acting ei-
ther independently or in the van of the fleet. A second was as the fleet’s defense against
enemy destroyers attacking with torpedoes. Despite their name, destroyers were not
considered an effective defense against enemy destroyer attacks because they were poor
gun platforms. In effect, they were ocean-going torpedo attack craft. A third was to deal
with enemy commerce raiders. At this time, the law of war prohibited “sink-on-sight”
attacks; merchant ships had to be stopped and examined, and only then sunk or seized.
The safety of their crews had to be ensured, which often meant that they had to be taken
onboard the raider. These legal conditions generally required that a raider be a surface
ship with enough speed and firepower to stop a fast merchant ship. An enemy would
deploy cruisers or, more likely, armed merchant ships. The first two roles were the most
important to the inter-war U.S. Navy, but for the Royal Navy trade protection was at
least as important.
Gaming demonstrated what everyone in the U.S. Navy already knew: The Navy had
far too few cruisers. Worse, it was impossible to place an upper limit on what it needed,
just as it was impossible to provide enough airplanes to support an Orange war. The
inter-war arms-control agreements functioned as justification for particular numbers of
cruisers under the 5:5:3 ratio (Britain: United States: Japan) applied to total battleship

The War College and Cruisers 109

and carrier tonnage (not numbers) at Washington in 1921. From 1922 on, the General
Board was responsible for a formal U.S. naval policy. It included the important political
requirement that the Navy be “second to none,” meaning parity with the Royal Navy. The
General Board also called for the United States to build the maximum possible num-
ber of the largest permitted cruisers armed with the most powerful possible guns. That
meant 10,000-ton ships armed with 8-inch guns. The numbers that the General Board
proposed were based on the 5:5:3 ratio, applied to the total tonnage of British and Japa-
nese cruisers. It seems to have been generally assumed that further arms-control negoti-
ation would extend the Washington Naval Treaty to cruisers and even to destroyers and
submarines.
Gaming offered only limited insights. It could certainly demonstrate the conse-
quences of the weakness of the U.S. cruiser force, particularly in a fleet-on-fleet battle.
Campaign games could show that the United States had far too few cruisers overall.
However, it was impossible to say how many would be enough. The games of this era
did not, moreover, provide insight into the range of cruiser types the U.S. Navy needed,
or to the proportions between types. In effect the ten “scouts” of the 1916 program pro-
vided the U.S. Navy with all the smaller fleet cruisers it was likely to get, the General
Board pressing only for construction of the much larger ships armed with 8-inch guns.
Here, fleet cruisers meant ships capable of dealing with (or supporting) destroyer tor-
pedo attacks.
The big cruisers favored by the General Board were well-adapted to function as
scouts, operating either independently or in the van of the fleet. To some extent they
offered a capability the United States had been denied when the Washington Treaty
had cancelled the U.S. program to build battlecruisers. However, big cruisers with heavy
guns were poorly equipped for other cruiser functions, particularly beating off destroyer
attacks on the battleships, since the powerful, 8-inch guns fired too slowly to be sure of
hitting rapidly maneuvering destroyers.1
U.S. Navy goals were not necessarily the goals of the U.S. government. The Wash-
ington Treaty came to be seen as the first step toward all-embracing naval arms control,
which was seen by successive U.S. administrations as a way both of avoiding war and of
reducing the crushing cost of armaments. U.S. President Calvin Coolidge convened a
follow-on conference at Geneva in 1927 specifically to extend the Washington Treaty to
smaller warships, specifically cruisers. The major navies, particularly the British and the
Japanese, were already building substantial numbers of large cruisers, and there was a
general perception that this new naval arms race had to be stopped. The General Board’s
hopes that Congress would provide enough such ships to make up for the gross U.S.
deficiency had been disappointed. Congress had authorized eight ships in 1924, far short
of what the General Board’s policy required. Any attempt at negotiation was complicated
by British insistence that the Royal Navy needed more small cruisers than the U.S. Navy
in order to protect the much more massive British seaborne trade. For their part, U.S.

110 Winning a Future War

The War College did not originate the big cruisers the U.S. and other navies built in the 1920s. Nor did it
evaluate them, based on their flimsy protection and their heavy armament. It certainly did appreciate the need
for fleet scouts. This is Houston (CA-30), a first-generation heavy (“treaty”) cruiser armed with 8-inch guns,
but very lightly protected to achieve high speed. In the background is the far better protected New Orleans
(CA-32). Both were photographed off San Pedro on 18 April 1935. (National Archives 80-CF-21337-1)

naval officers badly wanted sufficient numbers of modern cruisers before they were will-
ing to accept any sort of limitation. At Geneva, as at the Washington Conference, naval
officers were among the delegates. U.S. naval officers were widely credited with breaking
up the conference, infuriating President Coolidge.2
The Washington Treaty required that a new conference be called in 1930, when the
suspension of battleship construction was due to end. The conference met in London.
The sort of naval engagement that had crippled the 1927 conference was avoided by bar-
ring naval officers as delegates.3 They did attend as naval advisors, however. Rear Admi-
ral William V. Pratt, recently president of the War College and now commander of the
U.S. Fleet, was chief advisor.4 He was widely credited with resolving problems and thus
making the 1930 conference successful. In the past, he had been the Navy’s expert on
arms control. His key role in the 1930 conference made him extremely unpopular within
the service.5 Pratt had served as Admiral Sims’s chief of staff when Sims had command-
ed U.S. naval forces in Europe during World War I, and the British remembered him

The War College and Cruisers 111

warmly. His role at the conference was much appreciated by President Hoover, who was
well aware of the failure of his predecessor’s 1927 Geneva conference.
In the wake of the London Naval Conference, President Hoover made Pratt Chief of
Naval Operations. To the extent that this appointment was seen as a reward for service
at the conference, it cannot have helped Pratt’s popularity within the Navy. Presidential
favor gave Pratt more power than he might otherwise have had. To a substantial extent,
he could decide what the U.S. Navy would build. As CNO, Pratt removed the president
of the War College from his former ex officio position on the General Board. The War
College president retained his position within OPNAV, as a direct advisor to the CNO.
The War College itself appears to have had no direct impact on the U.S. negotiating
position.6 That is no surprise; the U.S. Navy did not change its preference for the most
powerful cruisers. The demand for parity was a political method of justifying construc-
tion of the maximum possible number of such ships. Conversely, the Coolidge and Hoover
administrations wanted to shrink the British cruiser fleet so that they could limit U.S.
construction of expensive cruisers within a demand for parity. Both U.S. administrations
doubted that the U.S. public would stand for an agreement leaving the British with a sig-
nificant superiority in cruiser numbers, even though it was obvious that the British needed
more cruisers to protect their huge merchant fleet (the threat to which, it was assumed,
was surface raiders, not submarines). The British had built their own cruisers armed with
8-inch guns, and they found such massive ships unaffordable in the desired numbers. Since
they associated the heavy gun with large (expensive) cruisers, their single main objective at
the conference was to preclude further construction of cruisers armed with 8-inch guns.
For trade protection, a much smaller gun, of 6-inch caliber, was large enough, because no
converted surface raider could be armed with anything more powerful.7
Probably the key factor in success was a very friendly 1929 summit conference be-
tween President Hoover and British Prime Minister Ramsay McDonald, during which
both agreed that differences would be resolved. British naval officers saw this as a sur-
render to the Americans. McDonald agreed to clauses, such as a dramatic reduction in
destroyer tonnage, that the Admiralty specifically rejected. He was anxious to reach
agreement both for idealistic reasons and because his country was already in severe eco-
nomic trouble, even before the onset of the Great Depression. McDonald found himself
accepting the American pressure to force down the size of the British cruiser force to
the point where the U.S. Navy could match it. The Admiralty admitted that it could not
quickly replace the numerous ageing cruisers left over from their World War I building
program.8 McDonald told his critics that numbers could be built back up after the treaty
expired. This time, too, the U.S. Navy was more favorably inclined toward cruiser limita-
tion because Congress had authorized 15 more cruisers in February 1929, of which five
were to be ordered each year from FY29 on.9
Given the new Congressional authorization, the United States might build 23 cruis-
ers armed with 8-inch guns. The British had stopped building such ships after comple-

112 Winning a Future War

tion of the fifteenth unit. Initially the U.S. delegation offered either parity with the Brit-
ish (15 each) or 18 8-inch gun cruisers, in each case with a tonnage allowance for 6-inch
cruisers.10 As ratified, the treaty allowed the U.S. Navy to build its last three 8-inch gun
cruisers, but their construction was to be delayed. As an incentive to the United States
not to build those ships, it could substitute 15,166 tons of 6-inch cruisers for each 8-inch
cruiser not built. Parties could also transfer up to 10 percent of their allowed destroyer
tonnage into cruisers armed with 6-inch guns. All of this left exactly the sort of questions
war gaming was designed to investigate. Were the last three 8-inch gun ships worth more
than the larger tonnage that could go into 6-inch gun cruisers? What sort of 6-inch gun
cruisers should the U.S. Navy build?
The high status of aircraft in the U.S. Navy was evidenced by the appointment of
Rear Admiral Moffett as a member of the naval staff at London. Because the initial U.S.
proposals were drafted by the State Department without naval input, they did not reflect
his thinking. Thus, the only carrier proposal initially offered by the United States was to
abandon the existing provision that anything under 10,000 tons would not count in the
carrier tonnage total. The Japanese had taken advantage of this clause by building the
8,000-ton carrier Ryujo. She had not yet been completed, so it was not yet evident that
she was too small and would have to be rebuilt. Once that had been done, she no longer
displaced less than 10,000 tons, although the Japanese chose not to admit as much. The
British agreed to ban small carriers despite their potential value in protecting British
trade. They were unlikely to build many such ships, given their economic problems, and
if anyone else did, such carriers would be a considerable threat to British seaborne trade.
The British took carriers far less seriously than did the U.S. Navy. The Washington
Conference had allowed guns as large as 8-inch caliber onboard carriers. The British
suddenly decided that a country could build air-capable cruisers instead of true carriers.
Such ships could be somewhat larger than the 10,000-ton cruisers allowable under the
treaty, and they might overwhelm existing cruisers. To prevent a cruiser from being built
in the guise of a carrier, the maximum caliber of carrier guns was reduced to 6.1 inches.
The British apparently did not imagine that anyone would build a large cruiser armed
with such weapons.
At the upper end of the scale, the British wanted to reduce the 27,000-ton limit agreed
at Washington to 25,000 tons. On that basis, the 135,000-ton total agreed at Washing-
ton, equivalent to five 27,000-tonners, could be cut to 125,000 tons (five 25,000-tonners).
This was anathema to the U.S. Navy. The two large carriers had taken up 66,000 tons,
and the remainder was just enough for five 13,800-ton Rangers. It seemed most unlikely
that the two huge carriers would be reclassified as experimental, hence instantly replace-
able. The British proposal translated into one fewer U.S. carrier. The U.S. Navy needed
more, not fewer carrier-based aircraft.
The carrier provision mattered in the context of cruisers because Moffett was vitally
interested in an end-run: the flight-deck cruiser. He wanted to use some of the allowed

The War College and Cruisers 113

It proved possible to improve cruiser protection (survivability) dramatically. Here, the second-generation
heavy cruiser Minneapolis (CA-36) fires in battle practice, 29 March 1939. It seems to have taken War College
analysis to demonstrate clearly how superior such ships were to contemporary foreign (particularly British)
heavy cruisers. Without the analysis of the type developed for war gaming, it was certainly clear that ships
like this were better than their predecessors; but only the analysis could show how much better they were.
(National Archives 80-CF-21343-2)

cruiser tonnage to build, in effect, fractional aircraft carriers. His chief, Admiral Pratt,
certainly understood: As Naval War College president, he had pointed to the desperate
need for aircraft in an Orange war—as demonstrated by gaming—and to the impossibil-
ity of providing enough within the terms of the earlier Washington Treaty.11 Moffett and
Pratt managed to insert the clauses they wanted in the final treaty.12 Up to a quarter of
allowable cruisers could be fitted with flight decks, and they would not count toward the
allowed total of carrier tonnage. That left another question that gaming could help solve:
would it be worthwhile to build flight-deck cruisers?
As noted above, the flight-deck cruiser had been conceived some years before in
connection with the 1924 Congressional authorization of the first eight 10,000-ton
cruisers. At about the same time, air power advocates were arguing that aircraft could
and should replace virtually all major warships. Cruisers became a battleground. The
Washington Treaty imposed no limit on the total tonnage of surface ships displacing
10,000 tons or less. On 31 March 1925, Secretary of the Navy Curtis Wilbur asked

114 Winning a Future War

the General Board whether, “in view of the propaganda now going on in favor of air-
craft,” some of the 10,000-ton cruisers should be completed either as small carriers or
as combination cruiser-carriers.13 The General Board rejected completion of any of the
ships as carriers. Not only was the board reluctant to sacrifice cruisers, but preliminary
studies by the the Bureau of Construction and Repair showed that 10,000 tons was
too small.14 The bureau similarly rejected the idea of a combination cruiser-carrier. The
Secretary agreed, and no such ships were built—for the moment. However, the idea of
a combination ship survived.
The proposal occasioned a formal comment from the War College, based on expe-
rience in tactical games.15 The college pointed out that the two functions of light cruiser
and carrier were fundamentally incompatible. To function effectively, a cruiser had to
place herself within gun and torpedo range of the enemy. A carrier had to avoid exactly
those weapons because she was inherently vulnerable. She would probably be unable to
function as a carrier once she had been involved in a gun action, because it would not
take much gun damage to disable her flight deck and to wipe out any aircraft on it. Her
aircraft could not affect the outcome, because planes would be unable to fly off or on, or
to survive on deck, during a gun battle. On the other hand, tactical exercises at Newport
certainly had shown that a larger number of smaller carriers offered advantages.
The situation changed once total cruiser tonnage was limited. Pratt returned from
London pointing to the treaty provision allowing construction of flight-deck cruisers as a
major opportunity for the United States; there was enough tonnage for up to eight such
ships. Each could accommodate a squadron of aircraft, so together the cruisers might
equate to two aircraft carriers beyond the limit assigned by the Washington Treaty. As
noted in the previous chapter, this possibility affected the choice of the next carrier de-
sign, justifying the choice of larger (hence less numerous) ships.

Evaluating Alternatives
In the aftermath of the London Naval Treaty, War College President Rear Admiral Har-
ris Laning wrote the General Board in September 1930 stating that for the past several
months the War College had been working on a table showing what it considered the
most effective fleet the United States could have when the treaty expired in 1936—what
might later be called a “balanced treaty navy.”16 The letter is interesting because it led to
an exercise showing how the war-game style of fleet comparison (and indeed of thinking
about problems) translated into proposals for construction. Laning was pleased that the
General Board’s program corresponded roughly with what he had in mind. Figures for
carriers (at that time four more Rangers) and submarines were fixed. The question was
how to use available cruiser tonnage. Laning wanted to use all the available cruiser ton-
nage and also to build destroyer leaders (a type not yet in the U.S. fleet) for five of the sev-
en squadrons of destroyers allowed under the treaty. He was particularly concerned that

The War College and Cruisers 115

the General Board did not intend to use all the allowable 6-inch gun cruiser tonnage. The
London and Washington treaties were due to expire in 1936. A replacement treaty would
be negotiated at a 1935 conference. It might well further reduce allowed U.S. cruiser
strength. Anyone trying to make such a cut would find it easier if the United States failed
to build all the cruisers it already could. Whatever was built or under construction in
1935 might define the new limit. Laning recommended keeping decisions as to details of
6-inch cruiser tonnage as fluid as possible, yet at the same time obtaining authorization
(and laying down all permissible ships) before 1936. The General Board clearly regretted
the loss of 8-inch cruisers at the London Naval Conference. It was not about to accept the
treaty provision offering nearly 46,000 tons of 6-inch gun cruisers in exchange for the last
three 8-inch gun cruisers. Without this tonnage, the treaty allowed the United States a
total of 143,500 tons of 6-inch cruisers. The ten Omaha-class “scout cruisers” that had
been ordered in 1916 accounted for 75,000 tons. However, because they had been laid
down from 1918 on, they would be eligible for replacement beginning in 1938.
The value of the powerful carriers had recently been demonstrated. The dramatic
success of the carrier Saratoga during the 1928 Fleet Problem showed that they could
usefully operate independently of the battle line. War games emphasized that point:
They should stay as far as possible from the battleships to avoid being caught by enemy
aircraft. A fast carrier had only one type of surface ship to fear: a fast cruiser. Conversely,
only cruisers were fast enough and powerful enough to be effective carrier escorts. To
protect a carrier, at the least they had to match enemy firepower. In 1930, the U.S. Navy
contemplated ultimately operating seven fast carriers. If each required a four-ship cruiser
division as escorts, the needs of carriers alone would account for 28 8-inch gun cruisers.17
Laning’s letter to the General Board was a new departure. In the past, since the War
College President had been an ex officio member of the General Board, most communi-
cation was informal. Admiral Pratt had just cut that connection. Probably as a result of
Laning’s September letter, General Board Chairman Rear Admiral Mark Bristol wrote
Laning on 14 October asking that the college work out tactical and strategic problems
using proposed types of ships, either those recommended by the department or those
suggested by the college itself—the latter being preferable.18 Bristol emphasized that he
wanted the War College to work out what was needed on the basis of its gaming; If games
demonstrated the worth of a proposed type, “the Department and the General Board
would then have a very sound foundation for preparing plans and developing such a type
of vessel and in the same way avoid working on types which the College has demonstrated
are not the best for the Navy.... [T]he College should suggest new types and improved
types by utilizing them in your game. Specifically such new types have been suggested
recently.” The latter presumably referred to the flight-deck cruiser.
As for the choice between 6-inch and 8-inch guns, the U.S. Navy saw the former
very differently from the British. To the British, 6-inch guns would match the heaviest
guns any merchant ship converted for commerce raiding could mount, making a cruiser

116 Winning a Future War

The Naval War College did not have much impact on the agreements reached during the 1930 London Na-
val Conference, but it certainly did affect what the U.S. Navy decided to build afterward. In convening the
conference, the British were concerned mainly to kill off the heavy cruisers, which they considered ruinously
expensive. They assumed that any navy building cruisers armed only with 6-inch rather than 8-inch guns
would adopt ships roughly equivalent to their own 6-inch gun cruisers, displacing about 7,000 rather than
10,000 tons (cost was about proportional to tonnage). The principal U.S. naval expert at the conference was
Admiral Pratt, who returned home having accepted the British view. It says much about the strength of the
analysis the War College was able to bring to bear that he reversed his position and accepted that the cruiser
the U.S. fleet needed was much larger. The result was the Brooklyn class. Honolulu (CL-48) is shown here in a 9
February 1939 photo. (NHHC NH 53562)

armed with such weapons an effective trade-protection ship. To the U.S. Navy, on which
the 6-inch gun had been imposed, its main virtue was that it could fire much more rapidly
than the 8-inch gun arming its large cruisers. That made the 6-inch gun most valuable
as a means of protecting the battle line from enemy destroyer torpedo attacks during,
before, or after a fleet action.
At this time, the General Board liked the idea of small cruisers, so it wanted five
ships displacing about 7,000 or 8,000 tons. Laning wanted the remaining 30,000 tons
devoted to three large 6-inch gun cruisers of “special type,” to be laid down in 1932, 1935,
and 1936. The first might well be a CLV, the other two of an unspecified type to be cho-
sen in a year or two, when U.S. needs would be clearer. The CLV should be tested before

The War College and Cruisers 117

more were built. Laning would use gaming to test the CLV, a well-protected 6-inch gun
cruiser, and other possibilities. He was careful to point out that the college was making
no claim as to the feasibility of the CLV. Assuming it was feasible, its tactical value could
be tested in a war game.
Laning’s September 1930 letter seems to have had some effect. The General Board
dropped its demand that no light (6-inch gun) cruisers be built until all the approved
heavy cruisers (8-inch guns) were completed. In December, it approved design require-
ments for a 10,000-ton cruiser armed with 6-inch guns, and CNO Pratt approved them
in January. Pratt also favored a small (6,000-ton) cruiser for distant scouting and for fleet
work. He may well have been influenced by the British, who had assumed during the
run-up to the treaty conference that 6-inch gun cruisers would be scaled down, hence
much less expensive, versions of the unaffordable 8-inch gun cruiser. It is not clear why
they thought so. British internal documents reveal no idea whatever of the logic of the
other maritime powers. It seems to have been assumed that they, like the British, would
be driven by a desire to reduce the cost of ships. The British were content to limit guns but
not cruiser size. They were soon unpleasantly surprised that both the Americans and the
Japanese built rather large 6-inch gun cruisers, the Brooklyn and Mogami classes, ruining
what the British clearly imagined had been a diplomatic victory.
In contrast to British reasoning, which was based on self-interest and a tacit assump-
tion that all governments shared the same rationale, the Naval War College examined
the logic likely to drive the other major powers, particularly Japan. That was the method
the college had long been teaching to students who had to prepare an “Estimate of the
Situation” before deciding what to do. The first stage was not what they wanted out of a
negotiation or problem. The first stage was to guess what the enemy wanted, and what the
enemy’s logic might be. Gaming might show that the guesses embodied in the “Estimate
of the Situation” were flawed, but simply writing an estimate offered great advantages.
The initial War College report (6 December 1930) focused on the proposed flight-
deck cruiser. It concluded that the ship should have the same speed and endurance as
existing heavy cruisers (32.5 knots, 10,000 nautical miles at 15 knots) and should be
protected at least against 6-inch (preferably against 8-inch) fire. On 18 December, CNO
asked for a formal comparison between an 8-inch gun cruiser and a flight-deck cruiser
armed with 6-inch guns. The War College submitted results on 2 January. They includ-
ed a comparison in which both types of ship tried to beat off a destroyer attack on the
battle line during a general engagement. A current 8-inch gun cruiser (Northampton)
was unable to sink all four attacking destroyers before one of them managed to fire her
torpedoes at the battle line. The CLV, with or without aircraft, did far better, because
its nine 9-inch guns fired much faster than the heavy cruiser’s nine 8-inchers. Either
gun would kill an approaching destroyer with a single hit. In a simulated night action,
in which the enemy destroyers were illuminated by U.S. destroyers at 8,000 yards, the
CLV sank all four attacking destroyers in three minutes, the 8-inch gun cruiser in three

118 Winning a Future War

to four minutes. In offensive screening, the CLV was far more effective. However, for
screening purposes, the CLV needed protection against 8-inch fire, even at the cost of
some aircraft.
At this point, the War College examined the much wider issue of what sort of cruis-
ers the U.S. Navy needed. CNO and the General Board envisaged building a variety of
cruiser types, as foreign navies did. Gaming showed that this would be folly: No U.S.
commander could be sure that his ships would encounter comparable foreign ships.
What a foreign navy used depended on what it was trying to do. “For this reason, the
development of special type cruisers based solely on the opposition they are to encounter
on a particular duty is not practicable [emphasis in the original].” The type of cruiser to be
used in a particular role should depend entirely on what the U.S. commander was trying
to accomplish. “We should design our cruisers primarily to accomplish a purpose for us,
but at the same time giving to each the maximum power for offense and defense.” Num-
bers were valuable, but in wartime, weak cruisers could quickly be created by converting
merchant ships. However, under the existing treaty there was no way to quickly increase
fighting strength. Ammunition allowances should be as large as possible, and gaming
indicated that radius of action should be at least 10,000 nautical miles at 15 knots. All of
this seems to have been an indirect way of killing Pratt’s 6000-ton cruiser.
On 11 December 1930, Admiral Bristol of the General Board formally requested
a War College study of the characteristics of future fleet cruisers, based on a list of du-
ties.19 The eight 8-inch gun cruisers the Navy had already laid down (the Pensacola and
Northampton classes) were lightly protected, as were their foreign equivalents. However,
the Bureau of Construction and Repair had managed to provide far better protection in
a new design, designated CL-38 at this time.20
C&R pressed for a version of the latest 8-inch gun cruiser (CL-37–38), armed with
15 6-inch guns (three quadruple and one triple mount) instead of nine 8-inch guns. It
argued that when employed with the fleet, she would be “worth half a dozen lightly ar-
mored, lightly gunned 6-inch gun cruisers such as are the prevailing style today in foreign
navies as well as our own.” Added displacement provided vital advantages not always
included in military characteristics: seaworthiness, sustained sea speed, and habitability.
Compared to a 6,000-ton ship, a 10,000-ton ship had every advantage at sea, in anything
but nearly smooth water, in maintaining her speed. Moreover, a larger ship could carry
more guns per ton. That would make each gun less expensive, since ship cost was roughly
proportional to tonnage.
If the three allowed 8-inch gun cruisers were built, 73,000 tons were left. It could be
distributed in various ways, assuming that all of it went on the same type of ship. How
could these ships be compared? The designers thought that the largest type (10,000 tons)
could be protected to the point that she could get close enough to an 8-inch gun ship for
her 6-inch guns to become effective. The War College’s gaming techniques were centered
on ideas of relative fighting strength. Its initial approach was to graph the relative fighting

The War College and Cruisers 119

strengths of the different possibilities at various ranges, and to compare them with the
existing U.S. heavy and light cruisers (Pensacola and Omaha) and with the latest pro-
posed 8-inch gun cruiser (CL-38, which became San Francisco). Fighting strength took
rate of fire into account, so for the 6-inch cruiser the War College tried both the rate of
fire realized in the Omahas and an increased rate of fire that it was already using in other
cruiser comparisons. All of this entailed an enormous amount of work, and as of 6 Janu-
ary 1931 the results had not yet gone to the General Board. There had not been enough
time to digest them.
First, the War College compared U.S. cruisers to U.S. cruisers. The 6-inch gun
cruisers performed surprisingly well inside 10,000 yards, where they could penetrate the
armor of an 8-inch gun cruiser. However, there was an important caveat: At a target an-
gle of 45 rather than 90 degrees (i.e., when the ships were not on parallel courses), which
might not be unusual in a cruiser-on-cruiser engagement, the new heavy cruiser (CL-38)
could penetrate the 6-inch gun cruiser without being penetrated. It was pointless to con-
tinue to pursue all the possibilities, so the analysts reduced the study to two 6-inch gun
schemes (one armored against 8-inch guns with twelve guns, and one with protection
against 6-inch fire and nine guns), plus the existing Omaha, and Pensacola. A comparison
with Orange cruisers showed how much better the new CL-38 was, compared to the
lightly protected (but more heavily armed) Pensacola. This seems to have been a distinct
surprise; the Naval War College officer wrote that “Lord, if I had only had this a year and
a half ago [mid-1929], when I sat at a meeting of the General Board, considering our new
cruisers, and I could find no sympathy there for my prefixed opinion that we needed more
protection on our cruisers.” A similar comparison with the Red (British) cruiser Sussex
again showed how effective better-protected ships (CL-38 and the 12-gun ship) were at
“nice cruiser ranges” of about 13 to 20,000 yards.21
So much for single ships; what of the numbers that could be built, given the
N-square law (combat power is proportional to the square of the number)? The War
College analysts compared three ships with 8-inch protection with four with 6-inch
protection (a total of 28,000 tons in each case). They were surprised to discover that bet-
ter protection overcame the N-square law at cruiser ranges of 11,000 to 17,000 yards.
When the Pensacola was replaced by the slightly better-protected Northampton, the val-
ue of superior protection was particularly obvious outside 18,000 yards. The increased
rate of 6-inch fire would devastate a poorly protected ship. When the two 6-inch ships
were put up against an 8-inch cruiser with good protection (CL-38) the result was very
different; at a 45-degree target angle, it would have been even more pronounced in favor
of CL-38. As a further experiment, the two cruisers and an Omaha were pitted against a
division of six destroyers. Omaha looked rather weak. It was obvious that numbers were
attractive, but fewer well-protected ships were a much better bargain. That effectively
killed the small fleet cruiser. The graphs were included in a formal reply to the General
Board dated 10 January 1931.

120 Winning a Future War

The War College was so effective in setting U.S. cruiser policy that cruiser design might be considered a mea-
sure of its success. In that sense, the outcome of the London Naval Conference of 1935 (which led to the 1936
London Naval Treaty) might be considered a measure of the downfall of the War College as a think tank bas-
ing its advice on gaming and game-related analysis. This time, the British succeeded in reducing the maximum
size of new cruisers to 8,000 tons, even though the War College had advised U.S. delegates to avoid exactly
such an outcome. Their argument was based on what war games tested: how well the fleet could carry out the
national strategy. Cruiser size does not really figure in accounts of the negotiation; the U.S. delegation was far
more concerned to beat down a British attempt to reduce the size of battleships and to develop a means of
controlling Japanese construction (the Japanese removed themselves from the treaty system). The War College
was not, it seems, even consulted as the design for a new small cruiser was developed. This is the outcome:
USS San Juan (CL-54) of the Atlanta class, photographed on 3 June 1942. (National Archives 19-N-31525)

All of this calculation, which employed methods developed specifically for war gam-
ing, had real consequences. In 1931 the U.S. Navy had eight completed 8-inch cruisers
and was building another seven. The War College analysis revealed just how weak those
ships were. The ideal solution was the new CL-38 design, but it was difficult to change the
designs of ships already under construction. Even so, that was done. Ships to be built in
Navy yards were reordered to the new design. The two ordered from private yards could
not be totally redesigned, but they were given considerably better protection.
The War College’s initial formal answer to the cruiser query was a study by the De-
partment of Intelligence, delivered in February 1931.22 It reflected gaming techniques,

The War College and Cruisers 121

both in its attention to what other countries were likely to do and in its methods of com-
parison. As the War College saw it, the question was how the United States might best
distribute the available cruiser and destroyer tonnages based on what the other two major na-
val powers already had and were likely to build. To get to this point, the War College looked
at the situation from the points of view of Orange and Red (primarily Orange) in terms of
the likely tasks of their fleets, as derived from war-gaming experience. To ask the question
about the policy a likely enemy would adopt before setting U.S. policy was the core of the
decision-making process the War College taught. This approach may seem obvious, but it
is striking that when the British looked at the same problem they made no such analysis.
Orange’s vital tasks would be to maintain command of the western Pacific and east-
ern Asiatic seas and to maintain the security of Orange sea communications—to prevent
attacks on Orange shipping. Orange would fight a maximum-intensity defensive war in-
tended to keep Blue from strengthening its position in the Far East. This was not a bad
description of the Pacific War as actually fought. Orange would have to find and attack
the Blue fleet before it got to the Philippines; if it arrived, its mobility would have to be
impaired. Blue lines of communication would have to be attacked. The Japanese might
occupy enemy territory. In that case their navy would have to escort shipping and pro-
tect it during and after a landing. The Orange navy would also have to protect Japanese
shipping to insure sufficient imports to carry on the war. Even though Japan could receive
massive amounts of material via protected sea routes from China, Manchuria, Asiatic
Russia, Korea, and Sakhalin, in a war lasting perhaps more than two years she would also
need southern trade routes for commodities such as oil and cotton. Japan was noticeably
weak in tanker tonnage, so these ships would particularly need protection.
Even though Japan’s war would be defensive, it would open with considerable of-
fensive action, for which the 12 Japanese heavy cruisers already built or building would
be suitable. The eight largest Japanese heavy cruisers were superior to all but the latest
U.S. heavy cruisers. The other four (Kako class) were inferior to U.S. heavy cruisers, but
compared well with the only U.S. cruisers normally attached to the U.S. fleet, the Oma-
ha class, due to their greater radius of action. At shorter gun ranges, the Omahas were
probably superior due to the higher rate of fire of their 6-inch guns. Combined with the
three Orange battlecruisers, the Orange heavy cruisers would make a powerful advanced
force. Six of the smaller Orange cruisers were already employed as destroyer leaders. At
this time, the Japanese were building what they described as 1,315-ton destroyers (they
were actually considerably larger), and within the available destroyer tonnage they could
fill out the same six-squadron destroyer organization.23 With the larger cruisers assigned
to distant operations and seven smaller ones needed as flotilla leaders, the governing con-
sideration for Orange would be numbers both for the fleet and for convoy duty. The sea
lanes involved were short, and nearby bases and fortifications would make it unnecessary
for the Japanese to build large cruisers. Orange had few ships suitable for conversion to
cruisers. The Japanese quota of heavy cruisers had already been exhausted.24

122 Winning a Future War

 By 31 December 1936, based on ships that would become over-age, Japan could lay
down 50,955 tons of new cruisers. The War College had a variety of alternative cruiser
designs from which the Japanese (like the Americans) might choose.25 The treaty allowed
Japan to build flight-deck cruisers, but that seemed unlikely given the numerous Japa-
nese-held islands suited to seaplane operation.26
All of this logic implied that the Japanese would build 6-inch gun cruisers roughly
equivalent to the (nominally) 7,100 ton-Kakos, which could mount as many as ten 6-inch
guns. That would immediately give them five ships. The Japanese later announced that
they were building four 8,500-ton Mogami-class cruisers, armed with 15 6-inch guns
each. Thus, the War College prediction was not too far off the mark, except that the
Japanese claimed that they could squeeze much more into the announced tonnage than
any Western navy could. No one seems to have suspected the reality, namely that the
Mogamis proved unacceptably flimsy and had to be rebuilt with much greater tonnage
(which was never announced by the Japanese).
Given the assumed Orange program, what should Blue do? Ultimately Blue would
want to gain control of vital Orange trade routes—the end-game blockade—while pro-
tecting its own. Possible Blue courses of action would be the “through ticket to Manila,”
or a fleet advance to an improvised base elsewhere in the Philippines, or seizure of an
intermediate base short of the Philippines. The fleet would have to undertake very exten-
sive scouting, and it would have to protect its lines of supply once it was established in the
Western Pacific. In any case, Blue had to be ready at all times to face a maximum attack
by Orange. War games showed that Orange generally tried to wear down Blue’s battle
line superiority (as provided by the Washington Treaty) using submarines, aircraft, and
light torpedo craft (possibly including small torpedo boats not limited by treaty) based in
the Mandates. They would have to be supported by Orange scouts. Blue would therefore
need scouts to locate and defeat the Orange raiding/scouting parties. Enough Blue scouts
would have to survive to ensure victory in a fleet action.
In wartime, when tonnage limits would no longer apply, Blue could build destroyers
in three months, as U.S. shipyards had shown they could during World War I. However,
cruisers would take at least a year. The Naval War College therefore argued that the
United States should take advantage of the 10 percent of allowed destroyer tonnage that
could be used instead for cruisers; every effort should be made to provide more cruisers.
What type should they be?
During the U.S. fleet’s advance to the west, it would need scouts. At the least, they
would cover the large areas immediately ahead of the moving fleet. They would also
screen the fleet so that the enemy’s scouts would be unable to determine the U.S. forma-
tion or even its precise course. With the advent of aircraft, the area that had to be scouted
and screened expanded enormously. To the college, the 10,000-ton flight-deck cruiser
seemed an ideal scouting platform, because its aircraft could cover such large areas.27
New technology offered it a considerable punch. It could operate dive bombers, which

The War College and Cruisers 123

already seemed to be extremely effective against any ship short of a battleship. As long as
aircraft from a flight-deck cruiser found the enemy first, the ship was more than a match
for any enemy cruiser. If the aircraft did not find the enemy first, the flight-deck cruiser
could be well protected. The college argued that flight-deck cruisers would provide air-
craft during the early phases of the war. They would free the big carriers to concentrate
on gaining and maintaining an initial advantage in the expected climactic fleet battle.
The Naval War College view was that U.S. 8-inch gun cruisers and flight-deck cruis-
ers would be used mainly for offensive screening, scouting, convoying, and commerce
raiding. Therefore, 6-inch cruisers would be used mainly on the flanks of the fleet to beat
off torpedo attacks by enemy destroyers and also to back up torpedo attacks by U.S. de-
stroyer on the Japanese battleships. Their high rate and volume of fire would suit them to
this role. Screening and convoying would be secondary roles. The existing and projected
8-inch gun cruisers already offered a margin over their Japanese counterparts.28
No matter which type of cruiser the U.S. Navy chose to build, the available tonnage
allowed for as many as eight flight-deck cruisers, which the War College’s analysis sug-
gested would be superior to 8-inch gun cruisers. Thus, if all 18 heavy cruisers were built,
it would be possible to justify all eight flight-deck cruisers, each superior to any 8-inch
cruiser for most cruiser duties—a remarkable achievement, as it would give the Navy 26
cruisers instead of the 24 the General Board had struggled for a year earlier.
The Naval War College report denigrated 8-inch cruisers as an artifact of the Wash-
ington Naval Treaty rather than a response to any particular tactical need; no one read-
ing it would imagine that the General Board had insisted on such ships in 1920, before
the treaty had been imagined. According to the report, these ships did not meet any
particular tactical need, and were wanted only to offset similar ships in other navies. It
seemed unlikely that more would be built in Europe.29The report argued further that
heavy cruisers, particularly the faster French and Italian ones, had turned out to be poor
sea boats. Too much had been sacrificed to cram 8-inch guns and protection into 10,000
tons. If such ships were given good protection, they would be slower than cruisers armed
with 6-inch guns. In any case, the United States could not get enough 8-inch cruisers.30
The Naval War College had already demonstrated that small 6-inch gun cruisers
were not worth building. It pointed out that by abandoning construction of 8-inch gun
cruisers at 15 (rather than the allowable 18) and by transferring tonnage from destroyers,
“we not only get better ships but more of them.” College studies showed that although it
was impossible to protect a cruiser against 8-inch fire at short ranges, a heavily gunned
6-inch gun cruiser had superior fighting strength to an 8-inch gun cruiser at ranges un-
der 8,000 yards owing to its higher volume of fire.31 Heavy cruisers might well encounter
6-inch gun cruisers at such ranges owing to low visibility or to tactical disposition. The
6-inch cruiser should have as many guns as possible, so that it could exploit their higher
rate of fire, twice that of an 8-inch gun. Some of those guns might be disabled before
the 6-inch cruiser got within effective range. To make certain it was superior once it was

124 Winning a Future War

within range, the 6-inch gun ship should have at least three times the initial volume of
fire as the 8-inch ship at the outset. That meant 12 6-inch guns against the usual eight
8-inchers in a British cruiser, for example, or 15 against Japanese cruisers with ten 8-inch
guns. “As the life of any cruiser will be short after penetration starts, volume of gun fire
will count heavily in determining which ship remains afloat.” Firepower was particularly
attractive in ships that might attack enemy trade, as the enemy would be forced to devote
more of his limited cruiser force to convoy protection. The War College pointed out that
protection could increase the life of a ship under fire and thus preserve her speed and
increase the total amount of fire she could deliver. This comment reflected the fire-effect
rules and hence the sense of how a ship would behave in combat. The trouble was that in
a fight with an 8-inch gun cruiser, the 6-inch gun cruiser had to close the range quickly
enough to retain firepower. A really fast 6-inch gun cruiser could not be protected at all,
so by the time she got within range she might have no firepower left.
The War College view, then, based on gaming experience, was that the U.S. Navy
should go as far as possible to build flight-deck cruisers and the largest 6-inch gun cruis-
ers it could get by exploiting all treaty allowances. The latter was the 9,600-ton Scheme
2. Available tonnage would amount to eight flight-deck cruisers and six Scheme 2s. The
resulting cruiser force would be considerably more powerful than Orange’s—as long as
all U.S. heavy cruisers were devoted to cruiser functions. In 1931, four heavy cruisers
were used as flagships. One was flagship of the Asiatic Fleet. A second was flagship of the
U.S. Fleet (the Navy was experimenting with the idea that the fleet commander should
not be onboard a battleship engaging enemy battleships). Another was flagship of the
Scouting Force; a fourth was flagship of the Scouting Force cruisers. All of them could
be replaced by large light cruisers, but that would leave only two such ships for fleet func-
tions such as beating off Japanese destroyer attacks and supporting U.S. destroyers. The
War College therefore punted: Instead of those two large light cruisers, it might pay to
invest the tonnage instead in six 6,500-ton cruisers, a type CNO had suggested for use as
flagships and for other fleet duty.
In marked contrast to the tactical evaluation of Orange cruisers in the context of a
Pacific War, the evaluation of the Red (British) cruiser force was political: The United
States had to maintain parity in tonnage and in fighting strength.32 It appeared that the
British force as of 31 December 1936 (when the London Treaty expired) could be pre-
dicted much more accurately than that of the Japanese.33 Alternative U.S. cruiser fleets
could be rated as to total numbers, volume of fire, and types.
The War College staff spent its Christmas 1930–31 break working on the evaluation
of the flight-deck cruiser. Laning himself wrote that he had long believed in the possibil-
ities of this type of ship, particularly given the obvious power of dive-bombing—and the
impossibility of stopping a dive-bombing attack. The problem, which the War College
was unable to answer, was how to trade off the air power exercised by a flight-deck cruiser
against gun power. That mattered, because a flight-deck cruiser was relatively expensive

The War College and Cruisers 125

in terms of guns per ton. However, even building eight rather than four flight-deck cruis-
ers would carry only a limited cost in total volume of fire.
Ultimately, then, the War College advocated building no further 8-inch cruisers and
instead building a mix of flight-deck and conventional 6-inch cruisers. In its view, a flight-
deck cruiser was likely to be more than a match for an 8 inch–gun cruiser. However, the
college did not advocate simply building the largest possible 6 inch–gun cruisers. Instead
it envisaged a trade-off between size and numbers. The college proposed a program of
four flight-deck cruisers, two 10,000-ton 6-inch cruisers (convertible to flight-deck cruis-
ers), and 18 8,250 ton-cruisers. This program would exploit the bonus tonnage available
if 6 inch– rather than 8 inch–gun cruisers were built as well as the transfer of 18,300 tons
of destroyer tonnage. Given the severe shortage of cruisers the study showed, the United
States should immediately build as many as possible of a new unlimited class of sloops
(up to 2,000 tons with 6-inch guns) to release real cruisers from convoy duty.34 It should
also prepare for rapid conversion of merchant ships to cruisers and carriers, with special
emphasis on carriers, because the flight-deck cruiser aircraft would need landing decks.
They would launch their aircraft before opening fire, but once they were firing they would
be unable to recover their aircraft.35
A War College report submitted on 2 April 1931 emphasized the role of torpe-
do-firing destroyers in any future fleet battle. The Japanese would use their destroyers
to overcome the treaty-mandated numerical superiority of the U.S. battleships. Con-
versely, in a fight against the British, who had treaty-approved parity in numbers, U.S.
destroyers would attempt to tip the balance. The War College added that “unless such
successful [destroyer torpedo] attacks can be prevented, they may prove the deciding
factor in an engagement.”36 Torpedoes mattered because each torpedo hit would inflict
more damage than several shell hits. Torpedoes could be delivered by destroyers or by
torpedo bombers or by submarines. The college emphasized destroyer torpedo attacks
because the other two means of delivery were unlikely to be as effective in a major fleet
engagement. Torpedo-bomber performance was poor (torpedoes were very heavy, and
aero engines not yet very powerful), to the point that in 1931 the U.S. Navy was serious-
ly considering abandoning them. Submarines were not fast enough to keep up with the
battleships in combat.
Further consideration of exactly what the battle line mission required further af-
fected the desired characteristics of any new cruisers. If, as the War College argued, an
enemy destroyer attack might decide the battle, any cruiser assigned to the battle line
needed enough fire power to break it up. Since any of the 6-inch cruisers being considered
could break up destroyer attacks, the deciding factor should be the ability to stop enemy
cruisers backing up the attacking destroyers. That strongly favored a 12-gun cruiser pro-
tected against 8-inch fire, since the enemy might use heavy cruisers to back his destroy-
ers. Comparison techniques developed for gaming showed that the 12-gun cruiser had
greater fighting strength than the British Sussex class below 19,000 yards or the Japanese

126 Winning a Future War

Nachi below 13,500 yards.37 It offered considerably more fighting power, ton for ton, than
any Japanese or British cruiser that might be built for some years to come, barring radical
changes in either country’s building program.
In June 1931, CNO Pratt reversed himself and recommended the heavily protected
12-gun cruiser. The Secretary of the Navy approved this recommendation several days
later. It is difficult to say just how much the War College and its game-based analysis
were responsible for the change, although Pratt was certainly sensitive to War College
arguments.38 Pratt certainly did not accept the War College argument that the last three
8-inch cruisers should not be built as such, or that tonnage be transferred from the de-
stroyer to the cruiser categories. Given the onset of the Great Depression, any such ideas
may have been dropped because they could not be realized. However, the War College’s
arguments favoring large over small 6-inch cruisers—arguments based on war-game ex-
perience and on a war-gaming style of analysis—did carry the day once the U.S. Navy
began building such ships. The U.S. Navy built seven large cruisers each armed with 15
6-inch guns—the Brooklyn class—plus two ships of a somewhat modified type (the St.
Louis class).
The War College was asked again in 1932 to analyze U.S. cruiser requirements. That
March, C&R claimed that there was “a strong and increasing sentiment” for a change in
the characteristics of future 8-inch gun cruisers. Pressure for further cuts in spending
and for further arms control would reduce capital ships to the point where large cruis-
ers effectively replaced them.39 C&R suggested that cutting cruiser speed by five knots
and replacing the 5-inch antiaircraft gun with the new (developmental) 1.1-inch machine
cannon would make it possible to add considerable armor.40 The Secretary of the Navy
asked the War College for an analysis.41 The War College rejected the C&R idea: Speed
was expensive, but cutting it back to 27.5 knots went too far.42 It also rejected the idea
of replacing the 5-inch secondary guns.43 Above all, abolition of battleships was so bad a
policy that the United States should reject it, and the War College should go on record
again (as in the past) as strenuously opposed to it.
The War College also affected cruiser design in another way. Up to this time, all
cruisers, including large ones, had been armed with torpedoes as well as with guns.
Should that be the case this time? Given the treaty limit on overall cruiser size, the
weight usually invested in torpedo tubes could be invested instead in antiaircraft guns or
in armor, or both. The issue came up in General Board hearings. While the War College
was working on overall cruiser characteristics, the board asked it to settle the torpedo
question based on game experience.44 In the games, cruisers in a general action had only
about a 50-50 chance of firing torpedoes. The chance of firing torpedoes in a one-on-one
encounter, which might often happen in wartime, would be even less. There was also the
question of whether a shell hitting torpedo tubes might detonate the warheads.45
This study may have been the source of the War College perception that it was so
important to deal with torpedo attacks by enemy destroyers. In games, destroyers suc-

The War College and Cruisers 127

ceeded in fewer than 2 percent of attempted attacks, but it turned out that many of the
misses were due to errors by players and to lack of experience in conducting torpedo
attacks. Thus, destroyers would probably do much better in reality.46
The War College reviewed experience in the major tactical game in each of the class-
es of 1923 to 1928; the major tactical game and the tactical phase of the major strategic
game of the classes of 1929 and 1930; and of the two major tactical games played to date
(January 1931) by the class of 1931, a total of 12 games. Not all had been carried through
to a finish in which one side was completely routed, but all reached a decisive stage. The
study illustrates the kind of data war gaming could provide.
A total of 311 cruisers were involved (108 Blue). Of these, 156 fired torpedoes (61
Blue, 95 enemy): 56.5 percent of Blue cruisers and 46.8 percent of enemy cruisers fired,
an average of 50 percent. At the moment of firing, the average percentage damage already
imposed on the cruisers was 29 percent, almost all of it by gunfire. Cruisers hit by tor-
pedoes were so badly damaged that they could not fire their own torpedoes. A total of
99 cruisers (23 Blue) were sunk without firing torpedoes at all, an average of 32 percent.
Of these, 53 percent were lost to gunfire and 30 percent to torpedoes. In many cases,
gunfire was a contributing factor in the ship’s loss, but the tabulation showed only the
primary cause. At the end of the games, 102 cruisers still had torpedoes onboard (42
Blue), one third of all engaged. The total number of light cruisers lost was 190 (57 Blue),
an average of about 61 percent. All of this understated the damage to cruisers, since it was
now known that gunfire could detonate torpedo warheads. Moreover, as gun fire control
improved, it seemed that effective gun range would increase to the point where torpedoes
would be less and less attractive.
Until that year, cruisers had rarely been the object of air attack; they were damaged
mainly by gunfire. The difference must have been the rise of dive bombing, which offered
a good chance of hitting a fast-maneuvering cruiser. In the two tactical games of the Class
of 1931, Blue launched early air attacks against the enemy cruisers, while the opponents
concentrated on Blue capital ships and carriers. Out of 27 hostile light cruisers in the two
games, seven were sunk immediately and three more sank later due to heavy bomb dam-
age. None of the ten sunk had the chance to use their torpedoes. This amounted to about
40 percent of hostile cruisers lost to air attack. In the larger tabulation, the percentage of
damage sustained by light cruisers up to the instant of firing torpedoes had to be estimated.
However, limitations of torpedo fire due to damage received made it possible to come close.
The General Board took this experience seriously: It was more useful to protect a
cruiser against air attack than to provide her with torpedoes for rare offensive use. On
that basis, the board approved removal of torpedoes from existing U.S. treaty cruisers to
provide space and weight to double their 5-inch antiaircraft batteries. Torpedoes were
eliminated from the designs of the new better-armored cruisers. U.S. tacticians devel-
oped night tactics in which cruisers relied on their long-range guns and left torpedo at-
tacks to destroyers.

128 Winning a Future War

 Other navies thought otherwise. The Japanese gave their cruisers heavier torpedo
batteries and spent considerable effort on torpedo tactics and on long-range torpedoes.
These efforts paid off during the night actions in the Solomons during World War II.
U.S. destroyers did poorly with their own torpedoes until late in the war, at Leyte Gulf
in October 1944. The record of the Naval War College suggests strongly that this failure
cannot be attributed to a lack of interest in torpedoes, but rather to poor execution of tor-
pedo tactics. Evidence of sustained U.S. interest in torpedo attack includes the continued
construction of destroyers with unusually heavy torpedo batteries.47

Cruisers at War: Three Years of Red-Blue Warfare

Gaming showed that the U.S. Navy could probably never have enough cruisers. That was
graphically illustrated by a series of large-scale games simulating a Red-Blue war.48 The
first game in the series (Operations III, Class of 1932) was specifically designed to evalu-
ate the flight-deck cruiser, and more generally the variety of cruisers then planned. BuAer
chief Rear Admiral Moffett had recently testified before the House Naval Committee,
and there was a real possibility that one or more would be laid down. After the game,
Captain Adolphus Andrews, the War College chief of staff, considered the tactical em-
ployment of CLVs its most interesting aspect. The CLV had been introduced in the 1931
Red-Blue game in a conventional scenario. It had been very valuable in strategic scouting;
three such cruisers formed and held a scouting line across the Caribbean despite strong
opposition by Red heavy cruisers. Falling naturally into the dual roles of cruiser and car-
rier, and not being as efficient in either case as the pure type, the CLV was not as useful in
ship-to-ship battle. In the 1932 game, Red had CLVs, conducting protective rather than
offensive scouting. One question was how such a ship would function in low visibility,
when she could not operate her aircraft. In this case, the CLV was placed behind the
scouting line of cruisers for protection, but an enemy cruiser found her anyway. Because
aircraft could not fly under some weather conditions, it was impossible simply to substi-
tute CLVs and their aircraft for scouting cruisers ahead of a fleet. However, CLVs could
back up heavy cruisers in a scouting line, providing the needed punch if more powerful
enemy scouting forces appeared.
Using CLVs as scouts was controversial. When first proposed, they had been envis-
aged as cruisers with an extra punch offered by their aircraft. The idea of armed scouting
had been rejected. Instead, it was assumed that bombers should be held back until the ob-
jective was found, then launched. In any case, a CLV did not have enough aircraft onboard
to substitute for a carrier; if it was so regarded, it would be better simply to build carriers
(however, carriers were limited by treaty, whereas CLVs could still be built in numbers).
In the post-game wash-up, Captain S. C. Rowan, head of the Operations Department,
pointed out that the value of the extra punch depended on the efficiency of bombing (pre-
sumably dive bombing) “which we at present don’t know.” Tests against the old armored

The War College and Cruisers 129

cruiser Pittsburgh were ambiguous. In good weather, the attackers made 60 percent hits,
demonstrating their skill with the Assistant Secretary of the Navy watching. When the
Secretary of the Navy came out to watch in bad weather, they made no hits at all.
To compare a CLV with a 10,000 ton all-gun cruiser, three separate engagements
were tried. Although a change in the fire-effect rules had just devalued the 1,000-pound
bomb the CLV aircraft could deliver, in good visibility the CLV sank one light cruiser,
two armed merchant cruisers, and a converted seaplane tender. In a second case, a mixed
cruiser force (four CLVs, two heavy cruisers, six armed merchant cruisers) faced four
Blue heavy cruisers and one large Blue light cruiser. In this case, one Red CLV sank
two Blue heavy cruisers using 1,000-pound bombs, but Blue cruiser aircraft using light
bombs were able to knock out the flight deck of one Red CLV (with all its aircraft). A
third case was a low-visibility engagement in which CLVs could not fly aircraft. A Red
CLV made contact with a Blue light cruiser at 12,000 yards and was sunk. A Red heavy
cruiser then arrived and sank the Blue light cruiser. It was clear that the sacrifices in ar-
mament and protection required in order to provide air facilities made the CLV inferior
as a gun cruiser to a conventional large cruiser armed with 6-inch guns.
The game was fought far from Blue’s bases. That emphasized the need for long cruis-
ing radius, seakeeping qualities, protection, and high sustained speed—as differentiated
from, and far more important than, high maximum speed. That, in turn, meant large
ships. Although the Blue cruisers were individually more powerful than Red’s, they suf-
fered badly during raiding (anti-shipping) operations. Four of Blue’s 15 cruisers were
sunk. The Naval War College Research Department pointed out that, considering how
few cruisers Blue had compared to the number Red had plus the size of the Red merchant
fleet (some of which could be converted into armed merchant cruisers), the loss of cruis-
ers would mean far more to Blue than to Red.
The Department of Research (which at the time meant Captain Wilbur van Auken)
reached what must have been a shocking conclusion:
This brings up the point of the desirability of planning for the combined
use of cruisers and auxiliary cruisers by Blue and Red—if a war can be
visualized where Blue and Red have such paramount interests together as to
be allied. If such might be the case, it is to the advantage of Blue and Red
to maintain a preponderance in cruiser tonnage. . . . In view of the pres-
ent world’s unsettled political situation and the common interests of Blue
and Red, it seems worthwhile to propose for serious consideration and
planning the question of ‘pooling’ all vessels of Blue and Red of all types
to be ready to meet any emergency against one or more powers allied and
prepare strategic and tactical plans. With Blue’s known weakness in cruis-
ers, and the delays in laying down more, it is vital to plan all phases of how
Blue’s Navy might be increased n power by aid from Red. It is a fact that

130 Winning a Future War

 our combatant ships cannot be increased after a declaration of war to any
great extent. During the World War, less than ten destroyers authorized
for construction under the “250” program were completed. It also a fact
that one battleship requires four years to construct after authorization at
the minimum; an aircraft carrier, three years; one new cruiser (with plans
completed), three years; one submarine of 1,700 tons, two and one-half
years; and destroyers, about three months (if and when yards are organized
and prepared to go into an emergency program). And now, ship yards are
not as numerous or prepared as in 1917.

If war should come within three years, it is manifest that the United
States Fleet cannot be materially increased and would be about as it exists
in combatant ships today, except all types would be three years older. It
is therefore seen how the Blue types stand relative to Orange and Red.
With no change to add trained fighting ships (for example against Orange
and her Allies) it would seem that strategic and tactical studies should be
made, games played and plans evolved, for the best possible use of all vessels
of all types, bases, and the logistics for Blue’s navy in a coalition with Red.
This is the only practicable method of Blue’s increasing her battleship, car-
rier, cruiser, or destroyer tonnage besides making available vessels for the
convoy of our overseas force. An even though such may never come, the
problem is one which merits the serious consideration of all officers of our
Navy and Army [emphasis in the original].
This must be one of the earliest references to such a possibility in official U.S. Navy
documents. Van Auken’s analysis is in the OPNAV War Plans Division set of papers
for Operations III, not in the War College set; it got to the Washington planners. Van
Auken’s coalition games were apparently not played; at least no records of such coalition
games prior to 1941 appear to have survived. Van Auken’s pessimism explains why rapid
U.S. fleet expansion, a key World War II reality, did not figure in inter-war gaming. He
repeated his call for coalition planning in a commentary on a later Blue-Orange game.
There is no indication that it had any impact; it was absolutely the opposite of the U.S.
official position. It says a great deal for independence of thought at Newport that van
Auken did not suffer for his comments.
Van Auken did not say so, but an important question was whether cruiser opera-
tions against enemy trade—an important part of any campaign against Japan—were still
viable, or whether such operations would have to be turned over to submarines. Several
later war games examined the submarine option.
Unusually, the three games involved portrayed a war in which the U.S. strategy
was to choke British trade, rather than the typical defensive war in the Western Hemi-

The War College and Cruisers 131

sphere.49 The basic U.S. strategy was to eject Red from the hemisphere while imposing
economic pressure as a way of depriving Red of food and raw materials. The war had to
be fought in such a way that Blue would not be so badly damaged that other countries,
such as Orange, could take advantage of the post-war situation. Naval and military action
would secure Blue’s vital sea areas, dominate Red territory in the Western Hemisphere,
prevent the passage of Red trade through the south Atlantic, and draw the Red fleet
into a general engagement on disadvantageous terms. Blue could certainly block traffic
through the south Atlantic, but the British could shift their traffic to the Suez-Medi-
terranean-Gibraltar route, outbound traffic passing down the West African coast and
around the cape. Both navies threw out cruisers across the Atlantic narrows, between the
Azores and the coast of Brazil. A clash between the British cruiser screen (intended to
protect trade further north) and Blue raiding groups led to the destruction of two Blue
heavy cruisers and the destruction of the flight deck of a Red CLV.
Tactical IV (played in December 1931 for the Class of 1932) was a continuation of
Operations III.50 It was a partial fleet engagement testing the effect of carrier superiority:
Red had four fleet carriers, Blue one. Blue was stronger in battleships (12 to four), but Red
was stronger in destroyers (60 in six flotillas compared to two destroyer squadrons and one
destroyer division for Blue). Red still had one flight-deck cruiser, but she had no aircraft on-
board. Blue’s fleet was slower than Red’s, and Red had reinforcements on the way. The forc-
es were so placed that Red could open with a night destroyer torpedo attack against Blue.51
The Red destroyers never managed to hit the Blue battleships, although they did sink one
heavy and three light cruisers and 12 destroyers (Red lost 25 destroyers and two leaders).
One Red half-flotilla got within six miles of the Blue battleships, but did not know it.
Before dawn, both sides launched big air attacks. Red’s carriers were on the flanks
and in the rear of his battle line. They sent out scouts, and then launched bombers and
torpedo planes that hit the Blue battleships. Red met considerable opposition from Blue
fighters and anti-aircraft guns. A later attack destroyed the flight deck of the Blue car-
rier using 116-pound bombs. Meanwhile, four Blue heavy cruisers caught one of Red’s
carriers and ultimately sank her at the cost of one heavy cruiser. Another Red carrier
had her flight deck damaged. Both sides lost many aircraft. The carrier engagements
once more demonstrated the vulnerability of carrier flight decks and also the danger of
encountering enemy heavy cruisers.52 The game also showed how important carrier speed
was: Blue’s 32.5-knot heavy cruisers easily ran down a 24-knot Red carrier (presumably
either Hermes or Eagle).
That night, both fleets launched night destroyer attacks, the two striking forces
meeting quickly after dark. In the melee, the side with superior numbers at the point of
contact won. Each broke through to the other’s battle line, albeit at a considerable cost.
The game was called soon afterward.
The scenario continued with Operations III (Class of 1933), which was paired
with Tactics III.53 Like the 1932 game, this one was concerned with the attack on Red

132 Winning a Future War

seaborne trade to exert economic pressure. Red might be able to occupy Portuguese
possessions in the Atlantic (the Azores and the Cape Verde Islands); Blue might gain
the use of Liberian ports. Blue would have liked to seize a base in the Azores, but could
not do so as long as Red was of roughly equal strength. Overall, Blue suffered because
it lacked an advanced base in the Eastern Atlantic. The issue of Red access to the Por-
tuguese islands led to a discussion of the status of the Mandated Islands in a Blue-Or-
ange war and the status of Canada in a Blue-Red war. The opinion was expressed that
it would be Blue’s policy to try to bring Canada into the war on the Blue side, despite
her connection with Red.
A “special situation” to be considered envisioned a collision blocking the Suez Canal,
causing much British shipping to be rerouted through the Cape Verde focal area, where
it would be subject to attack by Blue forces and to defense by Red forces in West Africa.
The Research Department considered the game useful as a reminder of how dependent
the British were on overseas trade, but did not add that the only country more dependent
on overseas trade than Great Britain was Japan, the likeliest future enemy. The game
demonstrated that a huge number of cruisers and aircraft would be needed to prosecute
a trade war against Red, and that it was futile simply to sweep enemy areas with large
forces. The game also brought out the need for detailed coordinated search plans.
Both sides made extensive use of air power. In this game, Blue had six fleet carriers
(three carrier divisions), although as yet the U.S. Navy had only two carriers, with a third
under construction and two more authorized. A fourth carrier division consisted of a
seaplane tender, an airship tender, and the large airship Akron.54 There were also seven
small converted carriers (two of which were initially troop transports), of which three
would be available on the West and four on the East Coast 120 days after mobilization;
and 14 converted seaplane tenders.
Both sides had flight-deck cruisers. The game again showed how vulnerable their
flight decks were once they were found by enemy air forces, presumably because they
could not put up extensive fighter opposition. Of three Red flight-deck cruisers sunk, one
was lost to submarine torpedoes and two to enemy cruiser gunfire. However, all three
Blue heavy cruisers and all three Blue flight-deck cruisers lost were sunk by bombing,
presumably dive bombing. Even so, flight-deck cruisers seemed to be justified by their val-
ue in sudden encounters with enemy cruisers, and particularly in their ability to quickly
develop contacts by air scouting.
In the corresponding Tactical III game (January 1933), each side sought a fleet ac-
tion to relieve domestic pressure for decisive action. The Red commander was ordered
to fight if the situation was favorable, but he knew that defeat would destroy the entire
Red Empire. He told his government that conditions were unfavorable, but did not know
whether it understood his fleet was inferior.55 Both naval commanders knew that the
Red fleet was the weaker of the two. In this game Blue did have a base in the Azores, but
its hold was precarious. Each commander wanted to avoid fighting within the radius of

The War College and Cruisers 133

land-based aircraft. For Blue that meant avoiding the area near Portugal; for Red it meant
staying out of the area around the Azores.
Blue estimated that he was superior in capital ships and probably in the air, and
about equal otherwise. He therefore chose to make an early air reconnaissance followed
by an early maximum-strength air attack to slow the Red battle line, followed by a de-
cisive general action. Blue’s air plan set priorities: the enemy carriers first, then cruisers
in the van (i.e., the enemy scouts), and finally the enemy battle line. In its critique, the
Research Department argued that Blue should have concentrated everything on the Red
battle line, “the foundation and most of the vital framing of the enemy’s structure. The
other arms form some of the substance, but all of the trimmings. In the give and take of
a naval battle I believe the structure which loses some of its substance and all of its trim-
mings is much more able to stand up than one which has its vital framing and foundation
weakened.” The critique conceded that bombs and torpedoes would not give such visible
effect on battleships as they would on cruisers and carries, but they would weaken the
enemy battleships and thus make gunfire more effective. Aerial torpedoes could slow the
enemy down and make it difficult for him to evade action.
Red had to weaken Blue’s battle line. He hoped to do that using submarines, but the
critique pointed out just how much they had to achieve; it seemed unlikely that Red had
enough submarines to do that. The idea of using submarines against enemy battleships
was of current interest at the War College. Red enjoyed an advantage in submarine speed,
so that its submarines could maneuver to remain in position relative to the Blue battle
line, whereas Blue’s submarines could not maintain their positions relative to Red’s battle
line. By the end of the game, submarines on both sides had made successful attacks. The
Research Department considered their hitting rate, 8.5 percent, far lower than that to be
expected in war. The game showed that if the enemy sought action, the U.S. battle line
could be maneuvered to draw the enemy’s battle line into submarine-attack range. The
threat of such attacks had been an important factor in British tactics at Jutland in 1916.
During the war, the British had built fleet submarines specifically to exploit opportuni-
ties created by the maneuvers by their own fleet. The Research Department concluded
that submarines would likely play an important role in future fleet actions. Tactics had to
be adjusted to exploit their capabilities.
The Research Department considered air operations one of the most interesting
parts of the game. “The air did in one hour what the destroyers had failed to do after
working all night.” Both Red and Blue air commanders were student officers who had
had recent fleet aircraft experience in command of Lexington and Saratoga. Their tactics
were what the fleet currently planned to use. The Blue air force attacked the Red carriers
140 miles away. It sank one Red carrier and destroyed the flight decks of three more.
This was the sort of attack, staged when ships were far out of sight of each other, that
seemed so revolutionary in the air-sea battles fought a decade later, for example at Coral
Sea and Midway.

134 Winning a Future War

 Either deliberately or by chance, Blue attacked when the Red planes were being ser-
viced aboard. This attack was quite expensive: none of the Blue planes got back to the
Blue carriers. Even after receiving tremendous damage from Blue, Red’s surviving car-
riers managed to make three torpedo hits on the Blue battle line. Red later wrecked the
flight decks of both Blue carriers. When the game ended, each side had four carriers, but
their flight decks had been so badly damaged and aircraft losses so heavy that they were
unlikely to play any further role in the battle that day. Blue lost a good proportion of its
spotting aircraft. Without enough spotters, its battleships would have had to fight at re-
duced gun range, but the game was called before the battle lines engaged. Dive bombing
sank three Red cruisers (including one flight-deck cruiser) and five Blue cruisers during
the tactical game. Such attacks also cost two Blue cruisers 50 percent damage.
The Research Department asked whether the game was realistic. Would the U.S. pop-
ulation tolerate removal of the whole fleet from the Atlantic coast? Would it tolerate staking
the safety of the country on a single decisive battle? The question recalled the experience of
the Spanish-American War. As the Spanish squadron crossed the Atlantic, the politicians
representing cities on the Atlantic coast demanded warships to protect their cities. Na-
val strategists had concentrated all the modern U.S. warships so that they could fight the
Spanish. Fortunately, the U.S. Navy had managed to assign obsolete ships, even Civil War
monitors, to the cities. Fortunately, too, the Spanish had not used the concentration of the
U.S. fleet as an opportunity to strike at one of the rich coastal cities. The U.S. Navy had had
no way of knowing where the Spanish squadron was headed. It took time to realize that it
was in Cuba. After that the fleet concentrated off its Cuban port for a decisive battle. Would
such concentration, particularly in distant waters, ever be possible in a future emergency?
Without it, the U.S. fleet could not be sufficiently powerful to win the decisive battle.
As it turned out, no real public-relations problem developed when the Pacific Fleet
was concentrated far from the Pacific coast in 1940. It deployed from its usual base at
San Diego to its war base at Pearl Harbor. To some extent, the public may have been re-
assured by claims by the U.S. Army that its aircraft and coastal batteries were sufficient
direct protection of the U.S. coast. Despite previous joint exercises with the Army War
College, the Research Department did not raise the possibility that the Army’s efforts
would make just the sort of concentration the Navy required a practicable proposition.
The last game in the series was Operations IV of the Class of 1934, paired with
Tactics IV.56 This time, Blue sought a base in Portuguese Guinea, in Africa (the Bissagos
Islands off the coast), from which its fleet could cover the entire western north Atlantic,
the south Atlantic, and the Caribbean, and also disrupt Red trade along the African
coast. Portugal was Red’s ally, which justified seizure of the base. This was considered
part of the second phase of the war, the attack on British commerce. Should the seizure
of the base succeed, the next step would be to seek control of British trade routes north of
Gibraltar. It was assumed that Red could not tolerate more than three weeks’ interrup-
tion of the trade between Gibraltar and the Channel.

The War College and Cruisers 135

 Geography favored Red, which had many bases in the theater of operations. Red’s
line of communications from Gibraltar would be short and well-protected, but Blue had
a long line of communications stretching back from the African coast. Blue was not only
far from home, but also far from the target area: 1,800 miles from Gibraltar, 2,200 from
Trinidad, 3,400 from New York. The situation was not too different from that in an Or-
ange war, with Blue penetrating deeper and deeper toward Orange through areas full of
Orange bases. Red’s only logistical problem was that it had been cut off from New World
oil supplies, hence had to rely on European and East Indian sources (the Middle East was
not yet a source of oil). However, Red had at least six months and probably a year of oil
stored at home.
To fight so far from home would be a tremendous undertaking. The writer of the
critique, Captain George B. Wright, pointed out that either Red or Blue would need
some compelling reason to undertake so vast a task as extend its naval power into the
other’s sphere of control in a war between them. He did not rule that out, and he pointed
out that governments might undertake such efforts “particularly if the risks of such an
undertaking are not clearly foreseen by responsible naval authorities.” It would be vital to
show what the risks were, “and if there is any reasonable way in which they may be avoid-
ed or minimized.” The main lesson of the games was to show how difficult either side
would find it to project its power across the Atlantic. Red projection was tested in games
in which Red tried to seize or defend a base in the Caribbean or in Canada. The current
game tested Blue projection in the opposite direction. Operations became prohibitive, or
nearly so, when the power on the offensive lacked both allies and a secure base capable of
both receiving and maintaining the fleet. Even so, Wright could imagine several possible
forces that would justify a transatlantic war: public opinion, economic pressure, social
conditions at home, international relations, or political factors.
Red’s fleet covered Red trade well enough that in order to block that trade, and win
the war, Blue had to destroy that fleet. Thus, Blue’s problem was really to bring the Red
fleet to battle. Equating the destruction of Red’s fleet to strangulation of Red’s vital com-
merce was equivalent to the assumption in the Orange war plans and games: The Orange
fleet had to be destroyed as a prerequisite. The other question in the Red-Blue game was
how to compel Red to fight even though both sides understood that Red’s fleet was infe-
rior, particularly in long-range gunnery.
It was possible that Blue’s seizure of a base off of the African coast would affect Red
public or political opinion in such a way as to force the Red fleet to fight, but that could
not be guaranteed. The Research Department pointed out that the previous year’s Blue
“Estimate of the Situation” had missed the key point of Blue’s battle line superiority, con-
centrating instead on operations near Cape Verde to control Red trade.
Blue estimated that Red would see little point in accepting battle anywhere south of
the Azores-Madeira line, since in that case it would be accepting the same disadvantage
of a long line of communications that Blue had accepted. Red would most likely attempt

136 Winning a Future War

to wear Blue down as it moved north in preparation for a fleet engagement. This was
much the attrition war the U.S. Navy assumed the Japanese planned against a U.S. fleet
heading west.
Red formed both a main body and a separate raiding force built around its three
battlecruisers and its carriers. The raiding force was intended to operate against Blue
carriers and other heavy units using its aircraft and night destroyer attacks. Not only did
the main Red raiding force attack on Blue develop into an air attack as intended; Blue’s
main counterattack was also by air.
Once the raiding force was heavily engaged with the Blue covering force, the Red
main body turned away. Only at the end of the game did it approach the Azores to fight
a decisive day battle against the main Blue fleet. The threat of the Red battleships forced
Blue to expose its battleships to harassing attacks while accompanying his base force
and train to the Azores. The Red battle line remained intact, but the Blue battle line
was exposed to air and submarine attack. This was much what the Japanese managed in
Blue-Orange campaigns: attrition, possibly without any decisive battle.
The key strategic point was that each commander had to understand what the mis-
sion really was. Was it simply to attack trade, or was it to gain supremacy by destroying
the enemy’s fleet? Students tended to concentrate on the immediate task of attacking
enemy trade. Few of them realized how much superior Blue was, due to a combination of
superior battle line and air strength. Several student plans might have been different had
they realized the actual relative strength of the two sides. That was a very serious criti-
cism since the whole point of much of the analysis done by the War College was to make
such relationships clear. No one mentioned dispersing shipping as a means of protecting
trade, yet, Wright wrote in the critique, “it is the cheapest and in war the most commonly
used form of protection.”
Like many other games, this one involved very large air forces. Red had eight carriers
to Blue’s seven, but Blue had many more aircraft: 798 to 479.57 Each side had three flight-
deck cruisers, each with 18 heavy dive bombers onboard. Blue also had six seaplane ten-
ders supporting 54 floatplane torpedo bombers and 54 patrol planes, compared to four
Red tenders supporting 48 patrol planes. Blue carriers were generally much faster than
Red’s (except for Langley, still a carrier at this time) and much longer-legged. How much
did the difference in carrier numbers count? The main point developed by the critique
was that massive numbers of aircraft were needed. When the enemy had adequate air
forces, only aircraft could develop contacts with screening ships into useful information
about the enemy’s position and disposition. There was little point in mounting air attacks
against an enemy with an approximately equal air force unless some particular effort was
made to catch his aircraft on deck. It seemed that long-range shore-based aircraft would
be very useful for this purpose, because their attacks would not reveal the position of
the carriers. To the Research Department, outstanding features of the exercise included
large-scale air attacks by both Red and Blue.58

The War College and Cruisers 137

 Neither side could prevent the other from scouting by air. Generally each carrier di-
vision (two carriers) located its objective with its own bombers and then sent out strikes.
That seemed necessary given the rapidly changing situation and the great dispersal of
enemy forces. Blue lost more aircraft when one of its carriers was torpedoed than by
all previous air operations. At the end of the problem, Blue had one intact carrier, one
with some underwater damage, and two seriously damaged with their flight decks out
of action. Three more had been sunk. Research attributed the relatively poor showing
of the Blue carriers to a fundamental mistake: Blue had used carriers to overwhelm the
modest Red shore defenses. In doing that, Blue had left the Red carriers full freedom of
action. Blue had forgotten the developing rule that there could be no security as long as
the enemy carriers remained.
Red found that a scouting flight out to 150 miles was not nearly as effective as had
been imagined, as it found only one of the Blue carriers. Red had been standing by with
carrier bombers and patrol planes to attack based on this one flight. On the other hand,
Blue needed three carriers to support his landing operation, because its fighters were
distributed among them. That brought the carriers close to the beach, hence vulnerable
to discovery and attack. One of the players thought that Blue should have used fewer of
his carrier aircraft to support the landing. Had he used some of them as scouts, he would
have found a Red force to the east.59 Air scouting was done mainly by carrier aircraft and
patrol seaplanes, not by cruiser or battleship aircraft.
As in previous games, it was assumed that whoever attacked an enemy carrier first had
an excellent chance of destroying all or part of her flight deck, rendering her ineffective. In
this particular game, one Blue carrier was sunk by initial air attacks and one had her flight
deck ruined. One Blue flight-deck cruiser was sunk by air attack and the other two had their
flight decks ruined. Two Red carriers were sunk and a third had her flight deck ruined; all
three Red flight-deck cruisers were sunk. Initial air attacks also sank four Blue heavy cruis-
ers, two Blue light cruisers, and two Blue destroyers. Red lost five heavy cruisers and six
destroyers, including a flotilla leader. Capital ships on both sides were hit, but not sunk. In
addition, Red submarines sank the damaged Blue carrier that had retained her flight deck,
nearly sank another (she was later sunk in a night destroyer attack), and badly damaged a
third. Blue battleships and cruisers were hit. One of the two damaged flight-deck cruisers
was sunk, together with three heavy cruisers (two of them damaged and hence slowed). The
last Red air attacks sank a Blue carrier and ruined the flight decks of two more.60
The action was somewhat unrealistic in that at least one large air battle involved
aircraft of such disparate types that they would not have worked together. For example,
there was an air battle between 30 Red bombers and 42 Blue observation aircraft from
Blue battleships. Normally the bombers would have so outperformed the observation
aircraft that they would never have fought each other at all.
Both sides used flight-deck cruisers as carriers rather than as cruisers (for screen-
ing). The Research Department pointed out that flight-deck cruisers had been conceived

138 Winning a Future War

as improved scouting cruisers that could use their aircraft to make and develop initial
contacts better than normal cruisers because they had more (and higher-performance)
aircraft. This idea was not tried out. In the game, these small carriers of “a distinctly of-
fensive type” proved useful. Their early loss on both sides could be attributed to the way
they were used—in easily located positions, which the enemy did not have to disperse to
find—rather than to their assigned tasks.
When Red aircraft struck the Blue main body, the Blue flight-deck cruisers (with
bombers onboard) tried unsuccessfully to help repel the attack; one was sunk and anoth-
er lost her flight deck, so that flight-deck cruiser aircraft in the air had only one ship left
on which to land. At this point, 40 of the original 54 bombers were in the air, and they
put the flight decks of two Red carriers out of action. Nine of them recovered onboard the
remaining flight-deck cruiser. They were launched again to sink the Red carrier previous-
ly crippled. Another 12 took off but missed the Red fleet; ten more were re-serviced and
took off. These last ten sank a Red flight-deck cruiser. Surviving bombers from the Red
flight-deck cruisers landed on Red carriers, reinforcing their air groups.
During Red air attacks on Blue and Blue air counterattacks, Red’s three flight-deck
cruisers were kept near each other and near the Red carriers; all were lost to air attack.
Their 54 bombers were launched against the Blue carriers, but this order was then can-
celled in favor of attacking a Blue heavy cruiser. Before they arrived, this ship had already
been sunk, and they were ordered back. En route they found and attacked Blue cruisers,
21 bombers being shot down in the process. The survivors helped defend the Red carrier
on which they were to land. Survivors eventually landed on a Red flight-deck cruiser.

Postscript: The Fate of the Flight-Deck Cruiser

The flight-deck cruiser was undoubtedly the most radical element of the War College’s vi-
sion of a future balanced U.S. fleet. In August 1931, the War College declared the cruiser
of paramount value in the “service of security and information” (scouting and screening)
and in sea control. The latter meant finding and destroying individual enemy ships. Their
aircraft should be first scouts, second dive bombers, and third (if practical) defenders. In
Blue-Orange games in 1931–32, these ships gained early information of the enemy upon
which major operations were based. The Naval War College evaluation was that even
if the flight-deck cruisers were sunk after carrying out their scouting mission, they had
already done their share in making the overall operation a success. Generally, they were
better able to hide than the big carriers due to their small size and their high speed. They
showed valuable flexibility. Against all that, their flight decks were just as vulnerable as
those of the big carriers, and without those decks, or in weather that prohibited flying,
they lost their special scouting value. Without her aircraft, the flight-deck cruiser could
not stand up against other cruisers with more powerful gun armament; she had to be
supported tactically by other cruisers.

The War College and Cruisers 139

 CNO Admiral Pratt, who took credit for the clause of the London Naval Treaty
that allowed the United States to build flight-deck cruisers, included one in the projected
1933 program.61 The General Board submitted characteristics in January 1931 and that
May it included a flight-deck cruiser in the projected 1932 program. This recommenda-
tion was repeated in November and December 1931. This proposal was serious enough
that the Bureau of Construction and Repair produced a set of contract plans for the ship,
tentatively designated Cruiser No. 39 (this hull number was actually applied to the heavy
cruiser Quincy).62
Support for the flight-deck cruiser was by no means unanimous.63 For example, in
September 1930, CinC Battle Fleet argued that to combine a gun cruiser with a flight
deck was “unsound in every way.” He pointed to the severe fire hazard presented by both
airplanes and their fuel, “so great that we shall have even under the most advantageous
conditions, a timid cruiser, one fearful of being brought into position where its guns are
to be used. It is my experience that airplane carriers demand for their protection, even in
problem work [i.e., fleet exercises] more potential offensive power than they themselves
can exercise….The more the Navy is reduced [by treaty], the more important it becomes
that every unit of the Fleet shall be offensive in its character, and the offense to be con-
tained within itself.”64 Moreover, it seemed that a single 4- or 5-inch shell could put the
ship out of action by setting her fuel or airplanes on fire. However, Admiral Pratt com-
pletely rejected these arguments.
The FY33 program was intended as the first in a ten-year program submitted by the
General Board in December 1931. Even after the first year’s program was cut to conform
to President Hoover’s proposed budget, it included one flight deck cruiser. This program
was abandoned due to the Depression. In September 1932, the General Board recom-
mended two 20,000-ton carriers for FY34. It omitted the flight-deck cruiser, which was
clearly less important. The two carriers were built as Yorktown and Enterprise after the
FY34 program veered back and forth in content. The flight-deck cruiser vanished from
the board’s formal construction plans after that.
The main sponsors of the flight deck cruiser disappeared in 1933. Admiral Pratt re-
tired as CNO. Rear Admiral Moffett was killed when the airship Akron crashed. Admiral
Laning left the War College, and flight-deck cruisers disappeared from its games. The
death of Admiral Moffett was probably by far the most important. He enjoyed enormous
political influence; he can presumably be credited with House interest in flight-deck cruis-
ers and also with the perception, in 1931, that Congress would authorize such a ship, but
not pure 6-inch gun cruisers. For its part, from 1932 on the General Board was more
interested in the two large carriers (the Yorktown class) then planned. It saw them, and a
further smaller carrier that could be built on the rest of the available tonnage, as a better
way to provide more aircraft to the fleet. There was no one of Moffett’s stature to press
the argument that even with all of the carriers, the fleet would not have nearly enough
aircraft, or that the fleet badly needed dispersed as well as concentrated air power. This

140 Winning a Future War

argument, moreover, was closely associated with simulations of an Orange war. By 1934,
War College influence was waning (see the next chapter).
In 1934, the War Plans Division, which received reports of war games, asked for
at least one flight-deck cruiser; the General Board sought fleet opinion. The ship, if any,
would be included in the FY36 program. In October 1934, Fleet CinC Admiral Reeves
wrote a letter strongly opposing the idea on the ground that carriers and cruisers “sim-
ply do not mix”; the proposed cruiser did not have a sufficiently long flight deck. “From
fleet experience, a flight deck 200 ft long will restrict flight operations to an intolerable
extent.” That such a deck would be extremely dangerous was shown by the numerous
barrier crashes onboard existing carriers with much longer flight decks leading up to
their barriers. Reeves was certainly pro-aviation; he was “completely cognizant of the
vital need for more and more airplanes in the fleet particularly for scouting.” His solution
was “vigorous promotion of the patrol plane program including the preparation of addi-
tional tenders and patrol plane bases.”
The question the General Board asked was whether it was worth building a flight-
deck cruiser for tests, or whether the concept could be tested using suitably modified real
carriers—but the need for such a ship declined as new carriers were completed. Not only
would five carriers be ready before a single flight-deck cruiser could be completed, but a
new conference due in 1935 might allow the United States more tonnage. Since the Jap-
anese had already announced their withdrawal from the treaty system, there was reason
to imagine that the 1935 conference might fail altogether, ending all limitation. In fact,
the new conference removed total tonnage limitations, but U.S. legislation continued to
key construction to the earlier limits, something no one could have expected. Any new
flight-deck cruiser would hardly be begun before the conference changed the rules. On
the other hand, the fleet had a pressing need for more conventional cruisers.
Moffett’s successor at BuAer, Rear Admiral Ernest J. King, proposed a new version
of the flight-deck cruiser in March 1935. He cited new high-lift devices as a rejoinder to
Reeves’s critique. King had nothing like Moffett’s clout. The idea was formally rejected
by the Secretary of the Navy on 29 March 1935 on the grounds the General Board had
stated: All cruiser tonnage should go into conventional types. CNO recommended that
no further steps be taken until “it is practicable and allowable to build types other than
those now authorized.” BuAer did not give up. It revived the flight-deck cruiser idea in
February 1936, when proposed disposal of the ten Omaha-class light cruisers to Brazil
would free up tonnage.65
The flight-deck cruiser idea survived as late as 1940, when the Navy developed a
large new cruiser program, but no such ship was ever built. It might be argued that the
World War II conversion of light cruisers to light carriers (CVL) may have been inspired
by the flight-deck cruiser idea, but these ships were never used the way flight-deck cruis-
ers were played in war games.

The War College and Cruisers 141

6. D

T he role of the Naval War College as the U.S. Navy’s primary tactical laboratory did
not last. In October 1934, the Naval War College was transferred from OPNAV
to the U.S. Navy school system administered by the Bureau of Navigation (BuNav, later
renamed the Bureau of Naval Personnel).1 In August 1932, its chief, Rear Admiral F. B.
Upham, argued that the college was no different in principle from the schools he already
administered—the Naval Academy and the Naval Postgraduate School.2 The argument
made here is that, at least up to 1934, that was not at all true: The Naval War College was
at least as valuable as a laboratory, thanks to its unique war-gaming capacity. The BuNav
argument had no chance of succeeding as long as the CNO was a former Naval War Col-
lege president who had relied on war-gaming arguments to support policy, particularly
air policy. It also seems likely that, as Naval War College president, Admiral Laning
would have argued forcefully for the special capability his school represented.
It would seem either that Upham was ignorant of the special role of the Naval War
College or that he was aware that it was becoming much less popular within the upper
reaches of the Navy. In August 1932, Admiral Pratt’s term as CNO still had nearly a
year to run, as did Admiral Laning’s term as Naval War College president. The president
succeeding Admiral Laning, Admiral Luke McNamee, clearly considered the laboratory
role vital. Pratt’s successor as CNO, Admiral William H. Standley, took the result of
the 1933 operations game seriously—he ordered a dramatic change in the war plan. That
makes the shift to BuNav under Standley’s tenure somewhat puzzling. The difference
may have been connected with the departure of Admiral McNamee from the college. He
requested early retirement to become president of the Mackay Radio and Telegraphy and
Federal Telegraph companies, which had important strategic roles. It is, of course, also
quite possible that Admiral McNamee decided to leave the college halfway through his
term because he saw the change of status coming.

142 Winning a Future War

 It is also possible that McNamee’s successor, Rear Admiral Edward C. Kalbfus, took
the change as an indication that the role of the college was due to shift in favor of its ed-
ucational mission. He seems to have emphasized that mission, not just inside the college
but also outside it, in the form of the correspondence course. In 1934, the fleet was just
beginning to expand, and it could be argued that war was coming. In that case, it was may
have seemed much more important to teach as many officers as possible to make sound
decisions (using the methods developed by the college), than to provide the Navy with
detailed advice as to the characteristics of its ships or the nature of its war plans. After all,
in the latter mission the college was in effect competing with other naval organizations.
The run-up to the London Naval Treaty illustrates one of Admiral McNamee’s last
contributions, which was based on war gaming. In January 1934, the Secretary of the
Navy asked the General Board for guidance as to the U.S. position at the forthcoming
conference, which would frame a treaty to replace the existing Washington and London
naval arms-control treaties in 1936. Advice was urgent because a U.S.-British preparatory
meeting was scheduled for 18 June–19 July 1934. Late in February, McNamee wrote a
memo for the board.3 Although the memo nowhere refers to war gaming, it is evidently
based on recent game experience. McNamee began with the “present disturbed state of
the world.” The “one clear outstanding fact is that the safety and security of our nation
will depend to a large extent upon an efficient Navy….” That was no surprise, but what
followed was: McNamee pointed out that without an adequate base in the Philippines, the
fleet could not in itself exercise decisive influence in the Far East—the Japanese were just
too strong. That echoed the conclusions of the 1933 operations game and its 1934 sequel,
in which the fleet could not hold Dumanquilas, its temporary base, while protecting a vital
convoy bringing it supplies. “A war between Japan and the United States alone under pres-
ent conditions would involve us in losses entirely out of proportion to any possible gain.”
McNamee was also apparently aware of Captain van Auken’s call for cooperation or
even alliance with the British. Although the U.S. Navy was not strong enough to restrain
the Japanese, “the British and U.S. navies combined could enforce any reasonable restric-
tions on Japanese policies.” Since British interests in the Far East were far greater than
those of the United States, McNamee thought “it would appear wiser for us to support
her where it is to our interest to do so, than to be placed in the position of fighting her
battles for her. The more clearly Great Britain understands that this will be our policy,
the surer we will be of obtaining joint action if and when it may become necessary.”
Without adequate defense, the Philippines were a liability, and any attempt to forti-
fy them now would create grave tensions with Japan. McNamee suggested some possible
offers to Japan in return for more favorable naval ratios. Overall, he called for decisions
prior to the conference, for example as to whether the Navy should be able to control
the western Pacific or should be limited to control of the eastern Pacific and the western
Atlantic. This was very much a consequence of campaign gaming, which showed just how
much naval power a western Pacific offensive would require.

Downfall 143
After 1934, the War College stopped providing advice to the U.S. naval leadership based on war gaming. It
is not at all clear what happened, but it does appear that the War College was moved from OpNav to the
navy educational system, and its educational role superseded its think tank role. At the time, it was led by
Rear Admiral Edward C. Kalbfus, shown here as an admiral at the change of command ceremony when he
relinquished command of the Battle Force. It is possible that the driving consideration was that the fleet was
beginning to expand and that the War College had to reach a wider audience. Kalbfus certainly thought that
way; he was very much interested in the college’s correspondence course. (NHHC NH 120063)

144 Winning a Future War

 McNamee also pointed out that U.S. policy of encouraging general reductions in
naval power by U.S. example had been a total failure. The United States could out-build
both Britain and Japan. Doing so would improve the U.S. strategic position and also pro-
vide shipyard employment. He was far from alone in pointing this out, and his vision was
realized in the Vinson-Trammell Act passed later in the year. This was not an inference
from gaming, but games did show just how heavy wartime losses might be.
Overall, McNamee recommended extending the London Naval Treaty with only
minor modification because it provided equality with the British, “thereby insuring to us
the maximum naval security attainable.” It defined a “treaty” navy “which Congress and
the people now understand as the meaning of ‘Adequate’ Navy, which they have always
been willing to support.” It put the United States in a stronger political position not to
ask for anything more; considering the political situation, anything else would be worse.
If proposed changes came from another power, the United States would not be caught up
in justifying unacceptable ones. Finally, McNamee thought that the British wanted and
needed a better ratio with Japan than the United States did, but it wanted the United
States to press for that ratio.
McNamee’s belief that Congress and the American people would accept a treaty
navy was certainly in line with the Vinson-Trammell Act, and it may have helped inspire
it. The act replaced separate authorizations (not appropriations) with a blanket authori-
zation for ships to fill the tonnages allowed under the Washington and London (1930)
Treaties as an effort to maintain a “modern Treaty Navy.” The Naval War College had
already studied the characteristics of a “balanced Treaty Navy,” but it is not clear whether
McNamee had that work in mind at the time.
For McNamee, desirable modifications to the London Treaty of 1930 included clari-
fication of the meaning of “flight-deck cruisers” to insure that planes could fly off as well as
on. He argued that the United States should accept the proposed abolition of submarines
if all countries agreed to scrap fully and at once, and to accept inspection. The destroyer
quota should not be changed “unless every submarine in the world is destroyed. They
might otherwise be supplied by an allied minor power.” McNamee absolutely rejected
any change in cruiser ratios.4
For the British, the most important goal in a new treaty would be a dramatic reduc-
tion in the size—hence the cost—of new battleships, to 25,000 tons with 12-inch guns.
To McNamee that was by far the worst danger in a new treaty. War-game experience
(not cited) showed how much bomb and torpedo damage battleships could be expected to
receive before they faced other battleships in the western Pacific. It took considerable ton-
nage to provide the requisite protection as well as protection against existing battleship
guns. McNamee argued that the battleship tonnage limit should be increased to 40,000
tons so that ships could be armed with the 16-inch/50-caliber gun, which was known to
be far superior to the 16-inch/45-caliber model arming existing ships, not to mention the
14-inch gun. “If the tonnage of a battleship is reduced it cannot retain the characteristics

Downfall 145
The decline of the War College (or, rather, that of its influence) might be considered a reaction to the air-mind-
edness shown by Admiral Laning in 1930–33. It took a long time to build the remaining carriers that could
be acquired within treaty limits, and the U.S. Navy did not accelerate carrier construction once treaty limits
(on total tonnage) had been removed by the London Naval Treaty of 1936. However, any very visible naval
program had to survive attacks by a very powerful pacifist lobby. Naval rearmament, under the Vinson Act of
1934, was justified as maintenance of a modern treaty Navy, and there was widespread sentiment within the
Navy that any further expansion had to be tied in some way to the earlier naval arms treaties and the force
ratios between countries they embodied. Any such policy had to reckon with the ratio between battleships
and carriers embodied in the 1921 Washington Naval Treaty, at a time when carriers were very much exper-
imental ships. New carrier construction, on a very modest scale, was embodied in the 1938 Vinson-Trammel
Act, which in turn was a reaction to Japanese aggression in China. Congress provided only enough tonnage
for two new carriers of Yorktown size, of which only one, Hornet, was ordered before war broke out in 1940.
The question is whether the naval leadership was interested in any more than that. Hornet is shown newly
completed, late in 1941. (NHHC NH 81313)

which are the chief arguments for the retention of the type.” McNamee also wanted to
keep the 10,000-ton cruiser—the fruit of the War College’s analysis a few years before.
“The keenness of our rivals to abolish it should alone be enough to arouse our suspicions.”
McNamee warned that the British and Japanese might collude; some British states-
men had written that in the event of American withdrawal from the Philippines (which
Congress was considering), the old Anglo-Japanese Treaty might be renewed. This was
probably the only unrealistic point he made.

146 Winning a Future War

 McNamee’s arguments seem to have appealed to the General Board. What actually
happened was complicated enough so that many of McNamee’s points were ultimately
lost. President Franklin D. Roosevelt was in a delicate position. There was enormous
public support for continued reduction in naval spending through a new treaty. However,
at the same time it was clear that (as McNamee wrote) the process of arms reduction
by example was failing, and the Japanese were building and modernizing as rapidly as
they could, to an extent U.S. naval intelligence could not know. Roosevelt’s solution was
simple. He sent Norman Davis, who had been negotiating arms-control treaties for some
time, to London with a proposal for another 20 percent cut in fleets. The President was
probably aware that no such proposal had the slightest chance of success. At the same
time, CNO Admiral Standley went to London as the U.S. representative who had the
best chance of securing agreement from the Royal Navy. It is not clear to what extent the
President was aware that the British were already moving toward rearmament in the face
of Japanese aggression in China, where they had important interests.5
When the Americans met the British for preliminary discussions, it became clear
that the British desperately needed more cruisers. In order to reduce overall tonnage,
they wanted to reduce the size of battleships and carriers, although they soon understood
that they would have to accept the larger battleships the U.S. Navy wanted. Their new
orientation showed in comments that the Admiralty was now in the ascendant (as it had
not been during the 1930 negotiations) and that although the British wanted to continue
the treaty system, they would let it die if they could not get a satisfactory one.6
From 1931 onward, a dominant “fleet faction” within the Japanese navy increasingly
demanded that the “insulting” 5:5:3 ratio favoring the two Western powers be scrapped.
When representatives of the three powers met in London in October 1934 to prepare
for the coming conference, the Japanese offered what they knew would be unacceptable
proposals, such as the abolition of “offensive” warships like carriers—the warships the
Western powers would need to defend their interests in the Far East. On 29 December
1934, the Japanese abrogated the Washington Treaty, as was their right. It was probably
clear that the treaty system was finished; to the Western powers the 5:5:3 ratio was an
essential guarantee that they could defend their interests in East Asia in the face of the
Japanese. Japanese delegates did attend the 1935 London Conference, but they withdrew
on 15 January 1936. The United States rejected parity, leaving the other Washington
Treaty powers to negotiate a new treaty. Italy withdrew to protest denunciation over her
aggression in Abyssinia (Ethiopia), leaving the United States, Britain, and France.7
Without any agreement by Japan to abide by fixed ratios of total tonnage, the Lon-
don Treaty of 1936 eliminated treaty ratios altogether, concentrating on limits on indi-
vidual ship characteristics. The United States managed to fight off the British attempt
to cut battleship size, but the British did manage to impose a new upper limit of 8,000
(rather than 10,000) tons on cruisers. That was very much against Naval War College
advice, but it was probably a necessary concession.8

Downfall 147
 The new treaty reopened the question of exactly what sort of navy the United States
needed. In 1930–31, the college’s answer was based on war gaming and on the sort of
reasoning it taught. In 1935, the General Board asked senior officers, including the new
president of the Naval War College, Rear Admiral Kalbfus, for proposals. Kalbfus was
certainly no stranger to the laboratory role of his college, having served as dDirector of
War Plans.
The shift in acceptable Naval War College advice seems evident in what Kalbfus
offered. He and CinC U.S. Fleet were asked to propose a post-treaty building program.9
Kalbfus took his time to answer; CNO’s questionnaire was issued on 1 April 1935 (with
a reminder on 6 July), but Kalbus’ reply was dated 1 October 1935. Given previous Naval
War College memos, it is striking that Kalbfus made no reference whatever to gaming
experience. He preferred to reason from basic political principles. The program had to
accord with national policy and it had to be shaped to gain public support. Neither could
be quantified. Kalbfus thought the public would probably be willing to extend the 5:5:3
ratio embodied in the earlier treaties. He emphasized his fear that anything more ag-
gressive would place the onus for starting an arms race upon the United States. This was
rather different from Admiral McNamee’s advice a year earlier. Kalbfus saw the ratio as
a guide to what was politically possible. The Navy had not even reached the limits set by
the earlier treaties. Once that was done, it could continue building to maintain the 5:5:3
ratio as other countries built beyond their treaty fleets. The main change that Kalbfus
envisaged was to use the number of ships rather than their tonnage as the standard, as
that would enable the U.S. Navy to build ships of the types it needed.
The closest Kalbfus came to using gaming material was a comment that the existing
fleet was satisfactory to fight a defensive war against Orange, but it was inadequate for
a transpacific offensive campaign. In that case, Blue would needed at least a 2:1 ratio.
Kalbfus considered it unwise, at least at the outset, to strive for that on a numerical basis.
He suggested instead that the United States should seek effective superiority on a unit-
by-unit basis. “Later on, should conditions arise to create a national sentiment in favor of
increasing the navy, the opportunity should be seized to build up to a point at which war
could, with reasonable assurance of success, be carried to Orange and Orange vital trade.”
Kalbfus recommended building three battlecruisers instead of three battleships. In
effect, he cited gaming experience, pointing to the long period of cruiser and commerce
warfare that would probably precede a decisive battle-line engagement. Without bat-
tlecruisers, the U.S. Navy had to depend on air strength to deal with the four Japanese
Kongo-class battlecruisers. “If for any reason our air forces were unable to operate in
an area into which enemy battlecruisers had projected themselves, the latter would be a
formidable menace to such light forces, convoys, and shipping of ours as might be there.”
To Kalbfus, sacrificing three new battleships was done to exchange “a tactical advantage
during a possible general engagement for strategical and tactical advantages on perhaps
numerous occasions during the pre-battle period.” If the new Blue battlecruisers were

148 Winning a Future War

It certainly seems arguable that the heavy emphasis on battleships in the late pre-war building programs came at
the expense of carriers—but it is also arguable that the existing U.S. Navy battle force was obsolescent and badly
needed replacement, whereas the existing carriers were relatively young. There does not seem to have been signif-
icant sentiment in favor of changing the balance between battleships and carriers. The Royal Navy had a much
larger carrier-building program (six ships), but all but one of its existing carriers were clearly obsolete. The main
pre-war effort began with the South Dakota–class battleships, of which Massachusetts (BB-59) is shown in January
1946. Gaming did not figure in their design, which initially called for slow ships (about 24 knots) unsuited to
carrier warfare. The ships’ speed was drastically increased only when the U.S. Navy became aware (through radio
interception of a report on trials of the battleship Mutsu) that the Japanese battle line was significantly faster. The
crash redesign worked, but its nature testifies to the lack of earlier analysis. (NHHC NH 46430)

powerful enough, they might work effectively with the battle line. The new battlecruisers
would have a speed of 30 knots; new battleships would have a speed of 21 knots, like ex-
isting U.S. battleships. If no battlecruisers were built, an increase in U.S. air forces would
be the main way to counter the enemy’s ships. That would mean more patrol planes.
A study of U.S. cruiser requirements, which has not survived in the Naval War Col-
lege document, showed a need for 53 to 60 ships— on the basis of a numerical ratio, not
a study of a particular kind of war through gaming. As of 31 December 1936, the United
States would have 37 cruisers. Orange would have the same number; the 5:3 ratio would
require 61 U.S. cruisers. If Orange scrapped over-age cruisers (an unlikely possibility), 28

Downfall 149
would be left, so the U.S. figure would be 47. Kalbfus recommended 57 as a conservative
figure. If the U.S. Navy also built battlecruisers, all 20 new cruisers could be 6-inch gun
ships. Beside the Brooklyn class (10,000 tons) the Navy was then building, there was also
interest at this time in 7,000-ton light cruisers—a type the college had explicitly ruled
out in its deliberations a few years earlier.
In calling for ten squadrons of destroyers, Kalbfus noted that two of them would
be used to screen carriers, which implies that he envisioned the carriers operating in
groups. This was certainly not the lesson of war games, in which carriers survived by
operating singly.
In the questionnaire (and in Kalbfus’s answer), carriers came after battleships/bat-
tlecruisers, cruisers, and destroyers, suggesting that they now enjoyed a low priority.
Kalbfus pointed out that Orange had a total of six carriers built and building, so in order
to maintain a 5:3 ratio, the United States needed ten. He pointed to the desirability of
having carriers re-service planes launched by battleships and cruisers, hence the potential
value of two small carriers to work with the battleships. This was certainly contrary to
previous Naval War College advice. Kalbfus tempered his advice: “[I]n view of the need
for striking power in our air force…it is considered that we cannot afford to build vessels
for servicing purpose alone within the 5:3 carrier ratio. In time of war, possibly converted
merchant vessels could be utilized for this purpose.” Kalbfus thus ended up calling for
three more big carriers like Yorktown and Enterprise. If the smaller Wasp was not laid
down until after the treaties lapsed, she should be re-ordered as a large carrier.
Kalbfus estimated that on the outbreak of war the Navy should have 12 patrol plane
squadrons (18 aircraft each) based on the needs of a step-by-step advance through the
Mandates, using these aircraft for patrol and screening. More would be needed as the
fleet advanced more deeply into the western Pacific. This might be read as based on war-
game experience, but not explicitly so.
Kalbfus’s list of desired new ships in priority order placed battlecruisers first, then
battleships, then submarines, then destroyers, light cruisers, and—finally— heavy carri-
ers. His proposed building program was two battlecruisers, two battleships, ten subma-
rines, ten destroyers, five light cruisers, one carrier, one minelaying cruiser, three tenders,
one repair ship, and one net layer (to protect an advanced base). Even this list does not
make it clear whether naval aviation was no longer considered very important. All U.S.
carriers were under-age and entirely modern. All U.S. battleships were over-age and ob-
solescent at best. Nearly all U.S. destroyers were obsolescent at best, and nearly all of the
submarines were obsolete.
The pressure to build as much carrier capacity as possible certainly evaporated after
1934. Whether that can be traced to the decline of the Naval War College is not certain.
Surviving records do not show intense interest in building carriers. In commenting on a
proposed ten-year building program in March 1938, the General Board mentioned that
“in view of the arising questioned value of aircraft carriers relative to increasing improve-

150 Winning a Future War

ment in characteristics of land planes, it is believed that if more carriers are desired they
should be built as soon as possible.”10 The sense may have been that a gap was opening
between the performance of land-based and carrier-based aircraft. The longer the Navy
waited for new carriers, the worse the gap and the lower the value of new carriers. How-
ever, it cannot have been clear what was going to replace sea-based aircraft. Moreover, in
1938 BuAer was actively trying to develop very high-performance carrier-based fighters
that would more than close the gap. It succeeded with the Vought F4U Corsair, which
won the 1938 fighter competition and flew in 1940.
Kalbfus’s call for battlecruisers resonated with the General Board. In 1935–36,
the U.S. Navy’s designers produced a variety of possible designs for the first two U.S.
post-treaty battleships. The General Board favored a 30-knot ship with nine 14-inch
guns (as signed, the London Treaty limited ships to that caliber). This was Kalbfus’s
battlecruiser. In addition to its ability to deal with its Japanese counterparts, it was
fast enough to work with carriers. The alternative was a slower but better-armed and
-armored ship. CNO Admiral Standley, acting as Secretary of the Navy in the Sec-
retary’s absence, rejected the fast design, adopting instead a 27-knot ship armed with
twelve 14-inch guns. It was built as the North Carolina (the 14-inch guns were replaced
by 16-inchers when the Japanese refused to agree to limit themselves to 14-inch guns).
The really fast battleship returned only in 1938 when the Iowas were designed, and they
combined speed, firepower, and armor in a package that was impossible on the more
restricted displacement available in 1936.11
As the world situation worsened dramatically from 1937 onward, U.S. Navy new
construction policy seems to have been less and less connected to the lessons of campaign
games. Comparisons of overall strength, generally in numerical rather than tonnage
terms, seem to have dominated. In 1938, faced with Japanese aggression in the Far East
and German aggression in Europe, Congress passed the second Vinson Act. The 1938
act simply increased the size of the Navy by 20 percent. The increase in carrier tonnage
was slightly greater, from 135,000 to 175,000 tons.12
The character of the Vinson Act and the absence of detailed reasoning leading up to
it suggest how badly the influence of the Naval War College had declined. To continue to
use the tonnages set at Washington in 1921 to set relative battleship and carrier strengths
was to imagine that almost nothing had happened in the interim to naval warfare, despite
rather dramatic improvements in carrier air capability. Neither of the other major navies
was following suit; both were building new carriers in considerable numbers.13 It might
be argued that the absence of the Naval War College laboratory showed in U.S. reluc-
tance to order more carriers.
To complicate interpretation, the shift in strategy (due to Naval War College analy-
sis based on gaming) changed the basis for air requirements. Until 1933, it was assumed
that the fleet would fight its way through the Mandates to reach the Philippines as soon
as possible. Once it reached the Philippines, it would fight the decisive engagement that

Downfall 151
would open Japan to blockade. En route it would face large numbers of Japanese aircraft,
which the treaties did not limit. Campaign games included enormous air battles, with a
high rate of wastage. It made perfect sense for Admiral Pratt to say that the fleet needed
vast numbers of aircraft in order to survive long enough to fight the fleet battle at the
end. Gaming data also showed that the direct contribution of aircraft to that decisive
battle would be limited, because it was difficult for aircraft to damage enemy battleships.
The aircraft could, however, strip the Japanese fleet of its other ships, adding to the U.S.
advantage in a more powerful battle line (air spotting would also make U.S. gunnery far
more effective).
It was one thing to defend the fleet as it passed through the Mandates en route to
the western Pacific. It was quite another to move into the Mandates island by island. The
Japanese would certainly defend their islands. Their aircraft would be able to hop from
base to base to concentrate aircraft: That was exactly how they tried to defend Saipan in
June 1944. However, once an atoll had been seized it could support both an air defense
force and long-range seaplanes capable of scouting and also—importantly—of bombing.
War games incorporated such aircraft (on both sides) as long-range bombers. Whether
these games reflected BuAer’s vision of seaplane power or were trying out a new tactic is
not clear. There was apparently no direct attempt to determine how effective a seaplane
force could be compared to the carrier force, or even how roles should be apportioned
between them.
The new step-by-step strategy also made scouting more important. A fleet thrusting
through the Mandates needed air and surface scouts to ascertain whether an attack on it
was imminent. A fleet seizing one island after another needed to know much more about
the position of enemy forces. Among other things, it had to decide which atolls were
most vulnerable to attack, and which the enemy might find it easier to reinforce. To the
extent that an advance through the Mandates might draw out the enemy’s fleet—again,
as it did at Saipan in 1944—it was necessary to know where that fleet was, even when it
was nowhere near the advancing U.S. forces. Step-by-step also emphasized the ability of
the seaplanes to relocate as the fleet seized successive island bases, since seaplanes could
operate from any sheltered body of water, such as the lagoon of an atoll. To do that, the
fleet needed far more numerous seaplane tenders.
In a world of treaties, the greatest virtue of a seaplane force was that it was entirely
unrestricted by treaty. There were several attempts to limit the total number of mili-
tary aircraft, but they all foundered on the reality that civilian aircraft, which could not
possibly be limited, could be converted for military use. As was previously noted, when
the war gamers at Newport wanted to test the idea of a long-range seaplane scout, they
provided commercial Dornier Do X flying boats, the largest of their time, to one fleet or
the other.14
The major change in the early 1930s was that seaplane performance improved to
the point where, for an interval, it seemed that seaplanes might be effective long-range

152 Winning a Future War

bombers. By that time, long-range seaplanes were already fitted to carry 1,000-pound
bombs, but performance was less than impressive.15 In 1933, however, BuAer decided
to take advantage of evolving aircraft technology to seek a combination of longer range
and higher performance.16 What BuAer got was so impressive that the resulting airplane
was designated a patrol bomber (VPB) rather than simply a patrol plane (VP): It was the
Consolidated Catalina (PBY).17 At least on paper, it was nearly as good as a land-based
bomber. In reality, big seaplanes could not compete with land-based fighters, but they
were effective in war games.
At least in theory, using seaplanes in a step-by-step advance would much reduce the
need for carrier aircraft. Instead of having to weather continuous attacks from the Jap-
anese-held atolls, the fleet would seize them one by one. Heavy seaplane bombers based
on atolls could neutralize air bases on other islands. Bases not seized outright could be
gassed repeatedly to make their use impossible.18 As noted previously, such neutraliza-
tion was a staple of war games.
The step-by-step approach changed the role of Marine Corps aviation. In 1938, the
General Board was asked to estimate what the Marines would need to seize island bases.
Marine aviators were expected to specialize in the support of Marines on the ground, so
it was assumed that an assault on an atoll would be led by Marine aircraft. That opened
the question of exactly how these aircraft were to reach the target island. The three pos-
sibilities were to place them onboard fleet carriers, or to carry them onboard freighters or
possibly merchant ships converted into auxiliary carriers, or to use aircraft with sufficient
range to reach the target area from land bases. The third was dismissed out of hand. Since
no converted merchant ships would be immediately available when war broke out, the
second entailed assembling knocked-down aircraft one by one and launching them from
the water, far too cumbersome a possibility. That left the fleet carriers. Marine aircraft
could go onboard each of the existing four large carriers as a fifth squadron. Nothing
came of this project; Marines found themselves supported by Navy aircraft in the initial
Pacific landings. Not until 1944–45 was there serious planning for carriers embarking
Marine aircraft to provide close-air support.19 It is striking that the step-by-step strategy
did not give rise to any pre-war plans for dedicated Marine Corps aviation carriers.
What does this say about war gaming and its influence? The step-by-step strategy
certainly reflected the effect of war gaming. It is possible that estimates of the level of
air support the landing Marines needed resulted from gaming. However, they were ex-
pressed in terms of yardage to be covered, which suggests that they came from map exer-
cises and staff estimates. Certainly no discussion of the Marine aviation issue mentioned
gaming evidence.
All of this confuses any interpretation of the decline of the War College as laborato-
ry. The radical changes in overall strategy and in the role of seaplanes make it impossible
to say that it was connected with distaste for naval aviation compared to, say, conventional
capital ships. On the other hand, there seems to have been considerable ambiguity about

Downfall 153
Anyone arguing that the War College declined because it was too air-minded must deal with the reality that
so much was invested in seaplanes and their tenders. Seaplane tenders had the significant political advantage
in that they were not limited by treaty or in any other way, yet they could support very significant offensive air
capability (which was, however, probably much overrated at the time). Many of the tenders, moreover, were
sophisticated ships with significant protective armor. This is Albemarle (AV-5), photographed on 30 December
1943 in the Atlantic, with a PBY Catalina on her deck aft. (National Archives 80-G-450247)

the role of the seaplanes. When the General Board held hearings on their characteristics,
all its witnesses pointed to scouting as their primary role.20 Yet BuAer often pointed to
the value of sea-based bombers. In 1935, it issued a specification for a four-engine flying
boat with a substantially larger bomb capacity.21 In 1938, it asked for much larger flying
boats, ultimately selecting the Martin PB2M Mars. They, too, were described as bombers
as much as scouts.22
The changing role of the seaplanes—and the step-by-step strategy—made it import-
ant for the fleet to move seaplanes forward as it advanced through the Mandates. In 1934,
the seaplanes were part of the base force at Pearl Harbor. They had mobile tenders, but
mainly to support them once the fleet reached the Philippines. There would hardly be time
for the seaplanes to deploy and act as the fleet pursued the “through ticket to Manila.” In
September 1937, however, the patrol planes were transferred to the Scouting Force.23
If indeed the seaplanes could deal with Japanese land-based air forces in the Man-
dates, then the role of the carriers would change dramatically. Instead of having to beat

154 Winning a Future War

off nearly constant enemy air attacks, they would deal mainly with the enemy’s carrier
force, either in the Mandates or in the climactic fleet action. The sort of unlimited carrier
air arm envisaged in the 1920s might not be needed. This argument seems not to have
been made explicitly, at least in surviving documents, but it may help to explain the shift
in policy that roughly coincided with the downfall of gaming as laboratory. This expla-
nation is an alternative to a simpler one: that with the death of Admiral Moffett, naval
aviation no longer had a powerful enough advocate forcing the whole fleet to accept the
somewhat futuristic view of naval aviation associated with the Naval War College.
If the carrier story is open to alternative interpretations, pre-war U.S. cruiser de-
velopment seems to show the collapse of Naval War College gaming input much more
clearly. In the fall of 1934, with the 1930 London Treaty still in force, the question of
replacements for the first two of the ten-year-old 7,500-ton Omaha-class cruisers came
up. About 20,000 tons was available, counting unobligated tonnage, the tonnage of the
two old cruisers, and savings on various ships. The Bureau of Construction and Repair
produced a variety of sketch designs for ships displacing 7,000 to 10,000 tons. The only
agency that could compare their fighting power was the Naval War College, using the
techniques it had developed to rate various ships. Its data showed roughly what it had
argued a few years earlier: The best bargain was a pair of 10,000-ton ships (the largest
possible). The result was two modified Brooklyn-class cruisers, Helena and St. Louis.24
Any decision as to replacement of the eight remaining Omaha-class cruisers was deferred
pending the outcome of the London Naval Conference.
Despite the General Board’s endorsement of the college’s recommendation that
any new cruisers be the largest possible, there was fleet sentiment favoring a few smaller
cruisers for battle-line work, presumably meaning to beat off enemy destroyer attacks
and to stiffen U.S. destroyer attacks. Thus, in March 1935, OPNAV asked the design-
ers for studies of smaller cruisers. In June 1935, they submitted sketches of 5,000- and
7,000-tonners.
At about the same time Commander, Cruisers Scouting Force (Rear Admiral Thom-
as C. Hart) argued that the major Naval War College proposal that torpedoes be omitted
from cruisers be overturned.25 He pointed out that all other navies armed their cruisers
with this weapon; “if this is a case wherein we claim all the other navies to be wrong and
we only to be correct, it would seem that for our own purposes our argument needs to
be very strong.” He pointed to the weakness of the Naval War College argument. It was
based on experience of major fleet actions (as gamed). It was well understood by this time
(also from gaming) that a major action would be the culmination of a campaign, in which
cruisers might well find themselves individually—or in small numbers—fighting enemy
ships. Torpedoes might then prove quite valuable. The college had not taken such combat
into account.
The U.S. negotiating position at London in 1935 followed the recommendation
to concentrate on large cruisers, but in the end the British managed to cut maximum

Downfall 155
cruiser size to 8,000 tons. What should the United States build after 1936? In the past,
the Naval War College would have offered a proposal based on gaming experience, or
perhaps a variety of smaller cruisers would have been tested in games. Nothing of the
sort happened. The only U.S. experience with modern cruisers in the allowed size range
was with the ten 7,500-ton Omaha-class light cruisers built under the 1916 program. In
1936, they were over-age and obsolete. It might be imagined that any new cruisers would
fill their existing roles. Some were flagships of destroyer and submarine flotillas. Others
were intended to help the battle fleet beat off enemy destroyer attacks, a role being taken
over by the larger Brooklyn class, whose more numerous guns fired considerably faster.26
The Brooklyns had been conceived specifically because the Naval War College argued
that smaller cruisers were not worth building for that purpose.
The Bureau of Construction and Repair was asked to lay out a spectrum of cruiser
designs within the 8,000-ton limit.27 All of them incorporated torpedo tubes. Initial-
ly, the General Board preferred the largest possible ship, with the usual combination of
6-inch and 5-inch secondary guns. Its proposed FY38 program was three such ships. It
turned out that no satisfactory two-caliber design was possible within the specified dis-
placement. The General Board chose the next largest possibility, a 6,000-tonner armed
with 5-inch guns. It could be justified as a fleet cruiser capable of working with or against
destroyers. Four could be built on about the same displacement as the three larger ships,
hence at a similar cost. None of this was based on Naval War College analysis; it does
not even appear that the college was consulted. The FY38 cruiser was very much the best
that could be achieved on an unsatisfactory tonnage. For FY39, the Bureau of Ordnance
promised a solution employing an embryonic dual-purpose 6-inch gun. Using that weap-
on, an apparently satisfactory 8,000-ton cruiser could be designed. The FY39 program
included two such ships. Once war broke out in September 1939, the London Naval
Treaty could be considered dead. The two 8,000-ton ships were replaced by modified
Brooklyns (Cleveland class). That must have been a considerable relief to BuOrd. Its twin
dual-purpose 6-inch mounting was not ready until after World War II began. It was
incorporated in the very large Worcester-class cruiser ordered in 1943.
Another cruiser development almost certainly did not involve the college or gaming.
In March 1938, the Secretary of the Navy (presumably meaning the CNO) proposed a
new type of large cruiser armed with 10- or 12-inch guns, hence capable of overcoming
all existing cruisers.28 Initial design studies showed that it would displace about 18,000
tons or more (26,000 tons with three triple-gun 12-inch turrets), hence would be well
outside the cruiser category as defined by the London Naval Treaty. It would be defined,
then, as a capital ship. One question was whether the U.S. Navy would willingly sacrifice
battleship tonnage (as provided under the Vinson Acts) to build such ships.
There is no indication that the idea came out of gaming. The General Board cruiser
file (420-8) includes a 4 April 1938 paper strongly advocating the new type of cruis-
er, written by Captain A. J. Chantry, who was then head of Preliminary Design in the

156 Winning a Future War

Bureau of Construction and Repair.29 Chantry had been a member of the Naval War
College Senior Class of 1936. His 1938 paper is in the form of the college-advocated
“Estimate of the Situation,” but it is far less analytical than the paper the college pro-
duced in 1931 to guide future cruiser policy. Given its timing, it may have been mainly
an attempt to back the Secretary’s request rather than an independent proposal that the
United States should build a super-cruiser.30 Chantry based his estimate of what the
Japanese were likely to build on
the national character of her people; the aspirations and aims encompassed
in her national policy; geographic factors; and in particular, her present
and future economic position…. In considering Japanese character it is
perhaps sufficient for present purposes to note that her people are zealots,
almost fanatical in type, and only removed in moderate degree from the
influences of feudalism. Nationalism is highly developed backed by de-
termination of the strongest character. The Japanese believe that they are
the children of Heaven and nothing can divert them from their national
idealisms. Life is held cheaply in support of their country. They are the
most resolute of peoples, the most fearless and determined adversaries of
those who seek to oppose them. They are secretive…. It is safe to assume
that [Japan] now seeks the hegemony of all China and areas to the north
and desires to extend her influence to the East Indies and South Seas as
opportunity presents…there is almost no limit to Japanese conceptions of
Empire. Coupled with the fanatical determination of her people it makes
Japan a “problem child” in the family of nations; one whose every move
deserves the most serious attention and consideration of major powers.
Some of Chantry’s tone may be traceable to the recent shock of the Japanese attack
on China in 1937, in the course of which the U.S. Yangtze River gunboat Panay had been
attacked and sunk, apparently deliberately. The invasion and such incidents were then
leading to the beginning of U.S. mobilization, in the form of the second Vinson Act and
less public debate on whether to increase production of naval weapons and equipment.
Japan was also in the process of rejecting attempts to convince her to abide by the limits
stated in the new London Naval Treaty. However, Chantry was going far beyond the
usual cool Naval War College analysis. War College papers did not dwell on national
characteristics the way Chantry did.
Chantry went on to point to Japanese dependence on imported raw materials; Japan
had amassed sufficient supplies to last for the early stages of a war, but would have to rely
on imports anyway. None of this can have been unfamiliar. Earlier analyses of likely Japa-
nese naval construction would not even have mentioned it. A list of courses of action open
to Japan repeated that Japanese secretiveness would be a major factor in her grand strate-
gy. Japan would counter opponents’ greater strength with “cunning,” meaning that Japan

Downfall 157
would develop “weapons of attack for which her opponents have no counterpart readily
available.” This last was an accurate prediction of Japanese policy, which emphasized the
development of individually superior weapons such as the huge Yamato-class battleship,
the Zero fighter, and the Type 93 “Long Lance” oxygen-propelled torpedo.
Chantry correctly predicted that the Japanese would build warships individually su-
perior to their foreign counterparts. He parted company with the Naval War College in
not considering overall Japanese requirements, hence looking at likely Japanese priorities.
Instead, he focused on one part of a possible Pacific War, a war against trade that the
United States would prosecute. The question was how the Japanese would defend their
own trade. This was very different from the issues typically raised in gaming. The Naval
War College certainly agreed that Japan would live or die based on its ability to import
what it needed, but that led to a U.S. strategy of blockade. Games explored raiding, not
as a means of crippling Japan, but rather as a way to draw off Japanese forces to allow the
U.S. fleet greater freedom of action.31
Chantry pointed out that a Japanese convoy strategy would fail if the convoy es-
corts could not deal with individually powerful raiders (presumably cruisers). However,
the raiders would be drawn into focal areas (where shipping routes crossed) as they ran
down their victims. By placing individually powerful ships in the focal areas, the Japanese
would gradually annihilate the raiders. It would not matter if the United States built
large numbers of cruisers, as long as they were individually inferior to those the Japanese
had. It happened that the focal-area defense had been planned by the Royal Navy during
the latter part of the 19th century. It was the alternative to convoys that the British ad-
opted at the time as a means of protecting their trade. It is not clear that Chantry or
anyone outside the Admiralty knew this in 1938; the focal-area approach may merely
have been the obvious one.
Chantry cited persistent rumors that Japan was already building cruisers somewhat
larger than the 10,000-ton limit embodied in the treaties. Other countries might do the
same; Chantry cited the German pocket battleships (essentially cruisers armed with 11-
inch guns) as well as very fast light Italian cruisers. He suggested that the Japanese would
use their extra tonnage to mount heavier guns—9-, 10-, or 11-inch calibers. In fact, the
Japanese cruisers displaced considerably more than 10,000 tons. No one knew that until
the end of the Pacific War. The Japanese used the extra tonnage to gain conventional
advantages like greater speed, but the ships did not mount heavier guns.
Chantry pointed out that having freed themselves of treaty restrictions, the Japa-
nese might choose to build somewhat larger and better-protected ships armed with more
8-inch guns . Alternately, they might jump to a super-cruiser type armed with much more
powerful guns. Japanese secrecy made it impossible to know whether super-cruisers were
planned. Chantry argued that the possibility could not be dismissed. Any of these ships
might indeed be extremely expensive, but the Japanese might well build them anyway.
Moreover, the Japanese had a national tendency to spring surprises on their opponents.

158 Winning a Future War

Chantry argued that the United States should anticipate the possible construction of Jap-
anese or other super-cruisers by building competing ships, the type the General Board
was considering. The U.S. Navy needed something superior to either the super 8-inch gun
cruiser or the semi-capital ship, since it could not be sure what the Japanese were building.
All of this was a very non–Naval War College argument dressed in Naval War Col-
lege clothing. Chantry never asked the usual Naval War College questions. What were
overall Japanese requirements? What were their priorities likely to be, given likely U.S.
action in wartime? There is no evidence that the college was ever asked for its views, or for
that matter whether it ever volunteered any.32 For the moment, enthusiasm for big over-
armed cruisers waned, because it seemed more important to use available tonnage for real
battleships (the Iowas were being designed). However, the super-cruiser idea returned
in 1940, when the Two-Ocean Navy Act of July 1940 dramatically increased available
tonnage and all restrictions associated with the London Naval Treaty had died. The big
new cruiser program included the six Alaskas, which certainly filled Chantry’s require-
ments.33 As for the Japanese, they seem to have decided to build their own super-cruisers
only in 1941, and then as part of their main fleet (they were to replace the Kongo-class
battlecruisers).34 These ships were never laid down.
In 1938, the General Board was considering a larger cruiser program. The individual
tonnage limits embodied in the London Naval Treaty could be modified by invoking the
escalation clause: If one of the signatories of the Washington Treaty (i.e., Japan) refused
to accept the limitations set at London, the others could modify the London Naval Trea-
ty. That actually happened with battleships (the upper limit was changed from 35,000 to
45,000 tons). A 26 May 1938 General Board memo suggested that it should also be ap-
plied to cruisers.35“The employment of cruisers in war, particularly against Orange, leads
to the assumption that we require a considerable number for assignment to trade routes,
convoys, and independent duties in general.” That implied a need for large rather than
small cruisers, and probably for 8-inch gun cruisers, the construction of additional units
having been banned by the London Naval Treaty. Moreover, the new triple 6-inch gun
in the Brooklyns had not yet been tested. The U.S. government did not raise the cruiser
issue when the escalation clause was invoked to permit construction of the 45,000-ton
battleships. Surviving documentation gives no indication of why, but it seems likely that
escalation was limited because there was still such strong anti-military sentiment.
Again, this possibility invited the sort of analysis that in the past had been done by
the Naval War College based on gaming experience. The 1938 General Board memo is
concerned with a simple numerical comparison of U.S. and foreign cruiser fleets, and
with possibilities within the tonnage available under the Vinson Act following the four
6,000-ton ships of the 1939–40 (FY38) program. Size therefore had to be traded off
against numbers, which was exactly the sort of consideration the college had taken into
account in 1930–31. This time, it was handled as a simple list of alternative possibilities,
without even the college’s comparison of fighting value.

Downfall 159
. C: G V
R   P

G aming had at least three possible functions. One was to explore possible wartime
situations in ways full-scale exercises could not. Military judgment based on expe-
rience could often foresee outcomes, but not when entirely new technology was involved.
This was the laboratory function performed by the Naval War College. The one entirely
new technology of the inter-war period was naval aviation. Gaming provided the U.S.
Navy with an invaluable sense of how that technology would affect war not only on the
tactical, but also on the operational and strategic levels. For students at the college, the
games offered a glimpse of what a future naval war would be like. No full-scale exercise
could have done the same.
A second function was to teach students how to fight. The Naval War College em-
phasized its careful rational way of deciding what to do. Not long ago, the Naval War
College placed Admiral Nimitz’s “Gray Book” (in effect, his command diary of the Pa-
cific War) online. The “Gray Book” can be considered an extended exercise in the “Esti-
mate of the Situation,” the decision-making basis that the college taught Admiral Nimitz
when he was a student.
In January 1942, a British officer criticized American naval officers for tending to
estimate the odds before making decisions.1 In effect, he was comparing what Newport
taught with what his own Tactical School taught. The lesson the British derived from
their World War I experience was that they should have shown far more aggressiveness
and initiative. It can be argued that this was an illusory lesson; the massive World War
I Grand Fleet was like an army, in which excessive initiative could have had devastat-
ing consequences. In any case, the Royal Navy’s tactical school awarded points for ag-
gressiveness, even when the decisions taken proved unfortunate.2 In effect, that was a

160 Winning a Future War

judgment on Britain’s enemies, that they could be bluffed because they were not nearly
as professional as the Royal Navy. That judgment proved quite correct when the Royal
Navy fought the Germans and the Italians. The U.S. Naval War College based its course
on the idea that the enemy would be as professional as its own students. The students
had to be taught to out-think a fully witting enemy. In comparing the U.S. and British
approaches to gaming and to decision making, it would be interesting to ask whether the
British approach would have been as successful against the Japanese.
Perhaps the greatest gap in Naval War College simulations of combat was the tempo
of an actual battle. The complexity of the rules made it impossible for move to follow
move fast enough to give students a sense of the confusion generated in an actual battle.
There certainly was an attempt to simulate the “fog of war,” but instructors pointed out
again and again that in games students had far more time to make decisions than they
would have in reality. In theory, full-scale exercises would have filled some of this gap, but
in reality they proved insufficient. That seems to have been particularly the case in the
night battles in the Solomons in 1942–43.
A third gaming function was to understand or even predict the behavior of foreign
powers. That was the most difficult of all, because it required that some players simu-
late alien ways of thinking. Could the U.S. government have guessed that, under intense
pressure in 1941, the Japanese would have chosen to fight even though their own analysis
showed that they would lose a Pacific War? Much the same might be asked about the
Japanese reaction to disaster. After the war, many Japanese said that they knew the war
was lost when the Americans took Saipan in June 1944. Yet it took Japan more than
another year to accept defeat. Would gaming have provided insight into what happened?
Such questions were outside the purview of the Naval War College. However, since 1945,
there have been attempts to game foreign governments’ thinking, generally at the level of
the State Department.
How well did the Naval War College envision the Pacific War? After the war, Admi-
ral Nimitz famously said that in the course of the games, at one time or another, someone
or other at the college experienced everything that happened in the Pacific, other than
the kamikaze warfare at the end of the war. The Naval War College archives of the time
included material on all the games, and at least some students must have taken advantage
of them. Moreover, the OPNAV War Plans Division, which actually planned the war,
obtained copies of much of the gaming material. Some of the game material survives only
in the War Plans Division files.
Did the Naval War College predict the course and character of the Pacific War in
detail? Of course not. On the technical level, it never predicted that carriers would dis-
place battleships as capital ships. Its estimates of weapon effectiveness showed that it
would still take either battleship firepower or torpedoes (typically launched by destroyers
or submarines and much less often by aircraft) to destroy an enemy’s battleships. That
was not entirely misleading. The number of carrier aircraft required to sink the huge

Conclusion: Games Versus Reality in the Pacific 161

War gaming prepared officers to wield carrier airpower, even if they were not aviators. Both commanders
at Midway, for example, were surface officers, and one of them, Admiral Raymond F. Spruance, went on to
command the fast carrier task force. Spruance had been a student and then an instructor at the War College.
He later testified to the value of war gaming in his own professional development. He is shown here as Com-
mander, Central Pacific Force, U.S. Pacific Fleet, 23 April 1944. (National Archives 80-G-225341)

162 Winning a Future War

Japanese Musashi in October 1944 were entirely unimaginable in the pre-war period.
Even this attack did not sink or even seriously damage other Japanese battleships in the
same formation.3 They survived to attack the U.S. escort carriers covering the Leyte Gulf
landing. However, the college did make it clear that carrier air power would be extremely
important, and it did show that it would be essential in assaulting Japanese-held islands
during the fleet’s step-by-step advance to the West.
On the strategic level, the college did not envisage the type of coalition war that
would lead the United States to defend Australia and thus to fight on Guadalcanal and in
the Solomons. These places hardly figured in pre-war thinking. The coalition games en-
visaged by Captain van Auken were never played. The college did play coalition games in
1941, when U.S. ships were already working with the Royal Navy to help protect convoys
and also to help deal with German and other Axis fleets. However, these games did not
touch on the real problems of such operations, such as differences in operating practices
and problems of potentially split command.
Similarly, no one envisioned the dramatic German success in Europe that opened
the Far East colonies to Japanese attack in 1941. No Naval War College game foresaw a
situation in which the United States would have to divide the fleet between Atlantic and
Pacific to deal with simultaneous threats in both oceans. However, there was at least one
War Plans Division study of a U.S. strategy to fight a Red-Orange coalition, which would
have had exactly that effect. In such a war, the Atlantic threat (Red) would constrain
what the United States could do against Orange, and vice versa. Moreover, the prefac-
es of the series of major Red-Blue games mentioned the possibility that Orange would
exploit U.S. weakness due to the fight against Red. That might be considered broadly
analogous to Japanese exploitation of the defeat of the European colonial powers in 1940,
and to the consequent weakening of the British. In 1938, the War Plans Division studied
a situation in which the U.S. fleet would move to Singapore and help protect the British
Far East empire against the Japanese while the British faced the Germans in Europe, but
that operation was not, as far as it seems, gamed. The War Plans Division study seems to
have been carried out in the context of secret staff talks between U.S. and British officers
in 1938–39.4
The various sorts of battles that the U.S. Navy fought in the Pacific certainly did
figure in the mass of games. They included not only the battle-line engagements that
might be associated with the pre-war Navy, but also many carrier battles, and even night
cruiser battles like the ones fought in the Solomons. The run of war games does read like
an encyclopedia of possible kinds of future warfare in the Pacific.
Most important, throughout the inter-war period the games educated the U.S. Navy
in the problems of a war projected across the vastness of the Pacific. No navy had ever
fought successfully so far from home. The only precedent, in the inter-war period, was
the disastrous Russian performance in the Russo-Japanese War of 1904–1905, which
culminated in the destruction of the Russian Baltic fleet at Tsushima. In that war, the

Conclusion: Games Versus Reality in the Pacific 163

Probably the greatest contribution the Naval War College made to victory in World War II was its demon-
stration that the strategy of steaming directly to the Philippines was bankrupt, even before the destruction
of the battle line at Pearl Harbor. This, in turn, led to the development of the amphibious force that was so
successful in both the Atlantic and the Pacific during World War II. Here, amphibious tractors (Amtracs) pass
the battleship Tennessee (BB-43), which is shelling Okinawa, 1 April 1945. (NHHC NH 42390)

problem for the Russians had been logistical—merely getting to the Far East in condition
to fight. After World War I, the U.S. Navy would face both the logistics and whatever
attacks the Japanese could mount as it passed through the island chains they had gained
as a result of the war. Even the logistics would be daunting. In 1919, when the U.S. Navy
first seriously considered a Far East war, most of the U.S. fleet burned coal. Ton for ton,
coal did not provide enough energy to move the fleet across the Pacific. Replenishment
was difficult at best. A U.S. naval war could not really be fought in the Far East until the
fleet had converted to oil fuel. That process began just before World War I, but it was very
much a post–World War I development. Oil fuel made it possible to project naval forces
and whatever they might convoy across the Pacific.
The war’s possibilities and consequences had to be envisaged in the context of new
weapons: large numbers of aircraft and also long-range submarines. Gaming was a good
way to do so. Looking back, we find it difficult to appreciate just how novel a Pacific naval
war would be. After World War II, the Royal Navy circulated an analysis of U.S. per-

164 Winning a Future War

formance in the Pacific that gives a sense of what was achieved. The writer likened the
sustained U.S. fleet operation in the western Pacific, which was based at Ulithi, to a war
fought off the U.S. coast using a base on the other side of the ocean at Gibraltar.
The U.S. Navy used war gaming to explore this kind of warfare and to make it fa-
miliar to its officers. Admiral Nimitz’s remark cited above may seem to refer to islands
around and on which successive generations of War College students fought. Some stu-
dents certainly did become familiar with places they would fight years later, but that
was a minor advantage. Much more important, gaming familiarized students with the
impact of the new aviation technology, new kinds of command and control, fleet subma-
rine operations, and also (from 1935 onward) with concepts such as amphibious assault
against defended islands. Like aviation, amphibious warfare was a new and nearly unex-
plored branch of naval warfare. For the U.S. Navy and the Marine Corps, it had little
significance until war gaming revealed that it would be essential as part of a step-by-step
advance across the Pacific.
Gaming forced officers to define and then to confront the most likely Japanese
strategy, which was to impose attrition as the U.S. fleet moved west. Exploration meant
working out what the Japanese would have to do to execute various alternative strate-
gies. To reach that conclusion, the Naval War College used games that explored a wide
variety of possible Japanese strategies. The extent of that exploration made it reasonable
to assume that the Japanese would follow the expected attrition strategy. This type of
exploration was necessary because other kinds of evidence of Japanese thinking were so
limited, thanks to tight Japanese security. The U.S. Navy monitored Japanese fleet exer-
cises—which turned out to reflect Japanese war planning—using broken Japanese codes
and also traffic analysis. However, the fruit of such communications intelligence was very
closely held, and it could not be used to shape war games at Newport.
The inter-war games embodied a particular approach to naval strategy, which the
gamers correctly understood applied to both the U.S. and Imperial Japanese Navies. It
emphasized the role of the main or battle fleet as a shield behind which other naval ac-
tivities could be carried out. That is why both navies looked toward a decisive fleet en-
gagement, which would determine whether either could freely use the sea. For example,
a convoy approach to shipping protection was viable only under the cover of a dominant
fleet. Freed of that dominance, the enemy would be able to destroy escorts and convoys.5
Through the inter-war period, the Naval War College—and indeed the U.S. Navy—
and the Japanese naval leadership considered battleships the deciding factor in battle.
The disparity in numbers of battleships enforced by the inter-war arms control treaties
made it likely that the Japanese would adopt an attrition strategy.6 They spent heavi-
ly on improving their ships, but their prospects were bleak. Until 1936, nine Japanese
battleships faced 15 U.S. ships. If the battleships were of roughly equal combat value,
the relative power of the two battle fleets would be proportional to the square of their
numbers: a ratio of 81:225 against Japan. Even if the three oldest U.S. battleships were

Conclusion: Games Versus Reality in the Pacific 165

The straight-through strategy was impossible because the fleet had no forward repair facilities; if it suffered
underwater damage (as was likely), it had nowhere to go beyond Hawaii. Once war began, a variety of porta-
ble base units was created, most spectacularly floating dry docks that were towed to combat zones. Here, the
cruiser Canberra (CA-70) and the destroyers Claxton (DD-571) and Killen (DD-593) share repair space in the
floating dry dock ABSD-2 (Advanced Base Sectional Dock) at Manus in the Admiralty Islands, 2 December
1944. (National Archives 80-G-304096)

not counted, the Japanese would face an 81:144 ratio. The war college understood that
their strategy would concentrate on changing those odds as the U.S. fleet headed west
through the Mandates. Inevitably, that meant using weapons not limited by treaty, such
as land-based aircraft. That, in turn, inevitably made U.S. naval aircraft more important
than they might otherwise have been.
The Japanese attack on Pearl Harbor can be seen as the ultimate attrition operation.7
In 1941, the U.S. Navy had nine battleships in the Pacific, three of the best modernized
ships having been transferred to the Atlantic. It also had two new battleships, which were
not yet entirely ready (North Carolina and Washington ). Once war broke out, the United
States would presumably have transferred all of its best ships to the Pacific, giving the
U.S. fleet there a total of 14 battleships compared to ten in the Imperial Japanese Navy,
with two more nearly ready. When Admiral Yamamoto Isoroku planned the attack on

166 Winning a Future War

Pearl Harbor, he wrote that his primary objective was to destroy at least four U.S. bat-
tleships. That would have eliminated any U.S. superiority, at least in numerical terms.
Yamamoto also said that he could “run wild for six months,” which would be roughly the
period before additional U.S. battleships already under construction could enter service.
Yamamoto was willing to sacrifice the carriers in the attack force to gain this level of attri-
tion. Even more than the Americans, the Japanese naval high command did not consider
carrier aircraft an essential arm in a battleship-on-battleship fight.8
The Pearl Harbor attack itself did not figure in any surviving War College game.
The gaming system based on estimates of the situation was in effect rigged against any
operation as risky as the Japanese attack on Pearl Harbor. Under the game rules, had the
defenses been alerted, the Japanese might well have suffered heavy losses. The Japanese
themselves thought it entirely possible that they would lose several carriers. The technique
taught by the War College forced players to balance possible gains against probable losses.
It did not take national psychology into effect. There is some evidence, for example, that
Yamamoto thought that the U.S. public would be so shocked by the loss of battleships at
Pearl Harbor that it would shrink from fighting a protracted war. Others in the Japanese
leadership took the opposite position: that the attack would enrage the U.S. public exactly
as it did. It is not clear why no one, Japanese or American, considered the sudden attack on
U.S. forces in the Philippines, which the Japanese considered an inescapable requirement,
would not equally enrage the U.S. public. The gap in Japanese perception of U.S. national
psychology might be matched by the inability of U.S. gamers to estimate the effects on the
Japanese of a U.S. carrier air attack on Tokyo, a feature of one of the games. One Naval
War College game did include a surprise submarine attack on the battle fleet at San Diego,
which reduced U.S. superiority over the Japanese battle fleet.
The games tended to focus participants’ attention on the way each side was likely to
shape its strategy to use its inherent advantages to offset its inherent disadvantages. For
the United States, the advantages were greater capital ship strength at the outset, thanks
to the ratios embodied in the inter-war treaties, plus a much larger industrial base. To
some extent, the advantages inherent in the treaties were offset by U.S. unwillingness to
build up to the allowed limit. In most types of ships in 1941 the Japanese enjoyed much
better ratios of strength than they might have. This disadvantage was alleviated after
1942. Since 1938, the United States had been building up under the second Vinson Act.
The great U.S. disadvantage was a combination of distance and the lack of secure bases—
and, even more important, repair facilities—in the Far East.
For Japan, the most important inherent advantages were distance from the Unit-
ed States and the geography of the Mandates in the Pacific. The disadvantages were
industrial. Game designers might doubt that the United States could instantly build
large numbers of new ships, but notes for participants often pointed to the much greater
strength of the U.S. aircraft industry. We now know that the Japanese were well aware
that they could not match U.S. numbers, and that they hoped to develop special weap-

Conclusion: Games Versus Reality in the Pacific 167

ons of various sorts as equalizers: To this extent, Chantry was right. The Naval War
College never seems to have envisioned this approach, even during its analysis of the
cruiser issue in 1930–31. Note that Chantry used Naval War College methodology to
reach his conclusion.
The Naval War College fought its games using existing technology, although it was
sometimes criticized on the ground that its view of air operations was on the optimistic
side. This argument was made particularly strongly once champions of air power like
Moffett and Pratt were gone. Thus, nothing in gaming reflected the vast jump in aircraft
performance after 1941, which was achieved largely due to the advent of a new generation
of engines. It can be argued that improvements in aircraft on both sides more or less can-
celled out. Air weapons used against ships and ground installations did not change nearly
as much. Thus, game rules involving dive bombing and air-launched torpedoes seem to
have been good predictors of wartime weapon performance.
The Naval War College did not predict the impact of electronics on the coming war.
The U.S. Navy began experimenting with radar in 1939, and became aware of British
experience using radar to control fighters in 1940. Neither development seems to have
affected gaming. Nor does either seem to have penetrated through to many U.S. Navy
decision makers like, for example, those on the General Board concerned with future
building programs. For carriers, the central fact of World War II was that efficient fight-
er control cancelled the pre-war assumption that, once found by enemy aircraft, a carrier
was probably doomed, or at least destined to be put out of action. A vital secondary re-
ality was the advent of navigational beacons that made it practicable for carrier aircraft
to operate much further from their ships. Thus, gaming as practiced at the Naval War
College between wars could not have predicted an air-to-air victory like that won in the
“Turkey Shoot” of the Battle of the Philippine Sea. Against that, at least one game (“Siar-
go”/Surigao) did include successful defense of a fleet against enemy carrier aircraft.
Some games carried important strategic lessons. For example, in several games
played in the 1930s, U.S. forces raided Japanese commerce, not so much because they
could inflict decisive damage, as because they could compel the limited Japanese fleet
to divert forces to shipping protection and thus not to deal with a simultaneous U.S.
fleet operation. This logic applied in 1942–43 when a U.S. submarine offensive against
Japanese trade began. The Japanese lacked the surplus naval strength to set up a convoy
system while maintaining their fleets. Eventually they were able to build escorts, but to
have done so before the war would have been to accept what they considered excessive
weakness in their main, shielding, fleet. The failure of the Japanese to protect shipping
is often attributed to cultural factors and an inability to see the wider naval picture, but
the pre-war games show that it was inherent in the logic of a Pacific war fought against an
adversary with limited industrial strength.
Overall, the games provided a correct assessment of a naval strategy the Japanese
would feel compelled to adopt—not because of some quirk in Japanese culture, but be-

168 Winning a Future War

The pre-war War College did not analyze a submarine campaign against Japan in any detail, but it did show,
in a series of games, that anything except unrestricted submarine warfare (which at the time was considered
entirely objectionable) would not be successful. Once war broke out, unrestricted submarine warfare against
Japan was authorized. Initially hampered by torpedo problems, it proved extremely successful. Barb (SS-220),
which made celebrated attacks in Japanese home waters, is shown in San Francisco Bay near Mare Island, 3
May 1945. (National Archives 19-N-83952)

cause of the logic of sea power that applied to both sides. The strength of the gaming sys-
tem, as it was used at Newport, was that it avoided any reference to cultural peculiarities
on either side. It was a clear-eyed means of understanding the logic of the situation. The
rationale was that any additional cultural factor would reduce effectiveness on one side or
the other. For the historian, what is remarkable is that so much that generally considered
culturally determined in wartime Japanese naval thinking can be attributed instead to
the logic of the situation. That is evident in the record of the war games; without this
record it would probably be missed. Conversely, understanding the logic of the situation
was a great advantage for a U.S. Navy, which might otherwise have misunderstood what
the Japanese were doing.
For example, Japanese submarine commanders generally avoided attacking mer-
chant ships and naval auxiliaries. They considered that their quarry was U.S. capital
ships. The post-war evaluation was that, infused by the spirit of the samurai, these com-

Conclusion: Games Versus Reality in the Pacific 169

manders considered attacking anything short of a battleship or carrier beneath them.
However, the logic of the situation was compelling. Japan could afford considerable losses
to merchant ships as long as the shield represented by the main Japanese fleet survived.
When the decisive battle was fought, the degree of attrition imposed on U.S. capital
ships would determine the outcome. Conversely, unless the U.S. main fleet was worn
down, the smaller Japanese fleet was doomed according to accepted theories of combat
power. It made good sense to limit risks to valuable submarines (which could not easily
be replaced, given the weakness of Japanese industry) to attacks that were likely to be
worth the trouble.
The U.S. Navy had no such priority, because it already enjoyed such crushing supe-
riority in capital ship numbers that it did not have to attrite Japanese capital ships before
the decisive battle. Sinking those capital ships would be helpful, but it did not carry near-
ly the same significance as it did for the Japanese. That was true even after Pearl Harbor,
because by 1941 the United States was building so many new battleships that it would
soon be able to overwhelm any Japanese force. It turned out that battleships were not
decisive to anything remotely approaching the extent that anyone imagined in 1941—but
the war games certainly did emphasize the considerations that actually affected the Jap-
anese. Moreover, pre-war games did show that a campaign against Japanese trade could
bleed the main Japanese fleet by forcing it to detach ships and valuable personnel and,
it turned out, scarce aircraft to protect merchant ships. That campaign might or might
not be decisive, but its diversionary impact would be considerable. Pre-war gaming had
suggested as much. As it happened, the U.S. submarine campaign was devastating, but it
would have been worthwhile even had that not been the case.
The correctly understood Japanese strategy in turn defined a particular kind of air-
sea warfare, which shaped the development of U.S. naval aviation in the inter-war period.
If that seems grandiose, it is apparent when U.S. naval aviation is compared to that devel-
oped by the Royal Navy. The British invented carrier aviation, and during the inter-war
period they invested heavily in it. Unlike the U.S. Navy, they associated it mainly with
a decisive naval battle. They therefore concentrated on finding their enemy first and at-
tacking first. There was no sense that a fleet would have to fight a lengthy series of air-sea
battles before it could even get to fight an enemy battle fleet. Gaming showed Americans
that this would hardly be the case. Simply to get to the Far East and a hoped-for decisive
battle, the U.S. fleet would have to beat off repeated heavy air attacks, many of them
mounted by shore-based aircraft. This latter point mattered. The inter-war Royal Navy
was convinced, incorrectly, that aircraft developed to operate from carriers would neces-
sarily be inferior to their land-based counterparts. That was acceptable as long as carriers
fought only other carriers. It was entirely unacceptable if carriers had to beat off repeated
attacks mounted from land bases. Gaming forced U.S. officers to confront exactly that
situation, and therefore to demand high performance of their fleet aircraft. That had
enormous wartime consequences.

170 Winning a Future War

 More generally, gaming forced U.S. officers to take into account many operations that
would have to be undertaken before any hoped-for decisive fleet action. They learned, for
example, that carriers could be subject to surprise attack by surface ships. Not only did
these valuable carriers have to be screened, but their commanders had to mount constant
scouting flights to insure against the possibility of surface attack. That may seem obvi-
ous, but it was not. In June 1940, the British lost the fleet carrier HMS Glorious to two
German battlecruisers because she had not put up scouts, hence could not simply evade
the attack. U.S. escort carriers were surprised off Samar in October 1944, but their com-
mander knew how to react. He had thought through situations that included this one.
The absolute need to screen fleet carriers against surface attack was involved. The escort
carriers were uncovered largely because Admiral Halsey kept his battleships with his fleet
carriers when he steamed north to attack what turned out to be a Japanese decoy force.
Provision had been made to detach the U.S. battleships in the event of a Japanese surface
threat, but Halsey understood that his battleships were integral with his carrier force.9
The comparison with the British is natural because the Royal Navy and the U.S. Navy
were considered roughly equal as the two leading navies of their time, and because for most
of the inter-war period both developed their strategies and their tactics specifically to deal
with Japan. Both shared the idea that the ultimate stage of a war would be a strangling
blockade, and that the prerequisite for such a blockade would be a decisive battle in which
the main Japanese fleet would be destroyed. The U.S. Navy seems to have gone consider-
ably further in laying out this rationale and thinking through the whole campaign.
That was not the war that either navy actually fought. The U.S. Navy was fortunate
in that the broad scenario for its strategy forced it to contemplate the sorts of combat it
encountered. This was because it had to plan on a passage through chains of islands that
Japan controlled. The Japanese attack on Pearl Harbor eliminated the battleships that
might have sought an early decisive battle, but in 1941 the United States was building a
new fleet of fast battleships, which would enter service in 1942–43. Until they arrived,
the air-sea war that preceded any decisive fleet battle in war games was a good model for
the war the Navy actually fought.
By way of contrast, the Royal Navy envisaged a relatively unopposed passage to its
war base at Singapore. Operating in that area, its fleet would threaten vital Japanese sea
lines of communication, so the Japanese would feel compelled to fight. To some extent,
this was a version of British World War I naval strategy, the difference being that the
Japanese, unlike the World War I Germans, had a vital reason to challenge the British
fleet.10 The consequence of the British strategy was that the key part of the war plan
was logistical rather than tactical; the fleet needed sufficient supplies to get to the Far
East. En route it might be threatened by Japanese submarines, but not by constant air
attack. Unfortunately for the British, however well they analyzed the problems involved,
this analysis or gaming did not prepare them for the sort of constant air-sea attack they
encountered in European waters during World War II. There was no way that a British

Conclusion: Games Versus Reality in the Pacific 171

Nimitz could say that thinking or gaming the central war plan could prepare the navy for
what it would later experience.
It seems, moreover, that the British did not undertake the sort of cross-analysis of
possibilities the U.S. Naval War College taught. Their Far East war plan seems to have as-
sumed that the Japanese would choose to steam south to meet the British fleet somewhere
near Singapore, far from the support offered by land-based aircraft. It is not at all clear
from later accounts that the British ever asked whether the Japanese would be so cooper-
ative!11 The key assumption seems to have been that the Japanese would find intolerable
the presence of a British fleet athwart the trade routes to the south, just as the U.S. Navy
assumed that the Japanese would find American possession of the Philippines intolerable.
However, there was a vast difference. In a war between only Britain and Japan (or
Britain plus European partners against Japan), the Japanese had an alternative source
of supply across the Pacific in the Western Hemisphere. The loss of the southern routes
would be a serious blow, but it might not be all that fatal. The Japanese might choose
to keep their fleet closer to home and to other Japanese resources. They could not do
the same if they fought the Americans, because in that case their supplies really would
come from the south. The Naval War College did consider the one alternative possibili-
ty, namely that the Japanese could rely instead on supplies from the Soviet Union. War
games did include cases in which the Japanese chose not to attack the Philippines. British
accounts of the Singapore strategy do not mention any alternative plan in the event the
Japanese chose not to take the bait represented by a British fleet based at Singapore. The
main problem they considered seems to have been how to protect Singapore until the
British fleet arrived there.
The issue of bait and timing became crucial during the last pre-war years. About
1936, the First Sea Lord, Ernle Chatfield, wrote that his greatest nightmare was a war
against both the European fascist powers and Japan. His only solution would be to send
a fleet east, destroy the Japanese fleet, and then move what was left back west to confront
the enemy European powers. This strategy was barely practicable because the French
would be allied to the British, and they would more than balance the Italians. The Ger-
mans were only beginning to build up. Before 1939, the British discussed the need to
build a two-ocean navy, but it was never affordable. Moreover, the Japanese did not coop-
erate. They waited until the Germans had overrun Europe, destroying a large part of the
Royal Navy in the process, and putting the French out of action. By 1941, the British no
longer had a big fleet that could be sent east to deal with Japan before returning west to
beat down the Germans. Even had the Japanese entered the war at the same time as the
Germans, the British still faced the question of how long they could maintain a fleet at
Singapore, waiting for the Japanese to take the bait it represented.
These are exactly the sorts of issues that campaign games at Newport explored, albe-
it for the U.S. Navy rather than the Royal Navy. That they have never been raised in the
70 years since World War II seems to be a clear indication that the British never had any

172 Winning a Future War

equivalent to Newport and its gamers—and that they suffered for that lack. The appar-
ent failure of British planning illuminates the success of the Naval War College gaming
system in preparing the U.S. naval leadership for the war it would fight.
The Pacific War opened with a stunning surprise attack at Pearl Harbor and with
an expected sudden attack on the Philippines. Many pre-war officers credited the Japa-
nese with a penchant for surprise attacks. The question of how to keep such a surprise
from being decisive was a theme of many war games. The problem was how to force the
Japanese to open the war with an overt act that would not be devastating. This act in turn
would justify U.S. counterattacks that might blunt the Japanese strike. In December
1941, the U.S. Navy sought exactly such a solution by commissioning some small craft to
patrol around the Philippines. The idea was that the war would open with a Japanese at-
tack there, and that the Japanese would feel compelled to sink the patrol craft before they
could land troops. In the event, none of the small craft was sunk, and the main surprise
attack included the Pearl Harbor raid. Some games included surprise attacks (such as a
submarine attack on the battleships at San Pedro), but they were not nearly as devastat-
ing as the one against Pearl Harbor.
That omission made less difference than might be imagined, because the games gave
a sense of how a protracted Pacific campaign would be fought. The battleships sunk or
disabled at Pearl Harbor would have been essential in the fleet action that the Naval War
College (and U.S. war planners) assumed would be the culmination of a trans-Pacific cam-
paign. Before that, the Japanese would concentrate on wearing down the U.S. fleet so
that their own less powerful battle line could win. The wearing-down campaign would be
fought mainly by naval aircraft and submarines. That is why war games, particularly those
depicting a full campaign or a large part of one, so often involved huge numbers of aircraft.
The game also involved large numbers of submarines on both sides. In 1932, when
he compiled lessons learned, Naval War College President Rear Admiral Harris Laning
wrote that Orange submarines had often exerted decisive power—meaning that the U.S.
Navy should invest heavily in anti-submarine warfare. The U.S. Navy installed sonars
onboard its new destroyers. That did not seem remarkable when World War II broke
out, but it seems notable that other navies with access to this technology seem not to have
done so on anything like the same scale. The British did install sonars on a large scale,
but mainly because they were concerned with protecting trade against a future U-boat
campaign, something of much less concern to the inter-war U.S. Navy. One indication
of relative U.S. unconcern with a submarine campaign against shipping was that, un-
like the Royal Navy, it made no attempt to produce convoy escorts after the 1930 Lon-
don Naval Treaty drastically reduced its stock of older destroyers. Thus, the U.S. sonar
program was designed mainly to protect the battle fleet against the threat of Japanese
submarines based in the Mandates—against a threat evident in gaming experience. The
sonar-equipped destroyers turned out to be very useful during the Battle of the Atlantic,
a problem not foreseen until 1941.12

Conclusion: Games Versus Reality in the Pacific 173

The pre-war Naval War College did not envision a new Battle of the Atlantic, but it did foresee large-scale
anti-submarine warfare because the Japanese submarine force was expected to be a substantial threat. On
this basis, the U.S. Navy spent considerable efforts developing sonar and anti-submarine tactics, although
once war came it was not nearly as advanced as the Royal Navy. U.S. industrial resources did provide large
numbers of escorts and escort carriers, as well as much-improved sonars. The destroyer escort Harmon (DE-
678) and another destroyer escort (unidentified) are shown alongside the destroyer tender Dixie (AD-14) in
Hawthorne Sound, New Georgia, in May 1944. (National Archives 80-G-238416)

From the early 1920s, the message was that the fleet would have to steam through
areas in which it faced constant attack by aircraft, submarines, and surface torpedo craft.
There would be no well-defined threat axis, because the fleet would find itself in the midst
of Japanese-held mandated islands. Prior practice, both in the U.S. Navy and abroad, was
to cruise in columns, in a rectangular formation. Such formations were easy to maintain,
but difficult to maneuver. Typically they could be converted into battle formation once an
enemy fleet was sighted. They were poorly adapted to a situation in which the fleet faced
constant threats from unknown directions. The alternative, invented by Commander
Roscoe C. MacFall as a War College student, was a circular formation. It may have been
tried for the first time during Tactical Problem IV (Red-Blue, played in November 1922
for the Class of 1923).13 This was also one of the first games critiqued by Captain Harris
Laning in his new role as head of the Tactics Department. Laning described the proposed

174 Winning a Future War

formation in considerable detail “because it seems to have possibilities. It is the only one
so far produced at the War College that permits a change of front of a huge fleet without
either causing confusion among the many ships when they rush to a new position, or
sacrificing security and ability to deploy.” A more conventional formation was difficult
to turn. In the long run, the most important virtue of the circular formation was that it
provided an all-round defense. The circular formation was used again in a Blue-Orange
game, and again Laning described it in great detail. One of MacFall’s fellow students,
Commander Chester W. Nimitz, was assigned as aide to U.S. Fleet commander Admiral
Samuel S. Robison when he left the War College. He sold MacFall’s idea, and by 1930 it
had been adopted as a standard formation. Meanwhile, it invariably featured in the staff
solutions to the estimates of the situation.
MacFall thought of the circular formation as a cruising formation rather than as a
battle formation, because battle meant a fight between two battle lines plus other warships.
However, the battles in the Mandates would generally pit the U.S. fleet against Japanese
aircraft and submarines. The circular formation was the battle formation in these cases.
Air-sea warfare in the Pacific was generally much more like the attrition war ex-
pected in the Mandates than like the culminating fleet battle envisioned as their se-
quel—hence, the importance of the circular formation. As it happened, the circular for-
mation was also very well adapted to carrier operations, and even to multi-carrier ones.
It reflected the Naval War College perception that a Pacific War would necessarily be a
protracted campaign.
The Naval War College envisaged three Japanese means of attrition. One would
be air attack, using either carriers or land-based aircraft, or a combination. A second
would be submarine attacks. A third would be mass surface torpedo attack, particularly
at night. Although the Pacific War did not turn out quite as expected, the Japanese took
a similar view: The college successfully predicted the methods they would emphasize. In
doing so it also predicted what many, looking back, see as peculiarly Japanese emphases.
Thus, the war games reflected an expectation that the Japanese would mount large
night torpedo attacks on the U.S. fleet. The night battles in the Solomons should not
have been nearly the surprise that they were. There seem to have been two reasons this
was not the case in reality. First, the pre-war U.S. Navy did not conduct extensive night
exercises because they were rather dangerous, hence politically unacceptable. As a conse-
quence, it did not gain the night expertise achieved by either the Imperial Japanese Navy
or the Royal Navy.14 That meant it did not come to appreciate the special requirements of
night combat. Worse, radar seemed by itself sufficient to solve the problem of night com-
bat, in effect to turn night into day. That was not quite what happened. Radar plus the
plot in a combat information center really did change night combat for the U.S. Navy, but
that took time to develop. Before that the Navy suffered badly in the Solomons. However,
it did not do so out of deep ignorance as to what night torpedo combat could mean—the
pre-war games made that fairly obvious.

Conclusion: Games Versus Reality in the Pacific 175

 Pearl Harbor clearly changed the war from the expected immediate step-by-step ad-
vance through the Mandates to something much more protracted, and far more reactive
to what the Japanese were doing. Thus, there was no direct war game equivalent to the
desperate fight around Guadalcanal or, for that matter, to Midway. However, there were
plenty of carrier-versus-carrier battles. There were also cruiser battles fought as part of
campaigns to attack Japanese trade as a distraction. There was considerable game (and
full-scale) experience showing that the U.S. battle line would have little or no part in a car-
rier-versus-carrier struggle. In that sense, Midway was fought as though Pearl Harbor had
not happened—except for the perception that further major losses would be disastrous.
Once the U.S. fleet had been rebuilt, the pre-war concept of a step-by-step advance
through the Mandates was revived. As expected, the Japanese held back their battle fleet
while contesting the U.S. advance mainly with aircraft. There was, however, a major sur-
prise. The war games showed that there would be huge wastage of aircraft on both sides.
For the U.S. Navy, the lesson was simple. As war approached, production not only of
aircraft, but also of pilots was stepped up. It was assumed that the Japanese would do the
same, although it was known that they could not remotely match U.S. aircraft produc-
tion. That was not true; until 1944, the Japanese persisted in training only small numbers
of new pilots. They also failed to rotate exhausted pilots out of combat. The effect on the
Japanese of large-scale pilot and airplane losses were therefore underestimated. It seems
to have been assumed that Japanese naval aviation could be eliminated only if the Japa-
nese carriers were sunk.
The effect of such misperception was evident after the Battle of the Philippine Sea in
June 1944. During the battle, U.S. commander Admiral Raymond Spruance considered
protection of the invasion of Saipan his primary responsibility.15 He was unwilling to
pull his fast carriers away from the island in order to pursue the Japanese carriers. The
Japanese found the U.S. carriers before they themselves had been detected. In the past,
this would have led to serious U.S. losses. By this time, U.S. fighters and fighter control
were so effective that the Japanese strike force was massacred. The air portion of the bat-
tle was called the “Turkey Shoot.” Because of the perception that no such massacre could
be decisive, Spruance was widely condemned. As Pacific commander, Admiral Nimitz
told the fleet commander for the next major operation, Admiral William F. Halsey, that
the Japanese carriers were his primary target, even if that meant sacrificing the landing
at Leyte Gulf.16 It seems not even to have been suspected that at the time the Japanese
carriers were effectively toothless, and that they were almost certain to remain so.
At Leyte Gulf, the Japanese used their carriers, nearly all of which had survived the
Philippine Sea battle, as bait to draw Halsey away from Leyte Gulf and the invasion
shipping, leaving it open to their powerful surface fleet. When the Japanese surface force
actually attacked, Halsey was roundly criticized. His pilots had attacked this force the
previous day, and had reported routing it. Halsey was later criticized for naively believing
the pilots, despite considerable experience that they often exaggerated. On the basis of

176 Winning a Future War

The one great failure of War College analysis was the kamikaze. No one seems to have realized that the Jap-
anese would have to resort to special tactics as they ran out of experienced pilots, probably because no one
aware of the huge U.S. pilot replacement program could imagine that any serious navy lacked an equivalent.
Once the Japanese had too few effective pilots, suicide tactics were the only ones that had much chance of
success; they were also consistent with Japanese culture, which glorified individual sacrifice through suicidal
attack. The War College tended to treat potential enemies as Americans who spoke other languages, but
shared U.S. thinking processes. The carrier Randolph (CV-15) is shown at Ulithi alongside a repair ship after a
kamikaze damaged the after part of her flight deck. (National Archives 80-G-344531)

this belief, Halsey felt justified heading north to attack the Japanese carrier bait. The Jap-
anese surface force turned around and struck Leyte Gulf. The invasion shipping survived
due to a heroic stand by U.S. destroyers and escort carriers off Samar. War gaming did
not describe anything like the complex attack mounted by the Japanese.17 However, it
did raise exactly the question Spruance and Nimitz faced: What was the mission (or the
priority) of the commander of a force covering an invasion?
The kamikaze surprise is another aspect of the same failure of perception. The
Japanese adopted kamikaze tactics because by late 1944 they no longer had sufficient
numbers of pilots who had any chance of attacking effectively using conventional tactics.
The choice they made certainly did reflect Japanese culture; no one in the West would

Conclusion: Games Versus Reality in the Pacific 177

have used suicide tactics on a large scale. However, Western accounts of Japanese culture
certainly did discuss their willingness to sacrifice lives freely “for the Emperor.” There
were even examples of deliberate suicide attacks earlier in the war. No one included mass
suicide attacks in war games, but mass air attacks were common. Given the difficulty of
estimating how well defenses would do, it is not clear that including the kamikazes would
have made an enormous difference at the gaming level. Moreover, by the spring of 1945
it seems to have been clear that kamikaze attacks were a rational response to the success
of U.S. fighters and anti-aircraft fire. It would actually cost fewer pilots’ lives (in their
function as kamikazes) to inflict a given level of damage on the U.S. fleet.18
What does all this say about gaming at Newport as preparation for World War II?
Above all, gaming was an excellent preparation for an entirely new kind of war. Gam-
ing teased out the ways in which aircraft could or would fight at sea in large numbers.
Full-scale exercises could and did test carrier-versus-carrier and carrier-versus-battleship
tactics, but they could not possibly simulate a protracted grinding campaign. Moreover,
gaming could show what would happen as temporary limitations on aircraft capability
were overcome. Looking back, we see officers who really did operate aircraft complaining
before the war that what the War College allowed was often unrealistic, but aircraft were
developing so rapidly that this was more virtue than vice.
The entirely new type of war could not be simulated on a full scale: No one was
going to see what happened if massed torpedo bombers encountered massed fighters, or
if they attacked a battle line at sea. There was some simulation, and there were dummy
attacks with practice weapons, but they were clearly unrealistic (among other things, no
one was shooting back). In all other cases, actual combat experience gave some sense of
what future warfare would be like, although all types of naval technology had moved on
considerably since 1918.
Gaming portrayed not just short battles but whole campaigns. Nothing but gaming
could have simulated this, and the virtues of gaming explain why so much game material
ended up in the files of the War Plans Division of OPNAV. An individual student would
experience a year of games and perhaps, as a staff member, another year or two. In the
1930s, the Research Department sometimes explored the lessons of multiple games, and
it produced a summary of highlights.19 However, anyone exploring the War Plans Divi-
sion files saw much more, and got a much better sense of how alternative approaches to
the fundamental Pacific problem played out.
Moreover, in the course of the games players tried many ploys that could feed into
future plans. For example, in one game the Blue commander withdrew one of his carri-
ers to make a surprise strike on Tokyo. Because the game did not take Japanese policy
makers into account, it was impossible to decide what the effect of the ploy would be. It
was, however, interesting enough that it was included in a series of notes on the games
produced by Naval War College’s research director Captain van Auken. It is impossible
to say whether it had any later consequences.

178 Winning a Future War

 Gaming, at least as it was practiced at the college, offered no insight into the way the
Japanese thought. The Japanese as they were simulated at the college were Americans
provided with the Japanese fleet and Japanese bases. That is why it is so impressive that
the sort of tactics the Japanese actually used during the war were evident in the games.
On the other hand, gaming at this level could not provide answers to questions like what
it would take to end the Pacific War. There is no evidence of much insight of that sort
anywhere in the U.S. government of the time.
Pre-war gaming only lightly explored the problems and potential of coalition war-
fare, because the pre-war United States was neutral, with a foreign policy that empha-
sized neutrality. As fought, the Pacific War was very much a coalition effort, but the most
important impact of the coalition was in the geography of the war. Defense of Australia
was a vital theme in the Pacific, and it largely accounted for the devastating fight around
Guadalcanal. It could not and did not figure in pre-war gaming (Australia was generally
a friendly neutral).
Gaming had far less impact on U.S. performance in the Atlantic War. No such war
was contemplated for most of the inter-war period (with the exception of various Red-Blue
exercises). It amounted to a replay of World War I, hence anathema to many Americans.
Moreover, the Atlantic War was far more a coalition affair, and also far more affected by
strategy for the land and air components of the war. Convoy warfare did not entail major
new operating concepts like those featured in the Pacific War. War games played in 1941
were clearly intended mainly to familiarize officers with the conditions of the undeclared
naval war in the Atlantic then being fought. All of this having been said, it seems unlikely
that the U.S. Navy would have been nearly as effective in the amphibious phases of the
Atlantic War had it not developed a form of amphibious warfare to fight in the Pacific.
What does all of this mean for war gaming? War gaming was best at exploring a
new kind of warfare, for which prior experience was not very relevant. That meant above
all air-sea warfare in the Pacific, with amphibious warfare a close second. War gaming
forced practitioners to think in campaign terms—not in terms of a single battle. It was
consideration of a Pacific campaign that led the U.S. Navy to understand what air-sea
battle would be like, and what it would mean. The great contrast with the Japanese and
the British is that neither thought through a protracted Pacific campaign. It is difficult
not to see war gaming as the difference. Nothing short of war gaming could have provided
a campaign perspective.
War gaming was also valuable because it allowed free exploration of possibilities,
away from any publicity or leakage. War gamers could envisage options that were en-
tirely unacceptable politically during the 1930s—and that were sometimes actually re-
alized during World War II, such as coalition warfare. Secrecy mattered, and it proved
possible to enforce.
Conversely, war gaming was a poor way to investigate the problem of how to end the
Pacific War—that depended very much on how the Japanese thought, which was just

Conclusion: Games Versus Reality in the Pacific 179

what the War College could not simulate. The assumption through all the games was
that the Japanese government would behave rationally; it would come to terms (presum-
ably short of unconditional surrender) when it knew that it was beaten. No war game
envisioned the bloodbath at Okinawa or the much worse one the Japanese planned for the
invasion of the Home Islands. Nor, incidentally, did war planners who had learned their
trade largely by gaming (simulation).

180 Winning a Future War

A

A: Playing the Games

There were two types of game, the large-scale (chart) maneuver and the smaller-scale tac-
tical game (board maneuver); all of the games were called maneuvers rather than games.
Notes in the account of Chart Maneuver I (Class of 1928) show how a large-scale game
was played. Typically each student turned in an “Estimate of the Situation.” All were
evaluated, but solutions (estimates and basic decisions) were produced by two staff mem-
bers as a basis for further play. No student solutions were collected for the 1928 game,
which began with a fixed set of orders for the two CinCs.
Normally solutions included radio frequency plans; in this particular case, Blue and Or-
ange frequency plans, indicating which ships talked to which commands, were provided.
Normally each CinC held a conference to explain to his subordinates his estimate
(commander’s intent), his decision, orders, and plans, and also such doctrines as he might
choose. He gave copies to all the subordinate commanders and to the game director. The
game director chose the length of each move. When ready for a move, the director would
post an announcement giving the serial number of the move and the date and hour when
it began and ended, and also the weather and visibility. He would set a clock dial to the
game time at the beginning of the move.
For the first move, each player turned in a tracing showing what his force would do.
A plotter transferred these movements onto the master chart in the game room. Once all
the first-move tracings had been submitted, a player could see whether or not there would
be contact between the forces. The assistant director for plotting would transfer students’
tracings onto his own chart showing what had happened during the move.
If there was no contact, the next move would be called. Even without contact, players
had to work up fuel and similar accounts, and they might send dispatches. In some cases
this work-up delayed play; the game director advised players to get their tracings (“flim-

Appendixes 181
sies”) in first, then dispatches, and only then fuel account and plans for the next move.
Players also had to turn in their “record of move” so that a history could be kept. The
flimsy showed what the player intended to do if nothing happened; the record showed
what the player actually did. There was provision for changing intentions, but only if the
player did not try to go back in game time.
Commanders of forces not in sight of each other were placed in separate rooms or
otherwise separated to keep each from knowing what the other’s force was doing. All
communication between individuals not assumed to be in the same ship was by message.
The rules included transmission and decoding delays.
If ships came into contact, either in sight of each other or in sight of smoke, the di-
rector or his assistants informed the commanding officers of the ships, and asked what
their action would be. The resulting movements were plotted. Players could ask for more
information, but were reminded that “every commander always suffers from lack of in-
formation.” The game continued once contact was cleared up. Play was often delayed by
the mass of messages sent upon contact; players were reminded to write up and hand in
messages as quickly as possible, because in reality there would be little time to waste once
in contact with the enemy. In some cases, players intended to send radio messages but
delayed until later moves, even though the messages affected action during earlier moves.
Players also sometimes took too long to decide what to do.
Once contact became general, action might be moved to the game board or the game
might be called. If the maneuver was to continue as a chart maneuver, contacts would be
handled and arbitrarily adjudicated by the director so that the maneuver could proceed.
The director would decide whether to play a situation out on the game board.
A record of each move was kept, and at the end of the game the records were all turned
in so that a narrative could be constructed. Accounts of games in this book are based on
such narratives, which in some cases included commentary on students’ performance.
The tactical game was conducted in much the same way, each move equivalent to
three minutes of real time unless otherwise specified.1 As in the strategic game, it was
important that players see no more than they would in reality. Typically, screens were
used to cut off visibility; alternatively some players could be placed out of sight of the
board and of any other player handling forces not in sight from his force. A player would
make his moves on tactical plotting sheets. They might be transferred either to the game
board or to the master plot, or to both. Commanders of submarines (individual players)
were permitted to see the board only when their craft were surfaced, and when permitted
by the director.
Players turned in special blanks at the end of each move. If a move blank or other
form called for a ship doing more than the rules allowed, the umpire would return the
form for correction. In the event of undue delay, the director might simply require a force
to continue what it had been doing on the previous move. No ships would be moved on
the board until the director ordered them moved. All would then be moved promptly.

182 Winning a Future War

 Once turned in, moves could not be changed except for ships whose action was based
upon the action of other ships, or in case a compelling but unforeseeable factor inter-
vened, such as a torpedo hit. Should a new move be ordered before damage had been de-
termined, the value of the ship’s own action would be appropriately revised when damage
was known, but the director would decide when to apply damage by sinking or losing
speed. Damage would be classed as above or underwater, each with its own effects. Dam-
age was expressed as percentage of original life in 14-inch hits. It was counted in units of
10 percent, rounded down (e.g., 0 to 9 percent was counted as 0). As soon as it was de-
termined, damage was reported to the commander under whose personal command the
ship came. All damage during a move became effective at the end of the move.

B: War Game Rules—Aircraft

The war game rules (“Maneuver Rules”) were key to the simulation the Naval War Col-
lege sought to create. They reflected the fleet’s understanding of current naval technology.
In addition to their use at the college, the rules were the basis for the umpire rules used
during full-scale fleet exercises. Thus, they affected the fleet’s experience of simulated
combat far beyond the War College. The fleet’s experience fed back into the rules, from
formal reports of the Fleet Problems (which survive in Naval War College files), from
replies to correspondence sent as the rules were revised each year, and from discussions
during game wash-ups, particularly of the annual “big game.” Members of the Research
Department were present at the wash-ups, and surviving copies of discussion mention
possible changes to make the rules more realistic. For aircraft, contentious issues includ-
ed the effect of fighter defense and the damage effect of dive bombing.
Most, but not all, of the annual editions of rules have survived in NWC files.2 It is possi-
ble to follow changes because each set of rules was a compilation of sheets, each of which,
from 1927 on, was dated. The first of the new-type sets of rules was issued in 1919. How-
ever, the earliest full set that survives is dated June 1923 (new sets were issued each June).
Notes to some of the games show that there were initial difficulties over the effectiveness
of air attack. To some extent, experiments against real warships in 1920–21 may have
resolved them, but important questions could not be resolved in this way. It might be
clear, for example, how much damage a 1,000-pound bomb would inflict if it hit, but it
was much more difficult to say what chance it had of hitting a moving, maneuvering ship
compared with, say, a 14-inch shell. Much the same could be said of air-launched torpe-
does. It was also difficult to evaluate anti-aircraft and air-to-air fire.
In most cases, the analysts at the college could point back to combat experience and
exercises as a basis for rules. For example, they knew the hitting rates of heavy guns in
gunnery practice, and they also knew hitting rates experienced in World War I combat.
They could use combat experience to estimate how much worse crews would do in reality,
compared to how well they performed in peacetime. Aircraft were a very different prop-

Appendixes 183
osition. No one could say, for example, how anti-aircraft fire would affect the accuracy of
a dive-bombing pilot. The analysts did try to use gun-camera footage to translate practice
dog-fights into estimates of combat performance, but they knew that this was far from
adequate, and many in the fleet told them so. Air-to-air tactics were still so primitive that
it was impossible to say how formations of fighters would fare against bombers. Aircraft
performance was difficult to factor in.
It might be argued that World War I experience was relevant. It appears that little
data had been collected during the war, for example on air-to-air combat. Wartime exi-
gencies had precluded systematic training of fighter pilots, so that it might be argued that
the much more highly trained future pilots would perform very differently than their
predecessors. There had been, moreover, little experience of the sort of mass-on-mass
air engagements that figured in the games. Evidence of the effect of anti-aircraft fire was
anecdotal at best, and future guns would have much better forms of fire control—hence,
were likely to be far more effective than those used during World War I. Naval War
College files seem to reveal no analysis of World War I air warfare data, such as it was.
Perhaps more remarkably, there also seems not to have been any attempt to base rules on
experience in the local wars of the inter-war period, particularly the Japanese war against
China and the Spanish Civil War. They were certainly studied, but it may be that hard
data were lacking.3
Accounts of war games sometimes feature air battles which seem entirely unrealistic.
For example, battleship and cruiser scout planes, with their limited armament and bomb
capacity, sometimes successfully attacked enemy carriers, wrecking their flight decks, de-
spite the presence of their fighters. In some cases, bombers and even low-performance
observation aircraft shot down fighters, because the rules made that possible.
For a modern reader, the game rules give a sense of how the U.S. Navy saw the in-
ter-war evolution of carrier aviation. This account, though considerably simplified, also
gives a sense of just how detailed the rules were and of how much effort the War College
had to expend to take them into account during a relatively fast-moving game. The rules
could not simply be codified into slide rule settings or graphs. Many of them involved
judgment, and players could seek advantages by using them creatively.
Until 1930, aircraft could be detected only by eye. The rules indicated at what range
aircraft could be seen. For example, a VT above the observer could be seen from five miles
away. In particularly high visibility, another mile was added. A surface observer could see
an airplane on land two miles further away. In 1930, sound detectors, which armies were
then using, were added to the game.4 These detectors were employed on land to defend
a base against air attack in one game (this one alerted defending fighters). It is not clear
whether they were ever considered by the U.S. Navy for shipboard use. The June 1941
rules still included sound equipment, by then quite obsolete, but also included the state-
ment that “a few installations exist of other (secret) means of detecting airplanes at a dis-
tance of several miles [i.e., radar]. The Director will rule as to their availability on request.”

184 Winning a Future War

 For simplification and comparison with game and wartime experience, this sum-
mary is limited to carrier aircraft. It therefore excludes extensive discussions of air
spotting for heavy guns (an important consideration in inter-war battle fleet tactics,
but irrelevant during World War II); seaplanes; and rigid airships. The rules were nev-
er intended as general guides to materiel and to tactics. The 1929 aircraft section, for
example, warned that
players should have a thorough knowledge of the characteristics, perfor-
mance, and use of each class of aircraft, and a general knowledge of air-
craft tactics.... These rules are based on the best information obtainable
and for average conditions but should not be considered literally as results
that will always obtain in actual practice.... For instance, the rules relat-
ing to visibility represent ranges at which objects are visible to the eye,
but in order to permit aircraft to employ the principle of surprise with
which they are endowed, it cannot be expected that such swiftly moving
objects will always be seen at the same definite range. The Maneuver Rules
concerning aircraft are laid down to give a fair basis for development of
tactics and strategy, and also to emphasize certain characteristics of air-
craft, such as their fragility of construction, liability to breakdown, and
short life, as compared to other naval materiel. At the suggestion of naval
aviators, breakdown penalties have been kept quite severe in an attempt
to duplicate war time operation with average material, and to discourage
excessive operation. It is well known that in peace time, airplanes can be
made more reliable than these figures indicate...effectiveness is all relative
and...the life assigned a battleship is probably as open to argument as is
the effectiveness of the air, submarine, or destroyer arm.
In the account that follows, different parts of the aircraft rules are treated separately
to show how they changed from 1923 onward. Strikingly, the 1939 rules opened with a
discussion of what sort of airfields would be needed by modern heavy aircraft, a natural
issue in games concerned with seizing and developing advanced bases, but one absent
from previous sets of rules.
The aviation part of the rules was Section J, which in the 1923 edition came last, after
submarines, and was 15 pages long. It included tables of aircraft performance. It referred
to rules in other sections, such as those for visibility (Section D). When he answered the
1935 call for revisions, BuAer chief (and later CNO) Rear Admiral King pointed out
that aircraft were the only naval weapon not yet really tested in war, so to some extent the
rules reflected opinion as well as experience.5 Questions of air-to-air combat and modern
anti-aircraft effectiveness were particularly difficult to answer.

* * *

Appendixes 185
The Airplanes
In 1923, the only surface attack aircraft were torpedo bombers capable of carrying either
a lightweight (1,650-pound) torpedo or one 1,500-pound bomb. Patrol planes (seaplanes)
were credited with four 230-pound bombs. All aircraft came in both land and seaplane
versions. Performance figures included maximum and cruising speeds; a VT cruised at
70 knots but could make 90 or 95 (landplane version); a fighter had a maximum speed of
130 or 135 knots. Data included the speed the airplane could make while climbing, and
its ceiling, span, and gross weight. Endurance was also given: 6.75 hours for a land-based
VT, 4 for a land-based fighter. All aircraft other than fighters had radios.
In 1926, for the first time, the game rules were based on the performance of actual
U.S. Navy aircraft. From then on, standard practice was to assume that both sides used
U.S. aircraft. In a very few cases it was accepted that a foreign navy had aircraft for which
the U.S. Navy had no equivalent.6 Overall, crediting foreign navies with U.S. types of
aircraft was a way of avoiding the problem of Japanese secrecy. No one really knew how
Japanese aircraft performed. In effect, the inter-war college assumed that U.S. aircraft
represented the state of the art and, true to the conservative practice of avoiding giving
the U.S. side any particular advantage, credited the Japanese with aircraft quite as good.
In the 1925 rules, the heavy carrier aircraft could function as torpedo bombers or
level bombers or scouts. VTs could carry one 1,000-pound bomb or two 500-pounders
or 30 25-pounders—in each case, enough to wreck a flight deck. Dive bombing did not
yet exist for the Navy, although the Marines were developing it. There were also single-
and two-seat fighters. Fighters could carry four 25-pound bombs.
The 1925 rules pointed out that aircraft should be organized into tactical groups be-
fore taking off, although they could be reorganized once in the air if the necessary signals
were available. In daylight without fog, planes could rendezvous in accord with specific
instructions, if they could see some recognizable object. The 1926 rules went further:
Prior to the game, airplanes should be organized into groups, wings, squadrons, divi-
sions, and sections as suitable, in accordance with standard procedure, and operated as
units (this was simplified to squadrons in the 1933 rules). Again, after taking off, tactical
units might be changed, however. Ideas about organization were linked with estimates of
how many aircraft could attack together or could join to fight an enemy’s air strike.

Carrier Air Operation

The 1923 rules specified how quickly aircraft could be flown off and landed on carriers:
one fighter every 15 seconds, but one heavier VT or VS or VO every minute. Either two
fighters or one VT could be brought up from the hangar in 2 minutes. To recover aircraft,
a carrier had to steady herself into the wind for six minutes, after which airplanes could
fly on at two-minute intervals (twice that at night). They could be struck below at the
same rate. However, a carrier could no longer launch or recover aircraft after receiving the
equivalent of a single 14-inch penetrating hit.

186 Winning a Future War

 The rules took the dangers of flying into account. In the initial rules, an airplane had
a 1 in 20 chance of being wrecked when landing on the carrier (1 in 10 at night). Umpires
decided how many aircraft were wrecked. In some games, so many aircraft were wrecked
that air reconnaissance became impossible. As airplanes improved, the wreck figures
were reduced. In 1925, the chance of being wrecked while landing in daylight was given
as 1 in 30 (1 in 20 in darkness), a 50 percent improvement. In 1927, rules were made more
elaborate and the chance of being wrecked while landing on a carrier depending on the
type of airplane and the weather. In smooth weather, a fighter had a 1 in 25 (0.04) chance
of being wrecked every time it landed. However, in rough weather this rose to 1 in 10. As
aircraft improved, the 1929 rules reduced the chance of being wrecked to 0.03 under all
sea conditions (0.06 at night). Comparable figures for a landing field ashore were 0.001
and 0.03. In 1933, the chance of being wrecked while landing on a carrier was reduced to
0.01, which was still ten times what a pilot faced ashore under the earlier rules.
In the 1925 rules, aircraft could not land on a carrier with bombs or torpedoes
onboard. This rule made for very high expenditure of weapons whether or not aircraft
found targets, with serious logistical consequences. In one game, bombs were in such
short supply that a badly damaged carrier transferred hers at night to undamaged carri-
ers. In another game, one of three carriers soon had no torpedoes left. In the 1930s, the
Bureau of Ordnance found ways to fuze bombs so that aircraft could land on carriers
while still carrying them. This improvement was reflected in the 1939 rules. Even then,
aircraft could not land on a carrier with their torpedoes onboard.
According to the 1925 rules, an airplane had to be serviced for at least half an hour
between flights if it had been out to less than half its endurance. It would need more work
after a longer flight, up to two hours if it had been out over three quarters of its endur-
ance. Catapults could be fired once every two minutes. Scouts should fly in pairs, because
if operating singly their chance of being unable to return was 1 in 20.
The time required to recover each airplane was crucial. A carrier could not fly aircraft
if she could not take them onboard when they returned. After flying out and attacking, a
returned airplane would have only limited remaining fuel, hence limited remaining time
in the air. The shorter the time to land airplanes, the longer they could remain in the air.
Reeves’s innovations were reflected in the 1929 rules. Aircraft could fly on at one-minute
intervals, and the rule that the flight deck had to be clear for an airplane to land was
eliminated. The 1933 rules cut the daylight landing interval to half a minute. At the
same time, launching intervals, which were already short, were further reduced. A fighter
could be launched every 10 rather than 15 seconds. The change for torpedo bombers was
much greater, from the earlier two minutes between takeoffs to 15 seconds. This change
much reduced the advantage a group of small carriers had enjoyed. If torpedo bombers
took so long to launch, several small carriers launching more or less simultaneously were
needed to place a large strike force in the air. A single large carrier could do far better if
she could launch the same aircraft very quickly. To the extent that grouping carriers was

Appendixes 187
inherently dangerous, the change in rules favored dispersing large carriers as widely as
possible. That had enormous impact. Note that it coincided with the appearance of the
two large U.S. carriers Lexington and Saratoga.
Other altered rules made sense given barriers and arresting gear. Now it was as-
sumed that
at least a quarter of a carrier’s planes would normally ride the flight deck.
Half could be accommodated there “in such a manner that either flying off
or flying on, but not both simultaneously, can be carried on.” A maximum
of nine aircraft could fly off as a group.
The rules changed in 1936 to allow for launch and recovery with aircraft on deck.
All of a carrier’s aircraft could be parked on deck, but in that case airplanes could take off
but not land (i.e., a deck-load strike could be set up). With three quarters of the aircraft
on deck, airplanes could take off (if the aircraft were all aft) or land (with all of them for-
ward), but not both simultaneously. However, with half the airplanes on deck, presum-
ably all amidships, aircraft could take off and land simultaneously. No more than half
the aircraft could be in the hangar at any one time. Such practices differed completely
from those in other carrier navies. Launch interval was given as ten seconds, and landing
interval as 30 (intervals would double in darkness). A new feature was that unscheduled
launching could occur five minutes after the decision to launch—as when enemy ships
were suddenly found—but only out to a radius of 100 miles. Preparations for longer
flights would take 15 minutes. The 1936 rules gave refueling time for different types of
carrier aircraft, the least being 20 minutes for a fighter. Rearming times were also given
for bombs and torpedoes.
Carriers had flight-deck catapults throughout the inter-war period. Although they
were mentioned, they were not the subject of any rules. No catapult launch rate was giv-
en. The 1939 rules pointed out that they could be used even if the carrier did not turn into
the wind, as long as the wind was forward of the beam. Otherwise the carrier had to turn
into the wind (five minutes) and the apparent wind had to be within 10 degrees of the
ship’s centerline. That might be a real advantage. Presumably this new rule was framed
because the new carriers had more powerful (and faster-firing) catapults, but carrier deck
catapults were not heavily used until late in World War II.
The 1941 rules, the last written before war broke out, added that carriers had arrest-
ing gear fore and aft, hence could land planes going either ahead or astern. Such gear had
been in place at least since 1939.

Bombing
Initial estimates of bombing accuracy were optimistic. For airplanes flying at up to 1,500
feet, it should take two bombs to make one hit on a large or intermediate ship, such as a
battleship or cruiser; up to 4,000 feet, four; up to 8,000 feet, six; and up to 12,000 feet,

188 Winning a Future War

ten (twelve for an intermediate ship). Airplanes dropping bombs from below 500 feet
would be destroyed by their blast. The 1924 rules recognized that the hitting figures were
valid only for target practice. They might decline sharply in combat, for example when
bombers met opposition. These modified rules added a table: When met by planes of
fighting strength considerably superior to the bombers, the hitting rate would decrease
by 40 percent. The umpire was to assign further decreases due to maneuvers by the target
ship or speed or anti-aircraft fire. It was assumed that bombers would face either fighters
or anti-aircraft fire, in the theory that ships would not fire with friendly aircraft overhead.
The next year these figures were changed again. Under target practice conditions,
a bomb dropped from 1,500 feet should have a 40 percent chance of hitting (less than
before); from 8,000 feet, 15 percent; and from 10,000 feet, 10 percent— all against
a large ship. Soon it was accepted that bombers would generally attack in formation,
their bombs forming salvoes equivalent, in theory, to the salvoes fired by a ship’s guns
firing together. As with gunnery, the chance of at least one hit depended on the spacing
of the projectiles, which in turn depended both on the spacing of the airplanes and
the precision with which their bombardiers released them. Much also depended on
how the ship moved—maneuvered—between the moment the bombers committed to
an attack and the moment the bombs arrived. Typically, a master bombardier had to
track the ship to estimate its course and speed before releasing bombs.7 According to
the 1926 rules, a 15-knot ship attacked from 4,000 feet would present no problems.
The faster the ship and the greater the altitude, the greater the chance would be that
radical maneuvers would throw off a bombardier. Thus, in 1926, it was estimated that
a bomber attacking a 35-knot ship from 10,000 feet would lose about a quarter of its
effectiveness. 8
In the 1925 and later rules, a single 500-pound bomb was credited with the ability
to destroy half a carrier’s flight deck, a figure retained in later sets of rules. The 1926
rules listed the weapons most suited to various targets, the 500-pound or 1,000-pound
bomb being best against carrier flight decks (but 2,000-pound bombs and torpedoes
were best to destroy a large carrier). Under the 1925 rules, a direct hit by a 2,000-pound
bomb would cause above-water damage equivalent to four 14-inch hits. One hit by a
1,000-pound bomb would cause the equivalent of two such hits. The supposed impact
of bomb hits was sharply reduced in 1931, in effect devaluing the new dive bombers with
their 1,000-pound bombs.
As a way of better understanding air weapons, the 1931 rules pointed out that the
different bombs were about the weight of different standard shells: 2,000-pounder for 16
inch, 1,000-pounder for 13 inch, 500-pounder for 10 inch. This was intended as a rough
guide to which bombs should be used against which targets. Two 1,500-pound or four
1,000-pound or eight 500-pound hits on a carrier were equivalent to one 14-inch hit.
However, it would take four 1,500-pound bombs to have the effect of one 14-inch hit on
a battleship, whose life might be ten to twenty such hits.

Appendixes 189
 Aircraft could also deliver chemical bombs. Under the 1931 rules, four such bombs
hitting a carrier would inflict the effect of one 14-inch hit. However, the effect of gas was
transitory, typically lasting only an hour. Six gas bombs hitting a battleship were credited
with the effect of one 14-inch hit. As noted previously, gas bombs were a fixture of most
inter-war rules, and in many games airfields ashore were kept out of action by repeated
gassing. Rules included considerable information about the effect of gas weapons and
there was much interest in ways of defeating gas attacks.
The 1929 rules introduced the key inter-war technique for air attack against small
moving targets: dive bombing. At minimum altitude for horizontal bombing, 1,000 feet,
a bomber was expected to have a 50 percent chance of hitting a large target. Normally,
it would never attack from so low an altitude, because it would be so easy a target for
anti-aircraft fire. If the level bomber attacked from 10,000 feet, just above maximum
anti-aircraft altitude, it would have only a 10 percent chance of hitting a ship, even if the
ship did not maneuver at all. Navies adopted strafing tactics to dislocate shipboard an-
ti-aircraft guns so that bombers could attack from lower altitudes. It was up to the game
director to decide how well such attacks would work.
A dive-bomber pilot pointed the nose of the airplane directly at the target and, to
some extent, the pilot could adjust as the target maneuvered.9 Anti-aircraft guns typically
could not hit a rapidly diving airplane at all, so the dive bomber could survive to release
its bombs at low altitude. Release altitude was typically set by the altitude at which the
bomber pulled out after releasing its bomb; the bomb was released about 600 feet above
that. Typically, the pull-out altitude was only 500 to 1,000 feet. Under the 1929 rules, a
dive bomber pulling out at 500 feet had an 80 percent chance of hitting (reduced to 65
percent if the bomber pulled out at 1,000 feet).
The 1931 rules handled bombing in a more sophisticated way. Level bombing was
given a slightly better chance of success (12 percent from 10,000 feet). However, anti-air-
craft fire was also rated more effective, so that the lowest altitude for such bombing was
2,000 feet, with a hitting rate of only 26 percent against a large non-maneuvering ship.
The rules now also took into account opposition by fighters during the bombing run.
How effective that might be depended on whether there were more or fewer fighters than
bombers. There was no attempt to go into greater detail as to the degree of superiority or
inferiority. A superior number of fighters could reduce the bombers’ effectiveness by 60
percent; an inferior number by 20 percent.
Dive-bombing accuracy was now estimated by reducing the level achieved in target
practice by 25 percent to take wartime conditions and anti-aircraft fire into account. By
this time, the figures given in 1929 were considered somewhat optimistic, so that in 1931
the chance of hitting a large ship after a 500-foot pull-out was given as 54 percent (the
corresponding 1929 figure would have been 60 percent). The higher the pull-out, the
worse the accuracy, but even a 2,000-foot pull-out was expected to give a 21 percent
chance of hitting a target. These figures explain why dive bombing was so exciting a tactic.

190 Winning a Future War

It seemed that if even a few dive bombers managed to reach an enemy ship, they were vir-
tually certain to score hits. By this time, too, U.S. dive bombers could deliver the heaviest
available bombs: 1,000-pounders.
At this point, it was accepted that the only effective antidote to dive bombers was
anti-aircraft fire, mainly by automatic weapons. Anything heavier required fire control
that could not follow the bomber through its dive. The effectiveness of a dive bomber un-
der heavy fire would be halved (light fire would cost only 10 percent). Comparable figures
for a horizontal bomber were 43 percent under heavy fire and 18 percent under light fire.
However, since the horizontal bomber was much less effective in the first place, it might
be that a combination of maneuver and anti-aircraft fire would negate it altogether. By
way of comparison, even halving that 54 percent hitting rate would make for many hits.
Another table gave reductions due to target maneuvers.10
Bombing data were further simplified in the 1933 rules, with unopposed dive bomb-
ers attacking a large ship simply being given a 40 percent hitting rate. That compared to
27 percent for level bombers attacking a large ship from 2,000 feet, and 13 percent for the
same target from 10,000 feet.
The value of bomb damage was significantly reduced. A 1,000-pound bomb hitting
a battleship had 0.58 of the effect of a penetrating 14-inch hit, the common currency of
War College damage tables. It had 0.75 of that effect against a carrier or a cruiser—which
would have less overall lifetime.
The figures for accuracy were based on actual results in the air section of the annual
gunnery exercises. A 1936 study of the rules tabulated results achieved by dive bombers
in 1929–33: 37.5 percent against large targets.11 On this basis, the rules gave dive bomb-
ers 40 percent hits. In 1936, the percentage of hits was reduced by a quarter—to take
the “fog of war” into account—to give 28 percent hits against large ships.12 In 1929–33,
the comparable probability for horizontal bombing was 18.18 percent, which fell to 17.9
percent for 1929–35. The practice targets were smaller than enemy capital ships, which
is why somewhat higher hitting percentages were assumed. For example, the 17.9 percent
hitting rate of 1929–35 would be corrected up to 19.25 percent against real ships. Taking
off a quarter for operational factors gave 15 percent.
When the rules were revised in 1936, bombing effectiveness was significantly re-
duced, because it was now evident that the explosion of its bomb would destroy an air-
plane as low as 1,500 feet. In the past, it had been assumed that the critical altitude was
400 to 500 feet. The minimum altitude for level bombing was increased to 3,000 feet, the
hitting rate falling accordingly. These rules omitted the percentages of past rules, instead
changing the percentages previously accepted. New bombing-effect diagrams separate
from the rules themselves made it possible to convert hits and near-misses into equivalent
above- and below-water 14-inch hits.
The 1937 rules largely reversed the relationship between dive and level bombing,
perhaps because the U.S. Navy now had Norden precision bombsights for level bombing

Appendixes 191
and also because it seemed that machine guns had bested dive bombers. Thus, a percent-
age of direct bomb hits under the full range of conditions showed 16 percent for dive
bombing a large ship, compared to 8 percent for level bombing from 9,000 feet and 15
percent from 6,000 feet (30 percent from 3,000 feet). However, another table showed
that dive bombing retained much more of its effectiveness at higher target ship speeds.
Effectiveness was halved for a ship making over 25 knots, but the equivalent factor for lev-
el bombing from 9,000 feet was 0.3. War experience would show that all of these figures
were rather optimistic, but that dive bombers could hit moving ships whereas horizontal
bombers generally could not.13
At the same time, the reduction in assumed bomb effectiveness was reversed. This
time a single above-water hit by a 500-pound bomb was considered equivalent to one
above-water and 0.25 below-water 14-inch hits; a single 1,000-pound bomb was equiva-
lent to two above-water and 0.75 below-water 14-inch hits.

Bombs Versus Carriers

Through the inter-war games, the single crucial fact of carrier warfare, highlighted by the
game rules, was that a few bombs could wreck a flight deck and eliminate a ship as a car-
rier. Early figures for bomb damage showed, among other things, that a single 500-pound
or 1,000-pound or 2,000-pound bomb—or even two 100-pound bombs or ten 25-pound
bombs— would destroy a carrier flight deck. Later rules differentiated between hits on
each end of the flight deck, leaving a ship with half her deck. The U.S. Navy placed ar-
rester gear at both ends of its carrier flight decks partly to increase flexibility (the carrier
could recover aircraft while steaming astern, her air group assembled at the after end of
the flight deck ready to fly off), but also so that the carrier would retain some capability
even if the other end of the deck was destroyed.
The 1928 rules included estimates of the time it would take to repair flight-deck hits.
A navy yard or base would require 30 days for permanent repairs to a flight deck ruined
by a 2,000-pound or 1,000-pound bomb, or ten days for temporary repairs. A tender
could do a temporary repair in 15 days, a ship’s company in 20 days. Corresponding
figures for a hit by a 500-pound bomb (half a flight deck) were 20 days and seven days at
a navy yard, 10 days by a tender, and 15 days by ship’s company. Aircraft could fly off a
temporarily repaired flight deck in twice the time normally required. With the forward
end of the flight deck destroyed, aircraft could fly off in one and a half times normal time.
With the after end destroyed, they could not fly on at all, but they could fly off under
normal conditions.
The 1937 rules were written as the first of the new carriers, the Yorktowns, were
completed. They had light wooden flight decks that would be far easier to repair than
the steel decks of the past. These rules reduced the effect of bombing flight decks, so
that both 1,000-b and 500-pound bombs would temporarily destroy only half a flight
deck. That was reasonable: The new U.S. carriers had open hangars that would dissipate

192 Winning a Future War

the effects of a hit. More important, the new rules allowed for quick temporary repairs
onboard a ship. Now, a ship’s company could repair anything less than two large bomb
hits in 12 hours, and two to four in 48 hours. However, more than four hits could not
be repaired onboard. Because a carrier could repair her own flight deck, it now mattered
how many airplanes survived the damage. Thus, the 1937 rules indicated that each large
bomb would destroy a fifth of the aircraft onboard the ship when it hit. This rule placed
a premium on flying off the maximum possible strike.
Note that these rules did not make the point that foreign navies, with their steel-
decked carriers, could not enjoy the new advantages. They would still be difficult to repair
after battle damage.
None of the rules, moreover, took into account the profound difference between U.S.
and foreign carriers. A foreign carrier whose hangar was integral with her hull, which
meant all British and Japanese carriers, would suffer far worse damage if that hangar was
penetrated. An open U.S. hangar would dissipate any explosion inside, but stout hangar
walls would not. Had this difference been clearer in the rules, they might have come clos-
er to predicting what actually happened to the four Japanese carriers at Midway.
The 1939 rules were even more optimistic from the carrier’s point of view. Under
these rules, a carrier should be able to repair fewer than two 500-pound bomb hits in four
hours, and fewer than two 1,000-pound hits in six; two to four large bomb hits could be
repaired in 36 hours (but, as in 1937, anything more than four bomb hits could not be
repaired onboard).

Torpedo Bombing
Under the 1923 rules, aerial torpedo tracks would be plotted by the torpedo umpire.
Half the torpedoes would run properly. Aircraft could lay smoke screens, an important
factor in game tactics.14 Not only could an aerial smoke screen hide ships, it could also
hide aircraft about to emerge to launch torpedoes. In one game, defending aircraft laid
a second smoke screen specifically to frustrate torpedo bombers emerging from their
own (the bombers’) smoke screen, preventing them from finding their targets in time to
release their weapons.
Presumably as a reflection of developing tactics, the 1928 rules called for concen-
trated attacks by torpedo bombers, if possible, simultaneous with bombing. The earlier
figure of half of torpedoes running normally was upped to three quarters, unless curved
fire (angled fire) was used, in which case only half should be expected to run normally.15
Of torpedoes running normally, under good conditions, at ranges of no more than 2,000
yards with favorable track angles and targets not radically maneuvering, 50 percent could
be expected to hit a large target (35 percent could hit a small one like a destroyer). Um-
pires had to decide how much to reduce hits under unfavorable conditions. Attacks by
hostile aircraft during the approach might “somewhat” reduce the efficiency with which
the surviving torpedo bombers completed their attacks, “but it is assumed that defending

Appendixes 193
planes discontinue their attack when attacking planes can come under effective gunfire.”
The anti-aircraft rules, which were not yet firm, would be applied to a torpedo bomber
coming under fire before it could drop its torpedo. These 1928 rules also gave greater
credit for radical target maneuvers.
In view of later experience, these rules much overstated the effectiveness of torpedo
bombers. Presumably reflecting sobering fleet experience, the 1931 rules cut the per-
centage of success for aerial torpedoes to 60 percent (40 for curved fire), but allowed a
maximum range of 4,000 yards.

Air-to-Air Combat
For air-to-air combat under the 1923 rules, the umpire would decide the outcome based
on numbers, types, and the previously mentioned Lanchester’s Square Law (fighting
strength was proportional to the square of numbers). Land-based aircraft were given an
advantage over seaplanes, formations were considered stronger than individual airplanes,
and airplanes flying low over water were considered stronger in defense than at high-
er altitude. Superior altitude was an advantage in attack, particularly if the attack was
a surprise. Estimated relative fighting strength equated fighters and torpedo bombers
and scouts (relationships with patrol planes and spotters were more complicated). When
many airplanes were engaged, ordinarily 10 percent would be shot down every three
moves (nine minutes).
As an example, the outcome of a fight between 50 fighters and 40 torpedo bombers
(total 90) lasting 18 minutes was initially determined by their relative fighting strength
(by the square law) of 25:16. In the first nine minutes, 10 percent would mean nine air-
craft in all shot down. The number each side lost would be proportional to its total num-
bers, meaning 5.5 and 3.5, respectively (rounded off as 6 and 4). On the next move, then,
the strengths would be 46 and 34 (fighting strength 100:55). On this move another 10
percent would be shot down, and losses would be 5.2 and 2.8, respectively (5 and 3). On
the third turn, that would leave 43 and 29 airplanes in the air on the two sides. This sim-
ple melee calculation assumed no coordination between airplanes—fighters at this time
lacked radios, although pilots could use hand signals. Fighter control of any kind was not
envisaged. Fighters and torpedo bombers were treated equally, presumably on the theory
that the rear-firing guns of the latter were at least equivalent to the fighters’ one or two
forward-firing guns.16
The rules became more sophisticated as the post–World War I U.S. Navy gained
experience with aircraft. The 1925 rules envisioned very short engagements: Airplanes
would not remain in firing range and position for very long. However, a single bullet could
bring down an airplane. This was very different from the cumulative damage envisioned
in ship-to-ship combat. Given superior performance or even position, moreover, a pilot
could break off (or avoid combat) at will. Planes of equal performance would engage only
if both were willing or if one was in a favorable (higher-altitude) position. The square

194 Winning a Future War

law was modified to take account of different performance by different types of aircraft:
One single-seat fighter was considered equivalent to half a two-seat fighter (because the
latter had a rear gun) and also to half a VT (again, because of the rear gun), despite its
performance edge. Because it was so difficult for several airplanes to attack together, the
square law was now applied to airplane-on-airplane combat, the maximum advantage be-
ing 4:1 (one two-seat fighter versus one single-seat fighter, for example). The rules pointed
out that estimates of relative strength were more applicable to formations, where relative
positions of aircraft gave a maximum volume of fire and thus mutual protection, than to
engagements between single aircraft.
Rules for air-to-air combat became more complex as air tactics developed. According
to the 1926 rules, the maximum number of flexible-gun aircraft that could form a single
fighting unit was 24; the maximum number of fixed-gun aircraft was nine. This is why,
in the wash-up after a game in the early 1930s, an Army Air Corps student spoke of 72
fighters, of which only 9 would dive to attack at any one time. This time, the cumula-
tive-hit model of damage used for ships was also applied to aircraft, which were assigned
lifetimes calibrated in machine gun hits: 20 for a single-seat fighter, 25 for a two-seater,
and 35 for a single-engine VT. That was realistic—a single hit could destroy an airplane,
but most single hits would not. Even during World War I, airplanes got home with bullet
holes in them. The hitting power ratios first used in 1925 were slightly modified, strength
being 2 for a single-seat fighter, 4 for a two-seat fighter, and 5 for a single-engine VT. It is
not clear why the VT was considered more effective than any sort of fighter.
The aggregate hitting power on each side would be the total of airplane life multi-
plied by airplane hitting power. The square law was replaced by a ratio of relative fighting
strengths. As in 1925, total casualties in a fight would be a proportion of the total number
of aircraft: 15 percent for fewer than 36 aircraft, 10 percent for 37 to 90, and 5 percent for
more than 90. Instead of proportioning losses based simply on numbers, they were based
on relative total fighting strength. There were further rules for one-on-one combat be-
tween “stunters”—highly maneuverable aircraft such as fighters—and non-stunters such
as a VT or VP. It would take four minutes for a stunter to finish a fight with a non-stunt-
er, or eight minutes to end a fight with another stunter.
By 1933, the fleet had learned far more about air-to-air combat, and the rules were
rewritten. Like other contemporary air forces, the U.S. Navy organized its fighters in
threes, of which two would attack a formation and the third would cover them.17 The
revised 1933 rules for air-to-air combat reflected new tactical experience, but it was still
assumed that a flexible gun used to defend a bomber was more effective than the two
fixed guns of an attacking fighter. This logic justified the current tactical practice of using
three-fighter groups attacking together, two of the fighters firing at the single target, and
the third protecting them. It was assumed that the fighters would attack from greater
altitude and that they would initially enjoy some surprise. The targeted bombers would
maintain their formation when attacked.

Appendixes 195
 If fighters attacked dive bombers and enjoyed a 2:1 advantage in numbers, the fight-
ers would lose 15 percent of their numbers, the dive bombers 20 percent. With 3:1 odds,
the situation would improve: the fighters would lose 10 percent, the bombers 25 percent.
Against slower but more heavily armed torpedo bombers, at 2:1 odds, the fighters would
sacrifice 20 percent to kill 25 percent of the bombers. In fighter-fighter combat, with the
numbers roughly even, each side would lose 15 percent in each attack. These rather pes-
simistic figures (for the fighters) reflect the reality that they did not yet much outperform
their targets—a reality that would change dramatically over the next few years.
When the rules were revised in 1938, it was still assumed that flexible rear guns on-
board bombers would be very effective. U.S. naval fighters still had only two guns of their
own, although that would soon change dramatically. The rules also allowed for attacks
by aircraft such as observation planes and bombers, with single fixed forward guns. They
were credited with half the effectiveness of fighters. No one seems to have pointed out that
there was a world of difference between a fighter pilot with lengthy gunnery training and
a bomber or observation pilot with little or no gunnery experience. Moreover, bomber
performance was still very close to fighter performance. On this basis, it was now assumed
that the defenders had a real edge: With 2:1 odds, the fighters would lose 30 percent to
shoot down 20 percent of the bombers.18 These figures were wildly unrealistic in that
bombers were rarely able to shoot down fighters in any numbers. Experience in Spain
and in China was showing that high-performance bombers often could penetrate to their
targets, but also that in doing so they rarely destroyed large numbers of attacking fighters.
Finally, in 1941 it had to be admitted that situations would be so varied that the
game director would have to assign permanent or temporary losses after considering the
conditions of the fight, allowing for factors such as altitude, formations, and positions
permitting mutual support. The previous table of relative losses was repeated unchanged.
However, the comment was added that when a group of airplanes was attacked by three
times their number, both sides would incur the expected losses and the attackers be driv-
en off; they could not resume their mission for 30 minutes. A group making an aerial
attack could not engage again for ten minutes.
There was no mention of fighter control. At this time, numerous U.S. official ob-
servers, mostly naval, were working in the United Kingdom and onboard British ships,
and their reports had given full details of fleet fighter operations and also of the fighter
control being exercised ashore. Presumably, the new rules reflected combat experience to
some extent, as U.S. files of the period show extensive reports from the observers.

Anti-Aircraft Firepower
Anti-aircraft guns were considered effective only up to a 10,000-foot altitude or a 5,000-
yard range. Effectiveness would vary with the state of the sea, light, and visibility, and
with the type, number, position, course, and speed of the aircraft, and also with the
amount of warning given. No figures were given in the 1923 rules. Later, the effectiveness

196 Winning a Future War

of a carrier was rated equal to that of two battleships or battlecruisers, or to that of six
light cruisers or 24 destroyers. The 1933 rules differentiated between heavy, moderate,
and light levels of anti-aircraft fire. Heavy fire could inflict 20 percent casualties on air-
planes bombing horizontally from 5,000 to 8,000 feet. Light fire could inflict 5 percent
casualties. Heavy fire could inflict 40 percent casualties on dive bombers (light fire could
inflict 10 percent). Coordinating dive bombing with a horizontal bombing attack would
halve the casualties suffered by the horizontal bombers. Experiments had not shown how
difficult it was to hit dive bombers.
The levels of anti-aircraft fire were associated with types of ships and with whether
the ships were also fighting a surface battle. When not under effective surface fire, a divi-
sion of battleships or cruisers or a carrier could develop heavy anti-aircraft fire. The next
level down (moderate) was associated with individual battleships and cruisers.
The 1935 rules went into greater detail about anti-aircraft batteries. It took one 1.1-
inch gun or two .50-caliber machine guns as equivalent to a heavy anti-aircraft gun, an
unrealistic rule because the machine guns and the heavier guns had such different char-
acteristics.19 Heavy, moderate, and light batteries were now defined in terms of numbers
of guns: 20 for heavy, eight or more for moderate, and fewer than eight (but more than
one) for light. At this time, the standard battery for a U.S. battleship or a recent cruiser
was eight 5-inch guns plus .50-caliber machine guns (each of the two big carriers had
12 5-inch guns). New battleships then in the design stage had 20 5-inch guns. The rule
stated that the guns of a division of ships, operating 1,000 yards apart or closer, could
be considered one battery. That explains why, in some games, carriers operated close to
battleships to share their anti-aircraft firepower.
In 1936, further conditions were added. Standard figures envisaged a ship under fire
by a similar ship, but not firing her own guns. If the ship was not under effective fire, the
effect of her anti-aircraft fire would be increased by 20 percent, but if she was firing her
own secondary battery, it would decreased by 20 percent. If the main attack was preceded
by at least nine light aircraft (e.g., dive bombers or fighters) bombing and strafing, the ef-
fect of a ship’s anti-aircraft fire would be halved. On the other hand, now the anti-aircraft
batteries of all ships within 3,000 yards could be grouped together. Half a ship’s anti-air-
craft guns were considered effective on each side. These rules stated that the machine
guns were effective out to 5,000 yards, each machine gun being considered equivalent to
half a larger-caliber (5-inch/25-caliber or 5-inch/38-caliber) anti-aircraft gun. This last
was clearly excessive; in the 1937 rules, machine guns were effective only within a slant
range of 3,000 feet. The 5,000-yard figure was the horizontal (not sloped) range to which
heavy anti-aircraft guns could fire.
By this time, the fleet was exercising anti-aircraft guns systematically, so the 1937
rules included tables of expected performance using a battleship as the unit of effective-
ness. Airplanes were assumed to attack in groups of no more than 60 (wartime experience
would show that practical attack groups were considerably smaller). The situation would

Appendixes 197
show how many units were firing at the group, and that in turn gave a figure for effective-
ness, which might be diminished by further factors. No more than eight ships could fire
at any one attacking group. Guns and machine guns were counted separately, so that a
battleship had a unit of guns and one of machine guns, but a carrier had two units of ma-
chine guns. An old destroyer armed with a single 3-inch/23-caliber gun was credited with
only 0.1 unit of anti-aircraft guns, which might be considered grossly optimistic. A loss of
40 percent would abort an attack. For example, a single firing unit could exact 60 percent
losses in a five-airplane group under ideal conditions, or 40 percent of a 15-airplane group.
Doubling up helped: Four units could wipe out a five- or ten- airplane group, or destroy
40 percent of a 40-airplane group. An additional table showed that by flying higher-level
bombers could reduce their losses to only 20 percent of the basic figure at 9,000 feet.
By this time, it was beginning to be apparent that it was difficult to hit dive bombers.
In their approach, guns could exact only a tenth as great losses as with level bombers,
because the dive bombers did not have to make a long, straight and level approach. When
diving, they would face only machine guns, which were credited with better results: The
standard figure would only be halved.
The 1937 rules also went into some detail concerning the effects of strafing and light
(116-pound) bombing by fighters. An effective strafing attack by 12 fighters could reduce
the value of a ship’s turret guns by a tenth for the rest of the day, or her secondary battery
by a fifth, or her anti-aircraft battery by 30 percent. Nine fighters strafing a ship just before
a bombing attack would halve her anti-aircraft effectiveness for that attack. Single fighters
armed with light bombs and machine guns could also seriously damage destroyers.
Anti-aircraft rules were revised again in 1938, the new assumption being that air-
planes would attack in divisions of six to nine aircraft, coming from one direction as rap-
idly as possible. “Although about 100 planes of different classes can be handled tactically
in the same group when cruising, it is assumed that during the attack they will split up.”
Groups would be strafers, level bombers, dive bombers, and torpedo planes all attacking
in separate groups. Only one group could attack one division of ships at the same time
from one direction, though two could attack two separate divisions at the same time.
By 1939, the fleet was using drones as anti-aircraft targets, with startling and de-
pressing effects. Existing fire-control systems seemed well suited to placing bursts within
lethal range of aircraft, but the drones, which were considerably flimsier than operational
aircraft, often survived just such explosions. The damage tables in the 1939 rules did not
yet reflect this uncomfortable perception. The rules even allowed the game director to
increase the effect of anti-aircraft fire by up to 30 percent in special or unusual circum-
stances. The 1939 tables were retained in the 1941 rules, the last peacetime edition.

Aircraft Navigation and Reliability

Aircraft navigation mattered in two complementary ways. First, it determined how far
aircraft could fly and still find their way back to their carriers. This was always consid-

198 Winning a Future War

erably less than the radius of action based on fuel supply. For example, in a wash-up
after a game in the early 1930s, a senior officer who had come from the fleet commented
that airplanes flying more than 25 miles from the fleet were often lost. These airplanes
could have been brought back from a much greater distance using radio, but no one was
willing to risk the carriers that way. In the late 1930s, both the U.S. Navy and the Royal
Navy developed carrier homing beacons which, in theory, solved this problem, but they
required considerable expertise, and they were not always effective in actual operations.
They did not feature in the evolving game rules.
Second, air navigation determined how effective scouting could be. The scout had
to know where it was when it spotted an enemy. Otherwise, no follow-up attack would
succeed. This problem was demonstrated, for example, at Midway, when the Japanese
fleet was found by land-based patrol aircraft, but carrier attackers had difficulty finding
it themselves, despite information available when they were launched.
An airplane generally had to estimate its position by dead reckoning based on its
known performance and its compass heading. The main problem was prevailing wind,
which would move the aircraft off course and might also affect its speed over the sea.
That was only a minor problem over land, but the sea was featureless. To a limited extent,
surface wind could be observed from wave action, but the wind at altitude might be quite
different. A carrier could launch balloons to measure wind directly overhead, but the
wind at various altitudes 50 or a 100 miles away often varied considerably.
Under the 1923 rules, aircraft dead-reckoning steering errors might be as great as
3 degrees if over 1,000 feet altitude, 5 degrees if over 3,000 feet, and 10 degrees if over
6,000 feet. The increase with height was because an airplane would estimate drift across
its path by observing the movement of the surface of the sea or by using navigational float
lights. Wind at altitude usually differed greatly from wind at the surface. Umpires would
indicate whether pilots would have the benefit of wind indication using a balloon. Since
War College problems generally involved scouting, such errors could have a considerable
effect on the estimate and decisions of a commander receiving news from air scouts. In
1925, the 3-degree altitude was 2,000 feet. The 1926 rules eliminated these figures, but
made the same point about dead reckoning.
None of the later rules quantified navigational errors, but in the game, aircraft were
increasingly important for scouting. Their navigational errors could have considerable
tactical and even strategic consequences, but they were not really taken into account.
Games certainly did take into account poor performance by scouts and poorly framed
messages, but they do not seem to have anticipated some of the gross failures marking
actual combat—in the Solomons, for example (the Savo Island debacle began with failed
air scouting).
The 1926 rules did include figures for the increased chance of breakdown due to
continued flying without overhaul or during a lengthy flight. During a half-hour flight,
the chance of breakdown was 1 in 200, but for a 12-hour flight (by a patrol plane, for

Appendixes 199
instance) it would rise to 1 in 15. These rules also pointed to the loss of morale due to the
load of continuous flying over time. The 1927 and later rules went into considerable detail
both for the number of hours pilots could be expected to fly each day and for the effect
of fatigue. For example, fighter pilots could not be expected to fly more than four hours
a day except for additional time in one 24-hour period. Multipliers were used to capture
various kinds of stress, such as the effect of overtime. For flight personnel, recovery from
fatigue would require three hours for every effective extra hour (with the multipliers). A
curve was included to show breakdown hazards due to increased flight time.
The 1931 rules added graphs of navigation, rendezvous risks and for chance of failure
for long flights (up to 20 hours for patrol planes), and a multiplier for the chance of success
depending on what percentage of a flight was made at cruising speed. The navigational
graph gave the percentage chance of failure for both flight duration and distance from a
ship; by this time, some shipboard aircraft had nominal ranges as great as 400 miles. For
example, a 150-mile flight or four flight hours gave a 25 percent chance of failure—which
was considered grossly optimistic in a game wash-up.
Operational losses were difficult to foresee and to calculate. In an effort to simpli-
fy the process, the 1936 rules gave loss penalties; once they added up to 100 percent
for a squadron, it would permanently lose one of its aircraft. Carrier plane losses in a
smooth sea were given as 0.5 percent (1 percent in a rough sea) per flight. Such figures
covered take-off and landing crashes, collisions, crashes in thick weather or high winds,
and forced landings due to running out of fuel. -

200 Winning a Future War

N

A Note on Sources
The Naval War College did not initially maintain its own historical archive. Instead, its
papers were retired to regional branches of the U.S. National Archives. In the 1960s,
they were brought back to the college to form its own collection. Most of the Naval
War College publications—the pamphlets describing games and those describing lec-
tures—seem to have survived. No correspondence describing the rationale for various
scenarios seems to exist. Moreover, there is no reason to be certain that papers for all
of the games have survived. For example, there is no trace in the college collection of
informal games played to test various technical concepts in the 1930s. The evidence of
such work is in OPNAV and other files citing college contributions to decisions about
ship and weapon characteristics.
Descriptions of games vary. Generally surviving materials describe the scenario and
also the staff solutions for the Blue and enemy “Estimates of the Situation.” Students
received the description of the problem and wrote their own estimates, but in order to
play a game a single estimate had to be adopted for each side. It included the central
decisions (in effect, commander’s intent) that would govern further actions. Once the
students received the relevant staff solutions, they were assigned subordinate commands.
During the game, a record (“History of Maneuver”) was kept on a move-by-move basis.
Ultimately, there was a formal critique, which might be retained in written form. Many
fewer histories and critiques than staff solutions have survived.
A second major source of material is the papers of the OPNAV War Plans Divi-
sion in the National Archives division at College Park, Maryland (NARA II). This file
includes lists of war-gaming material provided to War Plans by the Naval War College,
as well as most (but not all) of this material, and what amounts to a library of lectures
roughly parallel to those that survive at Newport.

Notes 201
 A third important source is the collection of OPNAV and General Board files in
the National Archives in Washington. It provides a few examples of direct influence by
the Naval War College on specific technical and policy decisions. This evidence is largely
lacking in Newport because the archive there seems not to include correspondence from
the president of the Naval War College to CNO.

1. Naval Transformation
1
The recent account of inter-war gaming at the War College, John M. Lillard, Playing at War: Wargaming and
U.S. Navy Preparations for World War II (Dulles, VA: Potomac Books, 2016) describes gaming and some of
its effects on participants, but does not connect gaming experience with wider issues of inter-war U.S. naval
development, such as the rise of carriers. He seems most concerned to debunk a previous, apparently off-
hand, dismissal of the games as ritualistic and unimaginative (or worse). The only other published book that
is concerned with the war-gaming experience is Michael Vlahos, The Blue Sword: The Naval War College and
the American Mission 1919–1941 (Newport, RI: U.S. Naval War College, 1980). The Blue Sword is mainly
concerned with how the War College helped weld U.S. naval officers into a coherent group sharing important
ideas in common. It includes a list of the war games and points to the significance of Operations IV of 1933, the
game that changed U.S. Pacific strategy. Dr. Vlahos is not concerned with other impacts of the games on the
larger Navy. Neither author seems to have looked beyond the archives at Newport to examine the way in which
the War College interacted with the wider U.S. Navy of the time, to the extent that Lillard seems unaware of
the 1930 London Naval Conference, an important watershed for the War College as an advisory organization.
2
For the War Plans Division, see, for example, Edward S. Miller, War Plan Orange: The U.S. Strategy to Defeat
Japan 1897–1945 (Annapolis, MD: Naval Institute Press, 1991), 17. A comparison between officers listed in
the 1 April 1941 Navy Directory and the roster of War College graduates shows only one not listed in the War
College roster: Rear Admiral W. R. Munroe, commanding Battleship Division 3 in the Pacific Fleet.
3
The torpedo story has often been told. The magnetic exploder, which the U.S. Navy called the “Index Device,”
was intended to detonate the torpedo below the enemy’s hull. It would blow in the bottom of a large ship, and it
could snap a smaller ship in half; such detonators (often acoustic in modern torpedoes) are now quite common.
The Royal Navy and the German navy developed their own magnetic exploders at about the same time as the
U.S. Navy and all encountered serious problems. In the U.S. case, the problem was exacerbated by defects in
the back-up contact exploder, which turned out to be too flimsy. The culprits were apparently a lack of pre-war
testing (partly because the exploder was too secret to expose, and partly because testing was too expensive for
a Depression-era fleet) and arrogance on the part of the responsible organization, the Bureau of Ordnance and
its Torpedo Factory. This episode, which drastically reduced the efficacy of U.S. submarines before the prob-
lem was solved in mid-1943, explains why modern U.S. torpedoes are so often tested against full-scale targets.
4
All forms of radio used before the late 1930s could be intercepted well beyond the horizon, sometimes thou-
sands of miles away. During the 1930s, navies adopted high-frequency radio, which could often be detected
an ocean away. The U.S. Navy seems to have led the shift toward it. The Japanese seem to have been the
first to develop effective high-frequency radio direction-finders, which made it possible for them to associate
messages with particular ships and then with particular movements. For its part, the U.S. Navy discovered
this Japanese capability through code breaking, and it adopted security measures to frustrate Japanese ex-
ploitation of U.S. high-frequency radio. In turn, the U.S. Naval Research Laboratory was asked to develop
a high-frequency direction finder—which it failed to do, claiming that no such device could work. OPNAV
insistence on this project was a rare pre-war example of the operational Navy pressing the materiel bureaus (in
this case, the Bureau of Engineering, which was responsible for the Naval Research Laboratory) to do what
they claimed was impossible. For details, see this author’s Network-Centric Warfare: How Navies Learned to
Fight Smarter Through Three World Wars (Annapolis, MD: U.S. Naval Institute Press, 2009), 23 and 268–70.
5
Covert U.S. observation of the 1930 Japanese maneuvers revealed just that. Worse, it appeared that the
Japanese had found effective means of defeating the U.S. plan. Plans had to be completely revised. See the

202 Winning a Future War

official history of the pre-war development of the Naval Security Group (RG 457, Box 108, SRH-355, Naval
Security Group History to World War II, prepared and compiled by Captain J. S. Holwick Jr., USN (Ret.),
June 1971, 71–76). According to the Security Group history, the analysis of the Japanese maneuver was done
by Captain R. E. Ingersoll, who was then in the OPNAV Division of Fleet Training (however, according to
a 1943 Security Group history, he was attached to War Plans). The understanding of the Naval Security
Group was that the framework of the U.S. war plan was changed as a result of Ingersoll’s studies, but no trace
of his analysis or report survived in its file copy of the 1930 report (page 75). Ingersoll had been head of the
Intelligence Section of ONI (Op-16B), which was later the Special Intelligence Branch. According to this his-
tory, the report did not go to ONI because CNO Admiral William V. Pratt did not trust its acting director.
For the 1930 maneuver, the Japanese managed to mobilize their fleet and its reserve so secretly that the U.S.
naval attaché in Tokyo was unaware of it. Secret Japanese mobilization was a feature of many later war games.
Reactions to the 1930 revelations included much more secure radio practices, including the “Fox schedule”
broadcasts (all messages intended for individual ships were broadcast by a shore station, not on a ship-to-ship
basis) and encrypted call signs. See Timothy S. Wolters, Information at Sea: Shipboard Command and Control
in the U.S. Navy, from Mobile Bay to Okinawa (Baltimore: Johns Hopkins University Press, 2013), 142–55.
6
In 1936, Sutherland Danlinger and Charles B. Gray published a book analyzing U.S. and Japanese naval
strategies, War in the Pacific: A Study of Navies, Peoples, and Battle Problems (R. M. McBride & Co.), on just
this basis.
7
The Royal Navy, which had invented the aircraft carrier during World War I, lost control of its aircraft to the
new Royal Air Force on 1 April 1918. British records show that the navy fought hard to regain control through
the inter-war period, finally winning in April 1939 (although the RAF retained control of all shore-based air-
craft). By the 1930s, British official writers considered the U.S. Navy far ahead of theirs. In the late 1930s, for
example, the Admiralty tried to negotiate an exchange of aviation information with the U.S. Navy, and was
much concerned that it might be giving away its technical secrets. The senior Admiralty technical advisor, Di-
rector of Naval Construction Sir Stanley Goodall, wrote that since the Fleet Air Arm was a constant source of
anxiety and U.S. naval aviation was considered the best in the world, the Admiralty should gladly take whatev-
er was offered. The inferiority of British naval aircraft is well known; British records show that those involved
tried hard to avoid admitting it. See the author’s Fighters Over the Fleet (Annapolis, MD: U.S. Naval Institute
Press and Barnsley, UK: Seaforth, 2016) for examples in the both pre-war period and early in World War II.
8
Each of the two carriers displaced about 36,000 tons. The Washington Treaty included a clause allowing
for the addition of up to 3,000 tons in order to provide existing ships with modern protection against torpe-
does and bombs. This was presumably an incentive to navies not to seek to build new ships that would be so
protected. Curtis D. Wilbur, the Secretary of the Navy in office at the time of the Washington Conference,
repeatedly said that there were 3,000 more tons “in the treaty.” His successors were somewhat nervous about
this interpretation, and General Board records reveal unsuccessful attempts to redesign out the 3,000 tons.
Between wars, the carriers were always rated at 33,000 tons with a note that this was exclusive of special
protection against bombs and torpedoes. An innocent reader might imagine that this was a matter of a few
hundred tons or less. All calculations of remaining U.S. treaty carrier tonnage counted the two large ships as
33,000 tonners. By way of contrast, the later Yorktown, which was considered a very satisfactory carrier, dis-
placed 20,700 tons, and the World War II Essex displaced 27,500. However, the designs of the big new carri-
ers were relatively inefficient, so they operated only about as many aircraft as later ships—initially about 100,
but fewer as aircraft grew larger. The Japanese Akagi and Kaga were roughly parallel conversions, of about the
same size. The British converted the “light battlecruisers” Courageous and Glorious into 20,000-ton carriers
with an official capacity of 54 aircraft (fewer in practice). The big U.S. ships were completed in 1927–28.
9
The initial impetus for Japanese carrier aviation seems to have been a 1921 British mission led by Sir William
Francis-Forbes (later Baron) Semphill; this was a logical choice given that the British had recently invented the
aircraft carrier. See for example Mark R. Peattie, Sunburst: The Rise of Japanese Naval Air Power 1909–1941
(Annapolis, MD: U.S. Naval Institute Press, 2001), 18–19. However, Japanese Fleet Admiral Osami Nagano
stated in 1945 that “I think our principal teacher in respect to the necessity of emphasizing aircraft carriers
was the American Navy. We had no teachers to speak of besides the United States in respect to the aircraft
themselves and to the method of their employment…. We were doing our utmost all the time to catch up with
the United States.” This statement is from Scot MacDonald, “Evolution of Aircraft Carriers: The Japanese

Notes 203
Developments,” in Naval Aviation News from October 1962. It was part of a series on aircraft carrier develop-
ment. MacDonald did not give his source, and it was not in the transcript of Admiral Nagano’s interrogation in
volume II of Campaigns of the Pacific War, published in 1945 by the U.S. Strategic Bombing Survey.
10
For an excellent account of the inter-war Navy, see Thomas C. Hone and Trent Hone, Battle Line: The
United States Navy 1919–1939 (Annapolis, MD: U.S. Naval Institute Press, 2013). The end date is 1939
rather than 1941 partly because with the outbreak of World War II in 1939 a powerful Atlantic Squadron,
which later became the Atlantic Fleet, was hived off from the unitary U.S. Fleet.
11
This may seem ludicrous given the World War II achievements of naval aircraft, but wartime experience
actually backs it up. Between 1939 and 1944, only two modern battleships were sunk by air attack (another
was sunk an anchor at Taranto, Italy): HMS Prince of Wales and the Italian Roma. The latter was sunk by a
guided bomb that could not have been delivered by a carrier airplane (the battlecruiser HMS Repulse does
not count as a modern battleship). During the same period, two modern battleships, Bismarck and Scharn-
horst, were sunk by gunfire, as well as the older Japanese Kirishima (a modernized battlecruiser). A fourth,
the Japanese Hiei (a sister ship of Kirishima), was disabled by cruiser gunfire and sunk by aircraft while
immobile. It is, moreover, arguable that the only modern battleship sunk at sea by aircraft using weapons
deliverable by carrier aircraft, Prince of Wales, fell victim to a lucky hit due to poor tactics on the ship’s part
(see the author’s British Battleships 1906–1945 (Annapolis, MD: U.S. Naval Institute Press, 2015) for de-
tails based on modern expeditions to the wreck). The first modern battleship sunk at sea by concentrated
air attack that did not exploit some tactical error was the huge Japanese Musashi, in October 1944—and the
number of aircraft involved would not have been imaginable before then.
12
The law was first stated in 1916. The unstated caveat was that, as the Battle of Jutland showed, there was
a limit to how many battleships could be concentrated usefully. That limit was, however, well above the
strength of the post-1918 U.S. battle fleet.
13
Even with infrastructure, forward basing was unattractive. Between the two world wars, the Royal Navy,
like the U.S. Navy, envisaged a campaign against Japan culminating in a fleet engagement. Ideally, the Brit-
ish would have their fleet in theater at the outbreak of war, and the Singapore naval base was built up for
that purpose. It had facilities far superior to those in the Philippines, but they were not enough. The British
repeatedly avoided basing even their three battle cruisers at Singapore, usually citing the lack of local labor
and poor facilities for the families of ships’ crews.
14
In the fall of 1939, the U.S. Fleet consisted of a Battle Force, a Scouting Force, and a Submarine Force. The
Battle Force consisted of the fleet flagship plus four battleship divisions, four cruiser divisions, two destroyer
flotillas, the carriers, and a force of light minelayers (converted destroyers) and minesweepers permanently
based at Pearl Harbor because they had relatively short range. The Scouting Force consisted of four heavy
cruiser divisions and the patrol planes with their tenders. Both the U.S. Fleet and the Asiatic Fleet had sup-
porting fleet trains (tenders and oilers). Earlier, the ships based at Pearl Harbor had been designated as the
Control Force. By late 1939, many of the carriers and cruisers were operating in the Atlantic. At this time,
there were also a special service squadron operating in the Caribbean, and a force (40-T) left over from the
Spanish Civil War. The ships that earlier had constituted the Scouting Fleet were now part of the Atlantic
Squadron, built around the three oldest battleships.
15
Figures from the Naval History and Heritage Command (NHHC) website, presumably as of the begin-
ning of that fiscal year (1 April 1939).
16
As understood during the inter-war period, the rules of trade warfare required that a raider visit and search
merchant ships to determine whether they were carrying contraband and thus were subject to seizure or
sinking. The crews of any ships sunk in this way had to be conveyed to a safe destination, explicitly not in life-
boats. That is why British merchant seamen from the ships sunk onboard by the German cruiser Graf Spee
were onboard the warship when she fought her last battle off of Montevideo in December 1939. No such rule
could apply to a submarine, which had no accommodation for captured crewmen. Ships owned by the enemy
were automatically subject to seizure or destruction. So were ships in enemy convoys (convoy as a trade-pro-
tection technique was certainly considered in inter-war games and also in inter-war U.S. Navy thinking).

204 Winning a Future War

17
It can, moreover, be argued that the successful U.S. submarine campaign was a variant of the earlier idea
that a U.S. battle fleet would shield the smaller ships enforcing strangulation via blockade. It seems rea-
sonable to argue that Japan was unable to adopt an effective convoy strategy to protect her merchant ships
because the struggle with the main U.S. fleet absorbed the resources involved. The Imperial Japanese Navy
was always painfully aware that should the main U.S. fleet be free of the threat of the Japanese fleet, the
convoys and merchant ships would soon be swept away. That was what the British had argued after World
War I: Namely, that the convoy strategy, which defeated the German submarines, succeeded because the
British Grand Fleet kept the German High Seas Fleet at bay, preventing it from sweeping up convoys. On
one occasion in November 1917 when British intelligence failed to counter a German sortie, two German
cruisers wiped out a convoy. They were far more powerful than the two destroyers escorting it. U.S. officers
had reason to remember because U.S. battleships seconded to the Grand Fleet were then assigned to cover
convoys. Several games envisioned submarine operations—including some unrestricted ones—against Jap-
anese trade as part of a diversionary raiding campaign (Operations I of October 1936 for the Class of 1937
and Operations III [Strategic] of November 1939 for the Class of 1940). Early post-1918 proposals by senior
submarine officers that an unrestricted submarine campaign should be included in any Orange war plan
were rejected on the basis of the way in which the German unrestricted submarine campaign had played out
during the previous war. The manifest illegality of the German campaign had helped tip the United States
into the war, with fatal consequences for Germany. Americans had reacted to the German campaign because
the Germans had been unable to (and had made no effort to) avoid sinking neutral (American) ships.
18
British records show that they were already preparing to ask for an arms-control conference. The four
post-war British battle cruisers (cancelled due to the Washington Treaty) were justified to Parliament as
bargaining chips in such a conference. A British official historian went so far as to suggest that the British
triggered the Washington Conference. See the unpublished Cabinet history, Y. M. Streetfield, “Ten Year
Rule 1919–1929,” CAB 101/295 in the British Public Record Office, Kew. The British record of the Wash-
ington Conference includes remarks by the head of the Japanese delegation indicating intense Japanese pub-
lic sentiment favoring some end to the expensive naval building program.
19
Before World War I, the U.S. Navy regularly asked for scout cruisers, but Congress was willing to authorize
only battleships and large destroyers. This record is slightly muddied by the General Board’s decision a few
years earlier not to ask for larger cruisers, because ten large armored cruisers had already been authorized.
These and contemporary large cruisers were rapidly outmoded as fleet scouts because they offered little or no
speed advantage over “dreadnought” battleships authorized from 1907 on. After that, the only satisfactory
fleet scouts were battle cruisers and small fast cruisers called scouts. Because only three of them had been au-
thorized, the pre–World War I U.S. Navy found itself using large destroyers as fleet scouts, an unsatisfactory
practice. The pre–World War I U.S. Navy considered but rejected the construction of larger scouts, including
battle cruisers. Designs for such ships survive in the National Archives files on U.S. Navy preliminary design,
Entry 449 of NARA RG 19. The big 1916 shipbuilding program, which to some extent triggered the 1921
Washington Conference, included ten large cruisers and six battle cruisers that would have gone some way
toward solving a tactical (but not strategic) scouting problem. However, cruisers had other roles. In the U.S.
Navy, they were needed to beat off enemy destroyer (torpedo) attacks against the battle line. In the Royal
Navy, cruisers were also considered an essential way of protecting global British trade. The British therefore
argued throughout the inter-war period that they needed cruisers both for the fleet and for commerce pro-
tection, hence a much larger cruiser fleet than the U.S. Navy should want. Unfortunately, no U.S. Adminis-
tration could lightly cede superiority in cruiser numbers to the British. To some extent this problem ruined
a 1927 attempt by President Coolidge to extend the Washington Treaty to cruisers. In 1929–30, the British
were willing to accept a form of parity because their economy could not support a program to replace their
large but aging cruiser force. U.S. naval officers painfully aware of the lack of U.S. cruisers used the apparent
imbalance in numbers to promote cruiser construction; Congress approved an eight-cruiser program in 1928
and then a 15-cruiser program. Stephen Roskill, Naval Policy Between the Wars, I: The Period of Anglo-Ameri-
can Antagonism, 1919–1929 (London: Collins, 1968), 505–16, describes how U.S. naval officers broke up the
conference. It seems to this author that the U.S. officers at Geneva feared that any limit on cruisers would en-
courage Congress not to authorize any of them. If that is true, the much smoother progress of the subsequent
London Conference (1930) can be explained by the Navy’s reaction to a Congressional agreement to buy badly
needed cruisers for the U.S. Navy. The need for strategic scouting in any war with Japan was probably obvious

Notes 205
to many U.S. officers, but war gaming would have reinforced that understanding. Roskill emphasizes the fun-
damental difference between the U.S. Navy, which needed large cruisers for strategic scouting in the western
Pacific, and the Royal Navy, which needed large numbers (hence smaller) for trade protection.
20
The Washington Treaty limited the total tonnage of battleships and aircraft carriers. Arms controllers want-
ed to extend those limits to all other types of warships and indeed to other arms. The 1930 treaty extended the
limits on total displacement to cruisers, destroyers, and submarines. The Japanese also managed to extract an
increase in their ratio. That Japan received the short end of the 5:5:3 tonnage ratio (factors of 5 for the United
States and the United Kingdom) and this became a rallying cry for Japanese nationalists. The ratio was vari-
ously justified on the ground that the Japanese would be defending their home waters, whereas any British or
U.S. attackers would lose much of their fighting power en route there; that the Royal Navy and the U.S. Navy
had to defend broader areas, including the Atlantic; and that the British Empire required additional defenses.
Later, it was claimed that nothing short of a 2:1 ratio could have insured U.S. success against Japan, so that the
5:3 ratio sufficed for Japanese defensive needs. U.S. records of the evolution of the U.S. position, which em-
bodied the 5:5:3 ratio, do not explain its origin. It was certainly not the result of gaming or of advice by either
the Naval War College or the General Board. Nor did the War College provide advice during the run-up to
the 1930 conference (or, for that matter, the 1935 conference that framed the 1936 treaty). By the early 1930s,
the U.S. view was that Japan was building every ship allowable under the treaties, whereas the United States
was not. The pre-1939 U.S. naval rearmament acts were justified publicly as a means of creating a modern
U.S. treaty navy, even though the 1938 Vinson-Trammell Act was signed after treaty limits on total tonnage
had been abandoned and Japan had withdrawn from the treaty system in 1937. That this language had to be
used as late as 1938—after Japan had invaded China and had sunk the U.S. gunboat Panay there—shows the
enormous strength of public sentiment that considered naval arms control a guarantee of peace. The treaty
language implied that ships would be replaced when they were over-age in treaty terms. In 1936, when the
first new U.S. battleships were approved, U.S. battleships were mostly over-age, but existing carriers were not.
21
This is evident from the Pratt papers at the War College and also from General Board papers on arms
control at the National Archives. Admiral Pratt was the General Board’s advisor on naval arms control
during the conference.
22
The Naval War College archive (RG 8) contains a 1935 paper replying to a CNO request for advice as to
what building program to pursue once the existing treaties expired at the end of 1936 (“Building Program
on Expiration of Washington and London Treaties,” UNC 1935–26, RG 8 Series 1, Box 37, Folder 8. CNO
(Op-12) sent a request for advice to CinC U.S. Fleet and to the president of the War College dated 1 April
1935, and on 6 July repeated his request. CNO’s assumption was that all restrictions would be lifted. In fact,
the London Naval Treaty of 1936 maintained (in some cases reduced) limits on individual ship size and arma-
ment, but it eliminated all restrictions on total numbers. The study was sensitive because there was enormous
political pressure to maintain and even to tighten overall limits (the U.S. delegation came to the 1935 London
Conference carrying just such a policy, even though by this time it was obvious that the slide toward World
War II had begun). The War College reply, signed by its president, Admiral E. C. Kalbfus, was dated 1 Octo-
ber 1935. It contrasts strongly with the game-based analysis of future cruiser construction completed in 1930,
and described below. The 1935 War College letter emphasizes the political requirement that any program had
to be acceptable to the public, and therefore framed it in terms of the 5:5:3 ratio adopted at Washington in
1921 and embodied in the later London treaty. It had to be accepted that the public would oppose any building
program that might be construed as the beginning of an arms race, although the public might be persuaded to
accept something more than 5:5:3 if that could be shown to be necessary. Since ratios had become acceptable,
the War College suggested that in future they should apply to numbers of ships rather than to tonnage, as that
would make it possible to increase the efficiency and strength of the Navy. The president of the War College
did point out that although a 5:5 ratio with Red was satisfactory, 5:3 (against Orange) was inadequate for a
trans-Pacific offensive; a ratio of at least 2:1 was needed. Since the public would reject any such ratio, the best
the War College could offer was to seek an effective ratio of that sort through superiority of individual ships.
Whether or not CNO agreed, that approach became impossible once the 1936 London Naval Treaty main-
tained limits on individual ship size and, indeed, reduced the allowable size of cruisers. The existing numbers
of foreign battleships precluded any great increase in numbers, but by 1937 the United States would have
seven over-age battleships, and it might be possible to replace five of them. The War College did propose that

206 Winning a Future War

three of the five should be battlecruisers (a type the United States lacked), matching the three Japanese battle
cruisers, “in view of the long period of cruiser and commerce warfare that will probably precede a decisive en-
gagement. . . . If for any reason our air forces were unable to operate in an area into which enemy battle cruisers
had projected themselves, the latter would be a formidable menace to such light forces, convoys, and shipping
of ours as might be there…in effect [this exchanges] a tactical advantage during a possible general engagement
for strategical and tactical advantages on perhaps numerous occasions during the pre-battle period.” To some
extent this reflected war-gaming experience, in which most of the action during a trans-Pacific advance oc-
curred before any main fleet battle could occur. It may explain the decision to adopt high speed for the first
U.S. post-treaty battleships of the North Carolina class. The president of the War College certainly did not
envision high speed for any other post-treaty battleships: He would have been satisfied with 21 knots, as in
existing U.S. battleships. As for carriers, current Orange numbers (four built and two building) would justify
ten U.S. carriers on the basis of the 5:3 ratio. At the time, the United States had four built, two building, and
one authorized, the existing four including the obsolete Langley, which would be reduced to seaplane tender
status when the authorized carrier Wasp was built. All the new carriers should be of the heavy offensive type
(Yorktown class). During the rest of the pre-war period, the United States would lay down one more carrier
(Hornet), a repeat Yorktown. Note that carriers were listed sixth in priority, after battle cruisers, battleships,
submarines, destroyers, and light cruisers, and before minelayers and auxiliaries—hardly the sort of strong
endorsement implicit in earlier War College comments on building programs. The War College paper includ-
ed a recommendation for 12 squadrons of patrol planes (16 aircraft each) at the outset of an Orange war, many
more being needed later on. This figure was based implicitly on gaming experience rather than on a numerical
comparison with the Japanese (Orange). However, the discussion of cruiser numbers appended to the letter
was decidedly not based on gaming. Instead, it was based on a breakdown of the U.S. fleet, plus some ratios be-
tween numbers of major units and desired numbers of cruisers. For example, it cited Admiral Jellicoe’s World
War I precept that there should be five cruisers for every three battleship (giving 25 cruisers), one cruiser as
flagship of each destroyer flotilla (two), and two for each carrier (12 for six carriers). Past studies were cited
as the basis for a requirement for another 21: nine to protect focal areas (trade protection), six to protect lines
of communication, and six to attack enemy trade. Other figures could be derived from numbers the Japanese
already had, this total being 53 compared to the 60 obtained from the breakdown of tasks.
23
Discussions of U.S. ships and the policies that created them are based mainly on the author’s series of illus-
trated design histories of U.S. warships: carriers, battleships, cruisers, destroyers, submarines, small combat-
ants, and amphibious ships and craft, all based on formerly classified U.S. official records, including records
of policy makers. All were published by the U.S. Naval Institute Press. I have also published a parallel series
of British design/policy histories covering carriers (and their aircraft), battleships, cruisers, destroyers, and
lesser escorts, both before, during, and after World War II, again based on official formerly classified papers.
24
The fatal hit, which was the first the Japanese made, was near the point at which the ship’s outboard propel-
ler shaft emerged from her hull. In retrospect it seems that this hit in turn was made because the ship’s com-
mander waited too long to begin the standard defensive measure of combing the torpedo tracks. It would not
have been logical for the pre-war Navy (and its competitors) to abandon the battleship as a capital ship—cer-
tainly not before 1943, when the Germans demonstrated that large land-based bombers armed with guided
bombs could sink the Italian Roma and disable the British Warspite, Roma being a fully modern battleship.
Sinkings at Pearl Harbor do not count because the ships sunk were not maneuvering and also because they
were not at a high degree of readiness, with many watertight doors open. The fleet at anchor was subject to
a much higher percentage of hits than would have been the case in the open sea, and the lack of watertight
integrity magnified the effect of the hits. The author has elaborated the Prince of Wales argument, using
modern examinations of the wreck, in his book British Battleships 1906–1945 (Annapolis, MD: U.S. Naval
Institute Press and Barnsley, UK: Seaforth, 2015). The sinking of the accompanying battlecruiser Repulse
was much less of surprise; she was relatively weakly protected, and had a weak anti-aircraft battery. British
records show that the sinking of the Prince of Wales was a great shock, as she was a modern ship whose tor-
pedo protection should have been proof against the Japanese torpedoes.
25
Games certainly taught U.S. officers that carriers could be sunk by fast battleships or even by cruisers,
particularly if they did not use their aircraft effectively. This helps explain why Admiral Halsey did not
detach his battleships when he took his carriers north at Leyte Gulf. It also raises the question of why the

Notes 207
commander of the carrier HMS Glorious allowed his ship to be attacked by the two German battleships
Scharnhorst and Gniesenau, which sank her off Norway in 1940. The implication of the loss of the British
carrier would seem to be that the British were unable to teach their officers, particularly carrier commanders,
about the facts of fleet life.
26
The connection between the air-defense role in the passage through the Mandates and the anti-air capability
of the new destroyers is explicit in General Board hearings that established their basic design features. The
dual-purpose guns were also justified by their ability to deliver plunging fire in support of troops landing on
fortified islands, another requirement underpinned by gaming. The new high-angle 5-inch/38-caliber guns
appeared for the first time in the Farragut class, the first post–World War I destroyers, which were designed
toward the end of the period of War College influence on warship designs. There is no direct evidence that
the heavy torpedo batteries of later destroyer classes resulted from Naval War College arguments, but their
adoption did coincide approximately with the War College’s accepted recommendation that torpedo tubes be
removed from cruisers. That left destroyers as the only remaining source of torpedoes in night surface actions.

2. The Naval War College and Gaming

1
See, for example, Miller, War Plan Orange, 15. An account of the inter-war General Board, emphasizing its
role in formulating official U.S. Navy policy and also in determining ship characteristics, is John T. Kuehn,
Agents of Innovation: The General Board and the Design of the Fleet That Defeated the Japanese Navy (Annap-
olis, MD: U.S. Naval Institute Press, 2008). Dr. Kuehn emphasizes the board’s continued espousal of a
mobile fleet repair facility, particularly one or more towable floating dry docks. The contention of the present
work is that the board became interested in such facilities because war games showed how vital they would
likely be. For details of General Board discussions of ship characteristics, see the author’s series of illustrated
design histories of various types of U.S. warships, all published by the U.S. Naval Institute Press.
2
This account of the 1908 controversy is based on John B. Hattendorf, B. Mitchell Simpson III, and John R.
Wadleigh, Sailors and Scholars: The Centennial History of the U.S. Naval War College (Newport, RI: Naval
War College Press, 1984), 61–64. As a junior officer, Sims became a disciple of British gunnery innovator
Captain Percy Scott, RN; his own work on gunnery in the U.S. Navy led President Theodore Roosevelt to
make Sims inspector of target practice. Sims was an early supporter of the all-big-gun battleship espoused by
British First Sea Lord Admiral Sir John Fisher. His reports in the naval attaché series at the U.S. National
Archives (RG 38) suggest that Sims enjoyed special access to the Royal Navy—to the revolutionary battle-
ship HMS Dreadnought, for example. He was widely considered a strong anglophile. Before World War I, he
made a speech before a British audience promising that if there was a war, the United States, as an offshoot
of Britain (part of “greater Britain”) would come to the aid of the British.
3
That is, the first U.S. battleship with all big guns and steam-turbine propulsion (the earlier South Carolina
had the guns, but not the turbines).
4
As indicated in the 1913 Admiralty report of the committee assigned to consider the organization and
training of the Royal Naval War College, held by the Admiralty Library in Portsmouth (Royal Naval His-
torical Branch). It includes an account of the U.S. Naval War College as it then existed, written by Captain
C. F. Sowerby, RN (page 42 of the British report). He began with the three-month summer conference,
“which they are very anxious to keep [as such]. To this conference are sent all sorts of questions by the Navy
Department. War plans are here discussed, elaborated, and in all probability practically settled. Schemes
for maneuvers, strategical and tactical questions are brought up, in fact everything in connection with the
handling and movement of the fleet. Admirals commanding are in constant communication with this confer-
ence. Lectures are given on naval history, strategy, tactics, international law, gunnery, torpedo, engineering
etc. A great stress is laid on the intimate connection of gunnery and tactics, thus the design of new ships
forms a topic of discussion. War games are of course played. This course is not compulsory, but officers are at
times told off to attend. This college is commanded by an Admiral, who has a staff of generally one captain,
five or six commanders or captains.” The U.S. Navy had, as far as was known, no war staff, and hence no
chief of war staff so-called “and it would seem the admiral commanding the War College was the nearest

208 Winning a Future War

approach to it they have; he was frequently at Washington.” At this time the Secretary of the Navy had four
aides, including an aide for operations whose office would soon be transformed into the Office of the Chief of
Naval Operations. Captain Sowerby wrote that the aide for operations was practically in the position of the
Royal Navy’s First Sea Lord, with the War College and the director of naval intelligence subordinate to him.
Sowerby added that a few officers, generally commanders and lieutenants, stayed on at Newport through the
winter working on various problems, attending lectures, and writing essays; as of a year earlier (1912), it was
hoped that those who showed special aptitude would be allowed to stay on another year. “Whether this was
with the idea of making them a war staff is not known.” This report is not otherwise identified. A typescript
version is designated T.21605.
5
This was more radical than it may appear. The old Bureau of Navigation was responsible for naval person-
nel, but its chief was also the Secretary’s primary advisor for operations. ( The bureau was also the conduit
for inputs from the fleet. For example, before 1909 its files show requests (which were often accepted) for
modifications to U.S. submarines. The result of the reform was that the Bureau of Navigation was reduced
to personnel functions, although for many years the Bureau of Naval Personnel was styled BuNav rather
than, as later, BuPers.
6
Spector, Professors of War, 142–43, and Hattendorf et al., Sailors and Scholars, 82–84. Daniels also appar-
ently rejected Fiske’s arguments for increased investment in naval aviation. Fiske had proposed the torpedo
bomber in 1912.
7
These were hardly junior officers. One was Captain Sims, who had helped ignite the 1908 battleship con-
troversy. The other three were on the War College staff: Captain J. S. McKean, Commander Yates Stirling
Jr, and Captain E. H. Ellis, USMC. Ellis was later a pioneer exponent of amphibious warfare in the Pacific.
McKean would later head the OPNAV material division and ended up as a Vice Admiral. Stirling was later
rear admiral and chief of staff of the U.S. Fleet. At this time, the War College president was a captain. Of
six officers in the second long course (1912–13), two were in that summer conference (1913), which seems to
have been a very limited affair (a total of only five officers). The full two-committee summer conference was
revived in 1914. The 1914 academic year was the first not to feature a summer conference, and also the first
to feature an enlarged long course (13 officers).
8
According to Hattendorf et al., Sailors and Scholars, 79–80, the first two long courses culminated in the
four-month summer conference. This was not the change ordered by Secretary of the Navy Josephus Dan-
iels, who succeeded von Lengerke Meyer after Woodrow Wilson became President. After a June 1913 visit,
he ordered the college to provide three different courses. An elementary course (three weeks) was intended
to teach junior officers tactics. A four-month “preparatory course” would add the first part of the long course
(Hattendorf et al., Sailors and Scholars, 80, see it as the short course or renamed summer conference). The
long course became the 16-month War College course, consisting of a four-month preparatory course fol-
lowed by 12 months of advanced work. There was also to be a correspondence course. Each of the two longer
courses was to have 20 students, and faculty was to be drawn from graduates. Ronald Spector, Professors of
War: The Naval War College and the Development of the Naval Profession (Newport, RI: Naval War College
Press, 1977), 125–26. In July 1913, Daniels personally ordered 50 Atlantic Fleet offices to attend a two-week
elementary course. The War College considered this too short, and President Rodgers replaced it with a
correspondence course. It was also difficult to arrange two four-month preparatory courses, as in each case
officers had to be detached temporarily from ships that were thinly manned. Rodgers asked Daniels to issue
a general order eliminating the two short courses in favor of the single long course. Aid(e) for Operations
Rear Admiral Fiske refused to endorse it, as Rodgers would soon be relieved by Rear Admiral Austin M.
Knight. Knight then endorsed Rodgers’s draft, and it was issued in January 1914. The new standard course
would last 12 months, and two such courses (15 students each) would be enrolled each year: one in January
and the other in July. Hattendorf et. al., Sailors and Scholars, 80–81. The correspondence course was formally
established by a 1 April 1914 general order.
9
During this period, American destroyers were seen primarily as ocean-going torpedo craft, which would
seek out enemy capital ships and attack en masse. At this time, the U.S. Navy had the most powerful de-
stroyer torpedo batteries in the world. Sims remarked later that although the theory was well understood,
no one in his command had any idea of just how the destroyers were to find their targets or of how they were

Notes 209
to attack. This was not a unique problem; a few years later, the Royal Navy found that it also had no idea how
destroyers operating independently were to find and attack major enemy units, particularly at night. Sims’s
remarks were reprinted as “Naval War College Principles and Methods” in U.S. Naval Institute Proceedings,
March–April 1915, 390–91. For his earlier skepticism, see Hattendorf et al., 76–77. Moreover, Hattendorf
et al. point out (page 88) that Sims was allowed to choose a staff (previously the destroyer flotilla had not
had one).
10
Sims was appointed president on 16 February 1917 and promoted to rear admiral on 19 March.
11
The War College program for the January and July classes of 1917 survives as file UNT in NWC RG 8
Series 1. It is mainly a diagram of which problems were to be solved. In January, the new class would have a
week of reading, after which it would have two problems (I and II), then a scouting and screening exercise,
followed by tactics classes at the start of February, followed by Problem III and then scouting and screening
again, then more tactical classes in March, followed by Problem IV and then Maneuver I (presumably a larg-
er-scale game). After that, tactics, maneuvers, and solutions would be interleaved. The curriculum explicitly
includes a logistics game (VIII) and at least three chart maneuvers (IX, X, and XI). Between January and
July, time would be divided between strategy (101 periods), tactics (109), scouting (68), and logistics (14),
with similar but slightly different apportionment between July and January.
12
After the war, the U.S. Navy published the papers of the Planning Section: Officer of Naval Intelligence
Historical Section, American Naval Planning Section, London (Washington, DC: GPO, 1923). According to
the preface of this book, contributed by Captain D. W. Knox, all planning for the U.S. naval force in Europe
initially came from Washington. Sims protested that he needed a planning section in London because it had
much better access to timely information, not only about Allied naval activities and requirement, but also
about the enemy. According to Knox, the London planning group was authorized by CNO Admiral Benson
after he visited London in November 1917. At that time, the Admiralty staff was being reorganized and a
planning section had just been created. In 1923, Knox was in charge of the Naval Library and Records and
of the U.S. Navy’s Historical Section; he was the direct predecessor of later chiefs of the Naval Historical
Center and of the current Naval History and Heritage Command. He was one of the first two members of
the London planning section.
13
The planning group was initially a section within the Operations Division of the Royal Navy’s War Staff.
This staff had been created to meet a political demand. In 1911, Prime Minister Herbert H. Asquith asked
his army and navy chiefs what they expected to do in the event of a war with Germany, which seemed entirely
possible. He rejected the “secret” naval war plan proposed by his First Sea Lord, and he accepted the view of
his Secretary of State for War that the Admiralty needed reform in the shape of a war staff like the army’s.
No one pointed out that the navy already had a staff in the form of the naval intelligence department, and
the new war staff was formed almost entirely from members of the earlier department. This demand was
the immediate basis for appointing Winston Churchill to the Admiralty; unfortunately for the historian, a
case can also be made out that the entire drama was window-dressing intended to move Churchill out of the
really sensitive position of Home Secretary (in charge of internal security), in which he had been less than
successful. The move led to reorganization of the Royal Naval War College to support the new war staff.
In army terms, the war staff was intended to help implement command decisions, not to ask wider-ranging
questions, and a war staff education on the army model was very much about how to turn high-level deci-
sions into detailed orders. Just as the army had its own staff college to produce staff officers, the navy was to
use its Royal Naval War College (created in 1900) to produce officers for the new war staff. That was quite
different from the earlier concept that, like Newport, the Royal Navy War College was intended to produce
better general-purpose officers. The origin and role of the pre-1911 Royal Navy War College were described
in a 1909 paper written after a visit by Rear Admiral R. P. Rodgers, USN, at that time Naval War College
president, in the NWC Archive (RG 8, Box 6, coded DNT). Rodgers reported that the war course having
proved its value, it was broadened and the War College made an adjunct of the naval intelligence department,
“which in England largely fills, under the First [Sea] Lord, the duties of a General Naval Staff.” In 1906, the
war course was renamed the Royal Naval War College and detached from the Greenwich Naval College to
Portsmouth. Under its first president, Captain H. J. May, RN, the War College used its gaming capability
to answer questions such as the relative value of speed and armor. Its answers were not entirely accepted, and
it seems that such issues were not raised after May died in 1904. The answers and the reactions are in ADM

210 Winning a Future War

1/7597, “Exercises Carried Out At the Royal Naval College, Greenwich,” dated 7 May 1902 (the war college
conclusions were also printed as naval intelligence department papers). Later evidence of pre-1914 British
naval tactical development seems to reflect the experience of full-scale exercises and post-exercise analysis;
war gaming or table-top exercises are never cited in the surviving papers. There is, however, evidence that the
Royal Navy continued to war game, as some pre-1914 rules have survived.
14
See William N. Still Jr., Crisis at Sea: The United States Navy in European Waters in World War I (Gaines-
ville, FL: University Press of Florida, 2007), 43. Still first quotes Admiral Lord David Beatty, CinC Grand
Fleet, complaining in April 1917 of the need for a more effective executive body including a planning body,
but he then emphasizes the role of Lloyd George in forcing reorganization. He states further that Lloyd
George identified planning as the weakest feature of the Royal Navy. It seems likely that the planning group
was created in response to outside pressure rather than to a demand from within the Royal Navy. The ev-
idence is that U.S. officers attached to the London planning section of Sims’s staff considered the British
planners weak and unwanted.
15
CinC U.S. Atlantic Fleet to Secretary of the Navy (Operations), 11 October 1917, page 6 of a general re-
port by fleet commander Admiral H. T. Mayo of a visit to England and France in August–September 1917
in OPNAV Confidential file C-47-9, NARA RG 80. “Until very recently there has been no planning section,
or was there any definite body of men charged with the function of looking ahead, or even of looking back
to see wherein lay the causes of success or failure, nor any means of furnishing the heads of the Admiralty
with analyses and summaries of past operations in order that decisions as to continuing old operations or
undertaking new ones might be reached with a due sense of “perspective” both as to past operations, and as
to the co-ordination of new operations in a general plan.”
16
This was not an idle question. In November 1917, two German light cruisers wiped out a British convoy
headed to Norway. The last German fleet sortie, in April 1918, was intended as a convoy raid. However,
German intelligence was poor: There was no convoy to raid. The planning section idea was imaginative, but
the German ships did not have the required range, nor were the Germans willing to make what amounted to
a suicidal gesture. However, ocean raiding of exactly this type was of considerable interest to the Germans
before World War II, and it was practiced during the war.
17
Still, Crisis at Sea, 45. Sims first requested a formal planning staff in October 1917. Secretary of the Navy
Josephus Daniels opposed the idea in line with his general rejection of the creation of any kind of general
staff. However, he referred the question to Admiral Benson, who included it in his list of items to discuss in
England in November 1917. Benson endorsed the idea, arguing (presumably based on Sims’s ideas) that a
U.S. planning group could sell U.S. ideas to the Admiralty. It was assigned office space adjacent to that of the
British planning staff. The first two members of the planning group were Captain Frank H. Scholfield and
Captain Dudley W. Knox. Comments by U.S. officers in the classified OPNAV records (RG 38 in NARA)
suggest that Still greatly overrated the power of the British planners.
18
See for example Hattendorf et al., Sailors and Scholars, 112–13. The War College resumed its pre-war prac-
tice of running two overlapping one-year courses, with classes entering in June and December. Beginning
in 1920 with the Class of 1921 (also described as the Course of December 1921), there was a single senior
course each year, typically consisting of about 30 officers, many of them captains (later some rear admirals).
The principal departments were Command, Strategy, and Tactics, Command being headed by the college’s
chief of staff. The college also ran a correspondence course. Both the Strategy and Tactics departments ran
their own games. Beginning in 1924, the college added a junior class (lieutenant commanders and lieuten-
ants) with its own staff and, in some cases, its own games. In theory, the junior class was intended to prepare
officers to command major ships; the senior class was intended to prepare them for fleet and major fleet or-
ganization command. This book concentrates on the senior games. The NWC archive includes some junior
class games, the avowed purpose of which is often to familiarize officers with the character of different ships
in the fleet.
19
According to a 1930 lecture on OPNAV organization, the original legislation provided for 15 officers to
assist the CNO. In December 1918, a total of 267 officers were assigned to OPNAV. Given the scope of
OPNAV tasks, an act of 12 August 1918 provided for an Assistant CNO (ACNO). OPNAV contracted

Notes 211
post-war, so that on 1 November 1930 it had 84 officers exclusive of those under instruction or on temporary
duty. Lecture, Rear Admiral W. H. Standley (ACNO), “The Functions of the Office, Chief of Naval Oper-
ations,” delivered at the Army War College on 5 September 1930, in OPNAV War Plans Division records,
NARA II.
20
Sims, “The Practical Officer,” 2 December 1919, RG 12, Box 1, Folder 20, NHC, 25-6.
21
Address to the Class of June 1919 (the first post–World War I class), NWC Archives, RG 8, file U-N-T
1919/136 in folder UNT 1919–1920.
22
Sims’s triumphant return to Newport is described in Hattendorf et al., Sailors and Scholars, 116–17. Sims’s
recommendations were reviewed and cut back by CNO Admiral W. S. Benson, who had worked with and
against Sims during World War I.
23
Hattendorf et al., Sailors and Scholars, 122–25, describe the controversy. Presumably Sims’s underlying
theme was that Daniels had unduly politicized the Navy, and that before the outbreak of war he had far too
enthusiastically followed President Wilson’s pacifist policy. It is difficult to separate Sims’s attack on Daniels
from the larger fight between the insurgents who had led the Navy in European waters and those who had
shaped the fleet before the war.
24
In theory, because it directly advised the Secretary of the Navy, the General Board was lead naval agency
during the various arms control negotiations of the inter-war period. It might be imagined that because
the War College alone had the technique to compare different fleets, it should have been deeply involved,
but that does not seem to have been the case. The General Board framed a U.S. Navy position prior to the
Washington Conference, but it was overridden by President Harding and the State Department. Important
clauses in the 1930 London Naval Treaty, particularly provision for the flight-deck cruiser, were the personal
responsibility of Rear Admiral William V. Pratt, who was U.S. naval advisor. He may have been inspired in
part by U.S. needs demonstrated by war gaming. U.S. CNO Admiral William H. Standley seems to have
been instrumental in negotiating the 1936 London Naval Treaty (the original U.S. negotiating position,
developed by the State Department, was completely overcome by events). War College analysis may have
helped the General Board head off a British attempt in 1927–33 to cut the size of battleships to cut their cost.
The British attempt is evident in British records, in the treaty records of the General Board, and in General
Board file 420-6 (battleships).
25
The existence of the struggle can be inferred from intense controversies over what may now seem minor
issues of ship design. The General Board won the fight over cruisers in 1920; its terms are clear from the
General Board cruiser file (420-8). General Board files (420-2) include Sims’s recommendation for the U.S.
1918 program. It is essentially a list of the types of ships then being built for the Royal Navy, adapted to U.S.
weapons and therefore somewhat enlarged (they included a U.S. equivalent to the then-revolutionary fast
battleship/battlecruiser Hood). The cruisers would have been equivalent to British fleet cruisers of the time,
the C and D classes. Sims and his London planners brought the cruiser idea home with them. Meanwhile the
General Board espoused big long-range cruisers armed with 8-inch guns for operations in the Pacific. The
issue had an unstated international dimension. In 1919–20, the British were trying to avoid a costly increase
in cruiser size. Sims and his planners were, in effect, supporting them—and, it might be argued, working
against the new kind of war plan the U.S. Navy required for the Pacific. The General Board won this fight.
It lost the next major fight, over battleship design in the mid-1930s. The fact of the creation of the OPNAV
War Plans Division clearly shows the loss of the General Board’s war planning role. The power of the in-
ter-war General Board is probably somewhat exaggerated for historians by the high quality of the records
it left (now in the National Archives in Washington), compared to the much sparser and less-organized
records left by various parts of OPNAV.
26
Inter-war presidents were Rear Admirals William S. Sims (16 February–28 April 1917 and then 11 April
1919–15 October 1922); Clarence S. Williams (1 November 1922–5 September 1925); William V. Pratt
(5 September 1925–17 September 1927); Joel R. P. Pringle (19 September 1927–31 May 1930); Harris
Laning (16 June 1930–14 May 1933); Luke McNamee (3 June 1933–15 June 1934); E. C. Kalbus (15 June
1934–15 December 1936 and 30 June 1939–16 June 1942); and Charles P. Snyder (2 January 1937–27 May

212 Winning a Future War

1939). Pratt was CNO between September 1930 and June 1933, having served as CinC U.S. Fleet between
appointments as president of the War College and CNO.
27
The extent of its information-gathering is evident in the surviving Research and Intelligence record group
(RG 8) in the War College Archive. Most of the files are in the same form as the ONI (largely attaché) series
held by NARA in Washington in RG 38. There was a rub. Apart from code-breakers, the inter-war U.S.
Navy had no source of secret intelligence about other navies. That is evident from the ONI and War College
intelligence files, for example in the very limited information gathered about the main potential enemy, Ja-
pan. This problem is indicated indirectly by the pride with which the code-breakers pointed to the one piece
of technical intelligence they obtained, the trial speed of the Japanese battleship Mutsu.
28
This change is evident in the War College’s roster of students and instructors.
29
These comments surfaced when the U.S. Naval Institute published the 1942 edition of Sound Military
Decision, which it called the “Green Book,” in its “Naval Classics” series (October 1992). According to Hat-
tendorf et al., Sailors and Scholars, 155–61, when he became President of the War College in 1934, Rear
Admiral E. C. Kalbfus devoted considerable energy to total revision of the pamphlet on the “Estimate of the
Situation” on the theory that its logic was the key to successful naval decision-making. At this time the Navy
was beginning to expand, diluting the War College’s influence except through its correspondence course. It
seems arguable that in Kalbfus’s view the college’s publications thus became much more important, since
most of the fleet could not benefit from actual gaming. Hattendorf et al. see Sound Military Decision as the
only handbook for decisions available to World War II naval officers. This is not a point made by Hattendorf
et al. in their description of Kalbfus’s effort.

3. War Gaming and War Planning

1
Unsigned booklet, “The Naval War College,” 10 June 1932, NWC RG 4, Box 55, folder 1. The booklet
was intended as an introduction to new officers. It was probably written by the War College President, at
that time Rear Admiral Harris Laning. Laning made this point particularly clearly in his memoir (Harris
Landing with introduction by Mark Russell Shulman, An Admiral’s Yarn [Naval War College Historical
Monograph No. 14, 1999]), 262: “student officers train for high command in war by actually applying their
knowledge of fighting to war situations; in war, no one condition is quite like another, and what might win in
one case could bring complete disaster in another.”
2
For example, the Naval War College archive contains a lecture delivered before the 1913 Summer Con-
ference by Captain William McCarty Little, “The Philosophy of the Order Form,” in RG 8, Box 20, coded
XCDO. At the time, McCarty Little was an important figure in the War College and he was instrumental
in having the first four students assigned to the long course. His point is that the order form describes a
complete course of action, hence can guide subordinates developing their own orders. This order form is
equivalent to the later “Estimate of the Situation.”
3
Spector, Professors of War, is particularly critical of what he sees as the narrowing of the War College in
his final chapter, “Epilogue: The Long Twilight,” which covers essentially the same period as this book. He
argues that, by the early 1930s, strategic problems were merely tests of isolated parts of larger plans (the ac-
count here of the impact of the “big game” of 1933 suggests the opposite). An alternative view would be that
this shift reflected the maturation of the Orange War Plan, the basic ideas having been worked out.
4
Spector, Professors of War, 117–18, credits introduction to Captain McCarty Little and his younger col-
leagues. He considers the origins of the system obscure, but does note a system of the same name used by the
Army War College, extensive notes having been taken by CDR William L. Rodgers during his tour there in
1907–1909. Spector credits Captain McCarty Little with taking up an English translation of German Gen-
eral Otto von Griepenkerl’s Letters on Applied Tactics as the basis for a new type of instruction. He formed
a committee with Lieutenant Commander C. T. Vogelgesang and Marine Major John H. Russell (a future
Marine Corps Commandant). Little made the important leap to a two-sided game.

Notes 213
5
The opposing argument is that too much initiative can be deadly. In framing instructions before Jutland,
Admiral Sir John Jellicoe was clearly worried that a battle might become a melee, in which case his own ships
might find themselves shooting at each other. That had nearly happened at Tsushima in 1905. Japanese
commander Admiral Togo had divided his fleet into two divisions. In the smoke and confusion, one nearly
engaged the other. We might now say that this clearly depends on how well everyone involved has situational
awareness. In an army, an over-energetic junior officer might advance so far that his flanks would be exposed,
and an alert enemy might well exploit that. A transition to objective-oriented orders remains a current goal
in army training.
6
The British ships were the large “Town”-class cruisers, which they felt compelled to build instead of the
far more affordable Leander class. The U.S. and Japanese ships were the Brooklyns and the Mogamis. As re-
counted later, the U.S. Navy decided to build larger rather than smaller cruisers on the basis of War College
analysis of what the other two navies (particularly the Japanese) were likely to do. The War College analysis
is described in greater detail below. British reasoning (and shock) is evident in various Admiralty documents
I have used for my history of British cruisers. British decisions on all types of major surface ships during the
inter-war period seem marked by wishful thinking and mirror-imaging. This is particularly obvious in the
discussion of what battleships the Japanese were likely to build once they withdrew from the treaty system
(announced in 1934). The British badly wanted to limit the size of future battleships so that theirs would
be affordable. They converted this wish into a totally false assumption that the Japanese could not go much
beyond the earlier 35,000-ton limit. The reality was the 62,000-ton Yamato. The War College technique was
an antidote; at the least, the War College would have asked what alternatives the Japanese might choose.
Admittedly the U.S. Navy also did not guess the Japanese would go as far as they did, but in 1934 there was
a serious attempt to guess what an unlimited Japanese battleship would look like.
7
Spector, Professors of War, 119, claims that the “applicatory method” was revolutionary in that it made the
creation and use of doctrine possible.
8
Jellicoe’s “Grand Fleet Battle Orders” survive in several files in the Archives at Kew, in versions reproduced
post-war for instructional purposes (ADM 116/1341–1343). These orders were supplemented by “Grand
Fleet Gunnery Orders” and “Grand Fleet Torpedo Orders,” as well as by “Grand Fleet Secret Orders.” The
apparent bulk of the “Grand Fleet Battle Orders” was increased by inclusion of brief comments on current
experience, for example, on the Battle of Dogger Bank. Jellicoe’s successor Admiral Beatty renamed the
orders “Grand Fleet Battle Instructions,” but they were still the sort of detailed orders Jellicoe had issued. It
is not clear whether the Royal Navy ever issued an overall tactical handbook between wars, although it did
publish a document called “The Tactics of the Gun” (no copy of which has surfaced) and some other detailed
handbooks. They contrast with the standardized formations and general tactical instructions issued by the
Office of Fleet Training and by the U.S. Fleet itself between wars. After Jutland, the Royal Naval College
tried and failed to find a tactical solution to the German turn-away. Gaming would probably have shown
that the notion of a submarine trap, which so occupied Jellicoe and which resulted in the construction of the
fast British K-class submarines, was impossible due to the limits of communications. Remarkably, the idea
that the British would lead them into a submarine or mine trap appears in the 1914 German tactical orders,
which the British captured and distributed in translation in October 1914. The 1915 version of these orders
has survived in a collection of German tactical documents in the Admiralty Library in Portsmouth, and the
January 1914 version (in translated printed form) is in the Public Record Office.
9
Jellicoe’s view that decentralization required training is evident in a memorandum on plotting as a basis
for divisional tactics, prepared by Captain F. C. Dreyer on the eve of war, a copy of which survives in the
Backhouse Papers in the Royal Naval Museum, Portsmouth. Backhouse, who was later First Sea Lord, was
Jellicoe’s flag secretary. Probably the strongest attack on the social system that placed obedience and def-
erence to authority uppermost is Andrew Gordon’s Rules of the Game: Jutland and British Naval Command
(London: John Murray, 1996).
10
It is striking that only one of the surviving U.S. games (Tactical Problem VI for the Class of 1922) sim-
ulated Jutland, although many classes heard or gave lectures about the battle. The British inter-war senior
course apparently devoted two weeks each year to Jutland, showing officers what had happened and what
would change using modern methods and weapons. The great difference between Jutland and most of the

214 Winning a Future War

big battles simulated by the War College was aircraft. Jutland involved a single British scout airplane, which
had no real impact because its radio report never reached the flagship. Inserting mass air forces would have
changed the battle out of all recognition. Records of the British presentation have not surfaced, but it seems
unlikely that aircraft were a major feature of the modernized version of the battle. In the British case, the
sole game played by the senior war class was a one-week campaign game simulating a war in the Far East.
Again, no detailed accounts seem to have survived, so it is not possible to say how much aircraft figured in
this game. It was probably a simulation of the movement of the fleet to the Far East as a prerequisite for the
expected fleet action there. Since the British fleet would have steamed through the Indian Ocean, it would
have been subject to submarine attack and to submarine-laid mines, but not to the sort of large-scale attacks
that figured in U.S. war games simulating the progress of the U.S. fleet to its own expected fleet battle.
11
Both pre-war and wartime tactical handbooks survive in the National Archives at College Park.
12
Hattendorf et al., Sailors and Scholars, 72.
13
The General Board records at NARA include a file (General Board 418) on the development of Battle
Plans Nos. 1 and 2. It begins with a 1903 file produced by the Naval War College, giving extracts from the
report of a special board on tactics. The paper is marked “Archives of U.S. Naval War College,” but it is not
clear that a copy survives in Newport. On 4 June 1904, the General Board formally requested a study of bat-
tle plans. The existing signal book allowed a senior officer to maneuver the growing U.S. fleet, but he needed
guidance as to what he should do in combat. Secretary of the Navy Morton ordered the War College to form
the summer course officers (rear admirals and captains) into a board to answer the three tactical problems
the General Board posed. The General Board includes the result: a 1 July 1906 tentative battle plan (Battle
Plan No. 1) prepared for testing by the fleet. The file also includes a 3 August 1906 letter to Commander
Niblack (who had written about the War College and officer education) from British Admiral Charles Beres-
ford. An attached memo recounted that when Beresford met Admiral Dewey on 4 May 1905, he “stated
that Admirals in the British Navy get no training in handling fleets before they attain flag rank, and, after
attaining that rank, some have little time and inclination to handle them. He found no fault with the method
of dealing with strategical questions, but believes that this lack of experience in handling tactical questions
practically is a source of great weakness to the Navy.” Beresford was proud of forcing his subordinates to
practice tactics, but pointed out that the Royal Navy had no regulation requiring such exercises: “[S]ome
Admirals do not maneuver their squadrons beyond the mere passage from port to port.” He wanted Dewey
to support his drive for such a regulation. For his part, the Secretary of the Navy ordered all commanders in
chief afloat to test and report on the new battle tactics which had just been prepared by the General Board
in conjunction with the War College. The General Board file includes these reports; Battle Plan No. 1 was
accepted. It and Battle Plan No. 2 were embodied in the Atlantic Fleet war orders when the United States
went to war in 1917.
14
The texts of the lectures have survived, and many of them offer insight into the typical views of the in-
ter-war Navy. Because the game material is massive, and hence relatively inaccessible, many writers have
taken the lectures as the core of the War College course. For example, the lectures present what seems to us a
very conservative view of the coming technology of aircraft. If, however, the view is Sims’s or Harris’s—that
the students would learn by doing—the details of the lectures are almost irrelevant.
15
Post-game discussions by War College President Sims were concerned entirely with the art of tactical
decision-making. As head of the Tactics Department, Harris Laning introduced comments on the tactical
lessons of the games.
16
Gaming in tandem with the 1927 Fleet Problem is described by Lillard, Playing War, 83–84. Presumably
to avoid political complication, the enemy was called Black, although the composition of the enemy fleet
indicated that it was British. Black was usually reserved for Germany, the most likely pre-1914 enemy.
17
The simpler pre-1917 games were often played in the fleet; in fact, one of the surviving sets of rules was pro-
mulgated not by the War College, but by the Atlantic Fleet. They were seen as means for the fleet to educate
itself and also to undertake tactical development.

Notes 215
18
The War College Archives contain several early sets of game rules. The 1901 rules contain both tactical
(including one-on-one) and strategic rules, a total of 21 pages. The 1905 rules (coded XMAT, in NWC RG
35) are more elaborate, 25 pages long (after an introduction), with special curves to make it possible to take
maneuvers into account. An attempt is made to indicate fire effect at increasing range. NWC RG 4, Box 90,
Folder 7 (XMAC) is a more elaborate set of tactical war game rules, with appended tables of ships of the U.S.
and German navies and merchant fleets, issued in 1912. It amounts to about 61 pages including the extensive
tables. The inter-war rules included no such tables; all ship data and fire effect data were separate. NWC
RG 8, Box 90 contains Lieutenant Commander H. E. Yarnell’s account of “Naval Chart Maneuvers 1916,”
meaning a dissertation on war gaming as it was then practiced (it is coded XMAC Accession 1916/174 in
the War College archive). Yarnell discusses the mechanics of gaming as it was then practiced, with rules for
items such as radio messages. No copies of the much-improved rules issued in 1919–22 have surfaced at the
War College, but some of the record groups involved are disorganized and may contain them. The record for
1923–41 in RG 8 is fairly complete.
19
War College files (RG 4, publications) include a massive treatise issued by the Intelligence Department in
June 1931, Construction of Fire Effect Tables 1922; Notes on Revision of Fire Effect Tables Up to March 1930;
The 1931 General Revision of Fire Effect Tables. Formulas and figures were adjusted to match new data. Mate-
rial on the construction of the 1922 rules is coded XOGF (RG 8, Box 96, Folder 7).
20
Rear Admiral Bruce McCandless, USN (Ret.), “The San Francisco Story,” in Proceedings of U.S. Naval
Institute, November 1958.
21
For example, in a 1939 game (Operations VI, Class of 1939) Hood (CC-4 in the Royal Navy list) was
credited with a lifetime of 17.7, compared to 16.1 for a Royal Sovereign, 16.6 for a Queen Elizabeth, and 19
for a Nelson. U.S. figures were 18.5 for a Maryland, 17.7 for a California (which had the same armor), 17.3
for the modernized New Mexicos, and 17.2 for Pennsylvania. Only the Nevadas (16.2) and older ships were
as weak, in these figures, as older the British battleships. All figures were in the common denomination of
14-inch hits.
22
Laning, An Admiral’s Yarn, 327.
23
“Superiority of the British Over the United States in the Fighting Strengths of the Capital Ships (Exclud-
ing Aircraft Carriers) Resulting From the Treaty Limiting Naval Armaments and Study and Computation
of the Relative Fighting Strengths of the British and United States Navies,” a letter from the President of
the War College. NWC RG 8 Series 2, Box 97, folder 5, coded XOGF/Safe. The letter of transmission to
the Secretary of the Navy is dated 13 October 1922; it refers to an earlier submission dated 12 October,
which is supplemented by the current one. A much more elaborate comparison (with the same conclusion)
was produced in 1924. The War College Class of 1923 gamed an engagement between the U.S. and British
fleets as its Tactical Problem II. During the Washington Conference, the General Board evaluated national
naval strengths on the basis of age and displacement, concluding that the ratio of British superiority was only
1.107. Sims’s analysis showed a crushing British superiority: an effective ratio of 1.44:1 at 15,000 yards, but
2.56:1 at 24,000 yards.
24
Lieutenant Commander H. R. Thurber, Suitability of Naval War College Instruction in Gunnery With Ref-
erence to Tactics, in folder NC (for War College), but in a War College jacket coded XOGF 1937-104, in the
CNO/SecNav classified files (NARA RG 80). Thurber was particularly worried by what he called the range-
band theory. Certainly it could be argued that armor and shell performance would define ranges at which
some ships could fight without being penetrated, but Thurber showed that the data from which the tables
were built was much less than credible. Data on foreign deck armor was nearly nonexistent. What little there
was, was grossly inconsistent. The width of the range bands implicit in War College tables was 2,000 to
6,000 yards, but the endless uncertainties Thurber cited would swamp that. For example, Thurber estimated
a difference of 9,000 yards in penetration between old and new armors. Grossly uncertain information on
Orange ships such as heavy cruisers did not help. Were the belts on those cruisers three or four inches thick?
The difference was equivalent to 6,000 yards for a U.S. 8-inch gun. Was it two inches thick? The actual figure
was about four inches, but no one knew that before 1945. Uncertainty as to deck thickness in the rebuilt
Japanese Kongo class battle cruisers equated to 10,000 yards (four versus seven inches over magazines, the

216 Winning a Future War

true figure being 4.7 inches). It is not clear how seriously Thurber was taken, but the survival of his report in
the CNO/SecNav classified file suggests that he was. Thurber’s analysis seems to have been the result of his
experience as a member of the War College class of 1936. He retired as a vice admiral.
25
War College archives include several studies taking supposed British means of increasing gun elevation
without modifying turrets into account, and the issue is also prominent in General Board records (mainly
420-6). Bulges were anti-torpedo structures built outside a ship’s hull, typically including air spaces. The
Washington Treaty certainly allowed their installation. Looking back, it is difficult to see why a fleet would
deliberately flood bulges, thus reducing its immunity to torpedo fire. The British did not change their own
interpretation of the treaty and therefore did not increase the elevation of guns onboard ships they modern-
ized in the late 1920s and early 1930s, but it is entirely possible that the chronically underfunded Admiralty
was using the treaty as an excuse to avoid expensive gun-mounting modification. For the U.S. side of the
controversy, and the U.S. battleship modernization program, see the author’s U.S. Battleships: An Illustrated
Design History. For the British side, and for arms control negotiations, see his British Battleships 1906–1945.
26
In 1936, the U.S. communications intelligence organization broke a message reporting the trials of the
battleship Mutsu, which was now credited with a speed of over 26 knots (the previous accepted figure had
been 23.5 knots). We now know that this was also the original design speed of the ship, unreported for some
reason through the 1920s. The real surprise, not known until late in World War II, was that reconstruction
of the earlier Japanese battleships had given them much the same speed as the Mutsu, which meant that the
Japanese fleet could have crossed the “T” of the U.S. fleet unless it was stopped in some other way. Through
the war, U.S. official handbooks credited the Japanese battleships with much lower speeds (so did the Brit-
ish). Presumably the high speed of the Mutsu did not figure in these handbooks because of the way it was
discovered. None of this turned out to matter during World War II, because (as the War College forecast)
the Japanese withheld their battle fleet pending attrition of the U.S. fleet (which did not, in the event, suc-
ceed). The most visible effect of the discovery of the speed of the Mutsu was redesign of the South Dakota
class. There is no evidence that the British did even this well.
27
Laning, Admiral’s Yarn, 263–64. Sims told Laning to incorporate what was being learned in his thesis on
battle tactics, which became a student handbook.
28
Van Auken had been a gunnery officer during World War I; he had run afoul of his fleet commander by
writing that failure to install a new fire-control system would doom the fleet in battle. He spent the post-war
decade in the naval education system, attending the Naval War College as a member of the Class of 1927. His
outspokenness shows in some of his written comments explaining war game lessons. Van Auken’s World War
I story is told by C. C. Wright, “Questions on the Effectiveness of U.S. Navy Battleship Gunnery: Notes on
the Origins of U.S. Navy Gun Fire Control System Range Keepers,” Warship International 41/3 (September
2004). After leaving the War College, van Auken commanded the battleship Oklahoma for a year.
29
U.S. war planners were certainly aware that they needed large numbers of auxiliaries to transport an expe-
ditionary force to the western Pacific, and, by 1929, printed War Plans documents included a list of required
conversions (this document is available as a microfilm at NHHC). They included aircraft carriers converted
from liners (XCV) and seaplane tenders converted from merchant ships (XAV) as the only ways of increas-
ing sea-based air power in a treaty-limited era. It also included armed merchant cruisers (XCL) to harass
enemy trade and to protect U.S. sea communications from enemy attack. The XAVs were dropped in 1935
both because there were too few fast U.S. liners suitable for conversion (particular ships were earmarked and
preliminary plans drawn) and because conversion would take too long—at least 180 days. The conversions
envisaged at the time were quite elaborate, resulting in ships resembling existing fleet carriers, but slower.
The need for fast naval auxiliaries was explicit in public explanations of the need for the Maritime Commis-
sion, created by the Merchant Marine Act of 1936. It was headed by former U.S. Navy Chief Constructor
Rear Admiral Emory S. Land.
30
War College records (RG 8) include a brief description of the 1933 design of a battle cruiser as a candidate
for construction once the battleship-building “holiday” initiated by the Washington Treaty (and extended by
the 1930 London Naval Treaty) expired at the end of 1936. Up to that time, proposals for such battleships
had envisaged conventional low-speed designs (the same was true of the British). War games highlighted

Notes 217
the operation of the Japanese battle cruisers, not only in fleet action, but also in cooperation with Japanese
carriers and in raiding operations against vital U.S. sea communications to the western Pacific. The General
Board absorbed the idea. Its proposed battleship sacrificed firepower for speed. CNO preferred a different
balance, a somewhat slower ship with more firepower. He won the fight because at the crucial time he was
also acting Secretary of the Navy. Even so, the resulting ships, the North Carolinas, had a rated speed six
knots faster than their predecessors. They seem to have been conceived as a fast wing for the future U.S.
battle line, because the next class of battleships, the South Dakotas, were to have traded speed (reduced to 24
knots) for better protection. That is not evident from what happened; the South Dakotas’ speed was increased
(via a radical redesign) because U.S. Navy signals intelligence picked up evidence that the rebuilt Japanese
battleship Mutsu was a good deal faster (over 26 knots). During World War II, a U.S. Navy summary of
the achievements of signals intelligence singled out this information as its most important pre-war coup,
although that is open to question. Unfortunately, correspondence files that might have demonstrated a link
between the 1933 sketch design and General Board policy do not exist; the General Board files (in NARA
RG 80) show the results of its thinking, but generally not its sources.
31
This change is evident in War College advice proffered in 1930. Unfortunately, the War College archive ap-
parently does not contain communications from the college president to OPNAV and other naval offices (or,
for that matter, correspondence relating to the scenarios chosen for war games) because it was reconstructed
in the 1960s from records that had gone to various Federal records centers. In this book, War College influ-
ence has been traced using mainly the records of OPNAV and of the General Board.
32
Series II of the War Plans Division (OPNAV Strategic Plans Division) files in NARA II is “Naval War
College Instructional Materials 1914–1941,” beginning with Joint Army-Navy Problem (Overseas Expedi-
tion) of 1923 in Box 14; the next in the series is Strategical Problem I (Joint Problem) of 1924. Boxes 28–34
are studies conducted at the War College. Another War Plans Division file is the receipts for war gaming ma-
terial received from 1937 onward. These studies range from fleet tactics to the strategy of economic warfare.
33
The U.S. Navy’s views as to which games mattered are reflected in the requests for gaming material from
the War Plans Division. Its surviving archive for 1936–41 includes numerous Orange-Blue games, but only
one Red-Blue game, from 1938. That game was presumably intended to support a study conducted that year
of which British bases the United States should seek to obtain for hemispheric defense. The only other game
in these files involving Red was a 1941 Atlantic exercise in which the United States and Britain were allied
against Germany.
34
Van Auken went further in his formal commentary on the big 1931 Blue-Red (i.e., against the British)
game (Operations III for the Class of 1932). He considered it essential to game and plan a cooperative war
by both countries against Japan and its allies. He repeated his comment the following year in commentary
on a Blue-Orange game (Operations II for the Class of 1933).
35
It also helped that the United States began naval rearmament in 1937 with the expiration of total tonnage
limits imposed by the Washington Treaty and the London Naval Treaty of 1930. However, the U.S. rear-
mament program was dwarfed by that of the British.
Operations Problem II-1927, with introductory remarks dated June 1927 by Captain H. E. Yarnell, who
36

wrote that “it is a situation which represents very closely what would happen under actual conditions.”
NWC RG 4, Box 34, Item 1285-F.
37
By 1929, War Plans documents included the proviso that if the Japanese did not see sense once the block-
ade began, it might be necessary to bomb Japanese cities. It was widely believed that the flimsy Japanese
cities were unusually vulnerable to bombing, particularly using incendiaries. This possibility did not affect
planning or gaming because the same islands that had to be seized to support an effective blockade could
also be air bases. The main effect of the possible change was to add another type of required auxiliary, an
aircraft transport for army bombers. No invasion of Japan was contemplated because it was considered most
unlikely that a sufficient army could be raised and transported across the Pacific. See Miller, War Plan Or-
ange, 150–52 and 162–64 for the use of air power; elsewhere in the same chapter, he estimates the numbers
of troops the Army thought would be needed.

218 Winning a Future War

38
Many accounts of the disappearance of Amelia Earhart, the famous aviator, connect her flight with a Na-
vy-sponsored reconnaissance project. Whether or not that is true, it reflects the well-known and badly frus-
trated Navy interest in the islands. Post-war investigation suggests that the islands were not fortified, but
that facilities such as seaplane ramps were built.
39
The U.S. demand for high performance is evident in BuAer’s draft fighter history (through 1945) in the
files of the Ships Histories Branch at NHHC (formerly the Aviation History Branch). The demand for
performance is discussed as the key factor in U.S. naval fighter development at least from the early 1930s,
resulting in pre-war designs such as the Buffalo, Wildcat, and Corsair. It was in striking contrast to the
consistent British view that low performance was acceptable because British carrier aircraft would face only
other carrier aircraft. The Admiralty position is described in, for example, a proposed Fleet Air Arm hand-
book produced in 1936 (ADM 116/4030). When the handbook was circulated to senior officers for com-
ment, Admiral Sir Dudley Pound, then Home Fleet commander, asked pointedly how a concept to defeat
Japanese carriers applied to the more likely war in European waters against land-based aircraft. The answer
from the director of Naval Air Department was simply that this possibility had not been considered. The
main British inter-war scenario involved a decisive battle between a British fleet sent east to Singapore and
a Japanese fleet coming from the north. The British would not meet Japanese aircraft en route to Singapore,
and they assumed that the only aircraft they would have to fight before and during the decisive battle would
be Japanese carrier aircraft. It is not clear what they planned to do after the decisive battle, when they would
have to operate near Japan in order to impose an effective blockade.
Miller, War Plan Orange, generally distinguishes between “thrusters” and “cautionaries” who moved the
40

war plan back and forth.

41
This shift is described by Miller, War Plan Orange, 139–42.
42
In Box 19 of the War Plans Division series, NARA II. This game does not figure, at least explicitly, in
Miller, War Plan Orange. He sees the shift based on a series of developments, including a new assumption
that the Japanese would fortify the Mandates and improvements that made aircraft much more effective.
However, the War College had been ascribing improved performance to aircraft for some time. Miller also
attributes the shift to a change in personnel from “thrusters” to “cautionaries.” It is impossible to say just
how the 1933 Operations IV game influenced thinking in the War Plan Division, but it seems that, at the
least, anyone rejecting “thrusting” could use it as powerful ammunition. Note that this is the only war game
described in detail by either Vlahos, The Blue Sword, or Hattendorf et al, Sailors and Scholars.
43
MacArthur was involved because U.S. war plans were joint documents approved by a Joint Board estab-
lished in 1903. The Army view, reflected in the first volume of its official history of the Pacific War, was that
U.S. Pacific strategy was about defending or else regaining the Philippines. That was very different from
the consistent Navy view that Japan had to be defeated, disgorgement of a conquered Philippines being
one of several consequences. See Louis Morton, Strategy and Command: The First Two Years (Washing-
ton, DC: Government Printing Office, 1961); the first chapter describes pre-war U.S. strategy. It is possible
that Morton’s concentration on the Philippines is connected with General MacArthur’s insistence that the
Philippines were the appropriate objective in October 1944 (the Navy would have preferred to seize a base
in Taiwan or on the Chinese coast). Morton describes the Army position: Although the Japanese could oc-
cupy most of the Philippines, it would be worthwhile to hold Manila Bay as a fleet base, or even to hold just
Corregidor to deny that base to the Japanese. This view does not correspond to the Navy view well before
1934 that Manila would surely fall and that the fleet would have to go to an alternatate port in the southern
Philippines. The situation was further complicated by congressional action in March 1934 promising the
Philippines independence in 1946. Morton does not mention MacArthur’s acceptance of the change in the
accepted Orange war plan, which dropped any attempt to reinforce the Philippines in wartime (it is dated
to 1936). After the war plan had changed, Army students at Newport commented that the change had been
inevitable, as they had doubted that the islands could be held. They seem not to have realized that the key
point (at least to the Navy) was that even if part of the Philippines could be held, the fleet almost certainly
would not have made it there intact enough to defeat the Imperial Japanese Navy—which was necessary in
order to go to the next step, the blockade. In effect, the students’ comment demonstrated how wide a gulf
existed between the services’concepts of a future Pacific war.

Notes 219
44
It might be argued that the war planners could simply add up opposing forces to compare them, or that
they could follow the earlier path of the “applicatory method” and lay out likely enemy courses of action for
comparison with U.S. courses of action. Simply adding up forces would not have sufficed to predict success
or failure; a campaign was just too vast for that. A force ratio might be used to predict the outcome of a single
battle, but not in detail. It could not predict how many ships would emerge how badly damaged, for example.
All of the war planners had been indoctrinated at the War College to see a two-sided game as far better than
the predictions a single planner or committee might make. A single planner or even a committee was just too
vulnerable to wishful thinking. Even after the War College had lost much of its influence, the war planners
continued to use its gaming data; they still wanted to avoid pitfalls. Moreover, gaming made it possible to
tease out unexpected possibilities. For example, during a 1932 game (Tactical VI, Class of 1932) a student
suggested a carrier air raid on Tokyo as a way of concentrating Japanese attention. The idea was incorporated
into a report on interesting developments in gaming, but it was not played out at the time—only because the
game had no mechanism for testing it against the team playing the Japanese.
45
The key advantage of seaplanes was that their tenders were not limited by treaty. It probably helped that
Congress did not consider them combatants. The seaplane carriers were modified tenders (AV). Their de-
signs incorporated carrier-type catapults. The aircraft they would launch were intended to provide Marines
with bomber support from lagoons near a contested island (they retained the AV designation). This concept
explains experiments with floatplane versions of dive bombers and probably also fighters. It was dropped
soon after war broke out, and the Marines could rely on numerous escort carriers. The new idea of a fully
mobile seaplane force was symbolized by the transfer of seaplanes from the Base Force at Pearl Harbor to a
new Patrol Wings Scouting Force. In 1939, there were only two modern tenders, the Wright and the ex-car-
rier Langley, but new ships were being built. Fourteen old destroyers had just been converted as prototype
AVPs, and nine minesweepers were also reclassified as AVPs. In 1940, the first purpose-built seaplane ten-
der, Curtiss (AV-4), was completed.

4. War Gaming and Carrier Aviation

1
The closest any World War I naval airplane came to participation was scouting during the Battle of Jutland.
It had no impact because the radioed report from the airplane was not transmitted onward by its carrier.
At the end of the war, the British had assembled a torpedo-bomber strike force, but the war ended before
it could see combat. Moreover, the strike force was intended to attack the German High Seas Fleet in its
harbor, not at sea.
2
During World War I, U.S. naval aviators had considerable access to British thinking, and they wrote exten-
sive reports, which among other things helped convince the U.S. Navy that it wanted carriers. Direct contact
with the British continued for a time after the war, as the British hoped that the United States would contin-
ue to help maintain the peace. Thus, U.S. files include contemporary plans of British carriers and, in 1922,
the British provided the U.S. Navy with details of its arresting gear. British files show that most exchanges
stopped in 1920, when it was clear that the United States would not be joining the League of Nations. One
great irony of the inter-war period is that the relatively large tonnage allowed for carriers under the Washing-
ton Treaty—135,000 tons for the United States and the United Kingdom—was the result of a British initia-
tive based on perceived fleet requirements, taking into account the very limited number of aircraft the British
believed that any one carrier could operate. In 1921, no other navy saw the need for a substantial carrier
force, and the U.S. Navy would probably have been satisfied with the conversion of the two battle cruisers.
3
The first approach may have been a set of data provided by Director of Naval Aviation Captain T. T. Cra-
ven dated 17 June 1919, replying to a 4 June letter from Lieutenant Commander H. M. Lammers. Land
(wheeled) aircraft would fly only from flying-off platforms on decks and turrets; they could not land back
onboard. Craven did not allow for fighters, but only for large (patrol) flying boats, spotters, torpedo bombers,
and reconnaissance aircraft. Of these, the spotters were the fastest, at 90 miles per hour; the slowest were the
torpedo bombers, at 70 miles per hour. Servicing time between flights was two hours for large aircraft or one
hour for smaller ones. Endurance was six hours for a reconnaissance plane, four for a spotter, and three for
a torpedo bomber. Aircraft could spot single ships 30 miles away and close formations 40 miles away. Max-

220 Winning a Future War

imum radio range for a land plane was 75 miles (they could receive messages at 250). There was no estimate
of hitting probability under various conditions. Craven suggested that the War College should pay attention
to lighter-than-air craft (blimps and rigids) that it had neglected in the past. NWC RG 8, Box 77, folder 6,
Accession 1919/561, coded XAEG.
4
The June 1923 maneuver rules are in NWC RG 4, Box 18, folder 20; they are marked 803-B.
5
NWC RG 8, file XMAR (1935-12) dated 26 February 1935, “Manuever Rules Pertaining to Aircraft and
Airplane Characteristics—Revision of.” Games often involved smoke screens laid by aircraft; King drily
remarked that students at the War College were much better equipped than pilots to lay them correctly.
6
The most striking exception was the British torpedo bomber–spotter–reconnaissance airplane (TSR), the
Swordfish. In games, it was accorded rather better performance than it had.
7
These were not new categories, although the designations were new. Converted carriers figured in 1920 war
games under the designation JC (J for carrier, C for conversion). This was a natural category; nearly all the
existing British carriers were conversions of one kind or another. The first British carrier with a full-length
flight deck, HMS Argus, had been converted from an incomplete liner laid down for an Italian company. It
seems unlikely that the XOCV designation implied the later idea of converting liners upon mobilization to
expand the U.S. carrier force.
8
Initial estimates of bomb accuracy were optimistic. For airplanes at up to 1,500 feet, it should take two
bombs to make one hit on a large or intermediate ship, such as a battleship or cruiser; up to 4,000 feet, four;
up to 8,000 feet, six; and up to 12,000 feet, ten (12 for the intermediate ship). Airplanes dropping bombs
from below 500 feet would be destroyed by their blast. The 1924 rules recognized that these figures, taken
from target practice, would likely be cut sharply in reality. For example, when bombers were opposed by
considerably superior forces, hitting would decrease by 40 percent. The umpire would assign further de-
creases due to target maneuvers or speed or anti-aircraft fire. Figures became more realistic as experience
accumulated; they were based on results in the annual bombing exercises. The 1925 rules assigned a 40
percent chance of hitting (i.e., less than before) to a bomb dropped from 1,500 feet; 15 percent from 8,000
feet, and 10 percent from 10,000 feet, in each case against a large ship. Anti-aircraft fire was expected to force
aircraft to attack from higher altitude and thus to lose accuracy (in 1931, the minimum bombing altitude
was 2,000 feet, and at that altitude a bomber would achieve only 26 percent hits against a non-maneuvering
ship). The penalty exacted by opposing fighters increased (a superior number of fighters could reduce bomber
effectiveness by 60 percent). In 1936, bombing figures were reduced by a quarter to take into account the “fog
of war.” Horizontal bombing was accorded greater accuracy in 1937, probably because the U.S. Navy now
had a precision bomb sight (the famous Norden bombsight). Now, a bomber was accorded 8 percent hits
from 9,000 feet (but much less against a fast ship). There was also an uncertainty in the effect of hits in the
14-inch shell currency used by the War College. In the 1933 rules, a 1,000-pound bomb hitting a battleship
was equivalent to 0.58 shell hits (0.75 against a carrier or cruiser), but in the 1937 rules a 500-pound bomb
hitting above water was considered equivalent to a 14-inch shell, and a 1,000-pound bomb was considered
equivalent to two 14-inch hits.
9
Thus, in the 1925 rules the chance of being wrecked while landing in daylight was given as 1/30 rather
than 1/20 (1/20 rather than 1/10 in darkness). In 1927, rules were made more elaborate, the chance of being
wrecked landing on a carrier depending on the type of airplane and the weather. In smooth weather, a fighter
had a 1/25 (0.04) chance of being wrecked on landing, but in rough weather this rose to 1/10. Figures were
again reduced in the 1929 rules, to 0.03 onboard a carrier in day under all sea conditions (0.06 at night).
Comparable figures for a landing field ashore were 0.001 and 0.03. In 1933, the chance of being wrecked by
landing on a carrier was reduced to 0.01.
10
NWC RG 4, Box 29, Folder 9.
11
The game did not include individual decisions by fighter pilots; it is not clear who realized that the obser-
vation planes were only a feint, and that not all of Blue’s fighter pilots should (or would) have been drawn in.

Notes 221
12
The post-game analysis pointed out that the barrier was effective only against aircraft coming from an
expected direction. Gaming was the way to guess unexpected tactics.
13
This presumably meant that more aircraft were being brought up from hangar decks and prepared on deck
(e.g., by having their wings unfolded). In 1924, the U.S. Navy had no idea of how that would work.
The game lasted this long because the Blue commander failed to take advantage of the situation and turned
14

away from Surigao.

15
Laning noted two important tactical innovations that year. One was the sustained Blue air operation; he
described the careful refueling arrangements as “building for the future.” The other was the circular forma-
tion, which the inter-war Navy later adopted, invented by Roscoe MacFall. Reeves was chosen as Laning’s
successor as head of the Tactics Department. Given Laning’s generally disparaging remarks about the tac-
tical choices made by other players, it seems very unlikely that he considered anyone except Reeves worth
choosing.
16
NWC RG 4, Box 26, Folder 2. In this game, Red’s fleet headed for the Caribbean to attack Blue trade
routes. Blue’s objective was to prevent Red from establishing itself in strength there. The objects of the Tac-
tics Department were to explore (1) how a fleet (in this case Red) could best conduct its train; (2) how the
Blue fleet could deal with Red’s superior speed and firepower; (3) how Blue could overcome Red’s superiority
in cruisers (in this case, 24 Red versus 10 Blue); (4) whether, given the same total number of aircraft, the
two big U.S. carriers were superior to six smaller Red carriers; and (5) whether cutting the Red fleet by four
capital ships would overcome its superior fighting strength. The game did not answer (1), because Blue never
tried to destroy Red’s train, despite its advantageous position. It did show that neither side could easily deny
the other air spotting. Blue was able to keep spotting even after Red destroyed the flight decks of its two
carriers and thus (in theory) gained “control of the air.” The only consequence of losing the flight decks was
that Blue spotters could not land back on the carriers to refuel, but by then the gun battle was over. Blue had
many more destroyers than Red. An important cruiser role was to break up destroyer torpedo attacks on the
battleships. In this case, Blue had so many more destroyers that some of them could be devoted to sinking
Red cruisers (at the end of the game Blue had more cruisers than Red). The post-game critique pointed out
that the Red commander evidently did not give sufficient consideration to the question of which side had air
control and thus could use airplanes for spotting. It made no provision for the possibility that Blue—and
not Red—would be able to spot from the air, which in turn would determine which side could fire at greater
range. “As this was a question yet to be decided by the outcome of the air engagement, it should have received
a most prominent position in the Red commander’s battle plan.” Red certainly intended to seize air control
at the outset and directed that every effort be made in that direction, but he should have reckoned with the
possibility that this plan would fail. Blue made no provision for using his air forces to observe enemy move-
ments, yet this was vital from the moment of contact. “All movements of enemy forces, their disposition, the
launching of air attacks, should be promptly known.” The critique, dated 14 January 1925, was unsigned, but
was probably written by Reeves as head of the Tactics Department.
17
Treaty rules included a proviso that new ships in categories whose total tonnage was limited could be re-
placed only to replace over-age ships in those categories (the treaty included rules as to lifetime). Langley was
not counted in future projections because, as an experimental ship, she could be discarded at any time to be
replaced by a new ship (or a fraction of one).
18
“Use of 10,000 ton Cruisers for Carrying Aircraft,” in General Board 420-7 file 1925–1931 (NARA RG
80), from War College President Rear Admiral C. S. Williams (writing as a member of the General Board)
to Senior Member, General Board, 23 April 1925. This letter rejects the idea of a combination cruiser-carri-
er, a concept that nonetheless survived into the 1930s.
19
General Board file 420-2 Serial 1313 dated 30 March 1926 in NARA RG 80 file 420-2 1926–1927).
NWC RG 8 Series 1, Box 37, folder 6: UNC 1925–1927. The cover letter from the President of the War
20

College is dated 11 April 1927, coded UNC (1927-83).

222 Winning a Future War

21
Summary of officers’ comments by Lieutenant Forrest Sherman attached to the letter from Admiral Pratt,
dated 1 December 1926. The officers were Captain J. W. Greenslade, Captain S. R. Moses, Captain B. B.
Wygant, Commander R. R. Stewart, and Lieutenant Forrest Sherman. Moses pointed to the advantages
of converting a suitable merchant ship into a medium carrier, noting that authority had already been given
to convert a second experimental (training) carrier comparable to Langley. Inputs were letters from the bu-
reaus of Aeronautics and Construction and Repair. Aeronautics strongly favored smaller carriers. Sherman
pointed to experiments he had conducted on scale models of the flight and hangar decks of the Lexington
and Saratoga using paper airplanes to show problems that could arise if aircraft spotted on deck did not
correspond to changing requirements (one conclusion was that the elevator arrangement of these ships was
defective). He pointed out that fighters spotted aft could take off over heavier bombers spotted forward. His
main conclusion was that the ships could not operate as many aircraft as were planned.
22
Light bombing meant the new tactic of dive bombing, which in 1926 involved fighters carrying 25-pound
bombs. These weapons were sufficient to wreck a flight deck and send a carrier back to a shipyard. In his sum-
mary comments, Sherman proposed building a non-watertight false deck 2.5 feet above the real deck. This
idea does not appear in the approved characteristics for Ranger, but she had exactly the sort of easily repaired
flight deck Sherman envisioned, the only difference being that there was no heavy structure immediately
below it. Subsequent U.S. carriers through the Essex class had this type of flight deck. Sherman also called
for the largest possible hangar. Having visited the Lexington, he was impressed by the potential fire hazard in
a hangar, and proposed that the hangar of any new ship be open and well ventilated. He mistakenly believed
that the British had drawn the same conclusion from their earlier carriers, including HMS Furious.
It is not clear why Sherman did not realize that the British had no choice; these were the only two existing
23

hulls they could convert under the limits in the Washington Treaty.
As areas, they were proportional to the square rather than the cube of dimensions. Displacement (volume)
24

was proportional to the cube.

General Board 420-7 (Serial No. 1362) file 1926-31, Characteristics dated 1 November 1927 (NARA RG
25

80). These characteristics were formally approved by the Secretary of the Navy on 7 November.
26
Memorandum dated 17 October 1927 in General Board 420-7 file 1926–1931 (NARA RG 38). Horne
called for a program of one 10,000-ton carrier specifically to support the battleships’ spotting planes (36
spotters and 36 fighters onboard); one 13,000-tonner (32 knots) for light cruiser planes (54 observation and
54 fighters); and four 14,000-tonners (29 knots) specifically to support bombers (32 bombers, 48 fighters).
At this time, the only design on offer was Ranger, which was sometimes credited with capacity for 104 air-
craft.
27
Ranger was authorized under the FY29 program and ordered on 1 November 1930. Her design was re-
vised after the two big carriers entered service. She had been conceived as a flush-decker, but their operation
demonstrated the importance of an island, and one was belatedly added. How late in the process that came is
evident from the fact that she retained the tilting funnels required by her original flush-deck design; trunk-
ing them into the island would have been too expensive.
28
Until April 1939, the Fleet Air Arm was part of the RAF, and the RAF leadership did not want to divert
scarce resources to it. However, the Royal Navy paid for the airplanes and the pilots, and it does not seem to
have understood that it needed so many more of both. This was partly, though not completely, reflected in a
failure to equip carriers with aircraft on anything like a U.S. scale.
29
The London Naval Treaty of 1930 tightened the definition of a carrier to include any ship intended to
launch aircraft. In addition, it included limits on the total tonnages of cruisers, a category not limited by the
Washington Treaty (except that maximum tonnage was 10,000 tons). As a consequence, installing a flight
deck on an existing battleship or cruiser would move it into the limited carrier category. Pratt extracted a
limited exemption for air-capable cruisers in connection with the primary British objective at the conference,
which was to end construction of heavy (8-inch gun) cruisers. The British thought that any navy building a
light (6-inch gun) cruiser would necessarily limit its size and that therefore the cost of maintaining their own
cruiser fleet would be reduced. This did not turn out to be the case.

Notes 223
30
“A Tactical Study based on The Fundamental Principles of War of the Employment of the Present Blue
Fleet in Battle showing Vital Modifications Demanded by Tactics,” February 1925, in NWC RG 8, Box 92,
folder 2.
31
Reeves never wrote down these conclusions; they are deduced from the conclusions included in the relevant
game and from what Reeves then did as Commander, Air Squadrons, Battle Fleet. Had the results of the
1925 game not been available, it might have seemed that Reeves would have been motivated by the need to
concentrate large numbers of torpedo bombers against initial targets in the enemy’s battle line. In game after
game from 1923 on, the question was whether the enemy’s flight decks should be the initial target.
32
Thomas C. Hone, Norman Friedman, and Mark D. Mandeles, American and British Aircraft Carrier De-
velopment 1919–1941 (Annapolis, MD: U.S. Naval Institute Press, 1999), 40.
33
Reeves’s new technique is described in the annual report of Aircraft Squadrons, Battle Fleet for 4 Septem-
ber 1926 through 30 June 1927. The Naval War College archives contain a two-page excerpt from the Weekly
News Letter of the U.S. Battle Fleet for 29 August–4 September 1926 reporting that nine aircraft landed on
in an average of one minute and 16 seconds each; “it is now possible with a squadron trained and flying fre-
quently to fly a squadron on and off at intervals not exceeding one minute thirty seconds.” This item is coded
XMAR and UANOP/1926-136 in NWC RG 8.
34
These figures were attained when Langley was fitted with both transverse and longitudinal arresting wires.
The latter were intended to keep an airplane from slipping to one side (the first British airplane to land on
a carrier was lost on a second attempt when it was blown over the ship’s side). The longitudinal wires were
eliminated in 1929, considerably reducing the landing interval. When he discussed aviation at Newport on
2 August 1929, Captain John Halligan Jr., commanding officer of the carrier Saratoga, said that once these
wires were gone he expected the landing interval to be reduced to 40 seconds. He had recently carried out
a dramatic air attack on the Panama Canal as part of Fleet Problem IX. NWC RG 8 Guest Lectures file,
Halligan folder, Lecture File (53). For this operation, Saratoga had 110 aircraft onboard, more than 90 of
them assembled (she was carrying Langley aircraft as well as her own). Halligan said that her flight deck
could accommodate more than a hundred aircraft ready for takeoff, including a squadron of heavy bombers.
No foreign carrier had anything like this capacity.
35
The author remembers the late senior British naval constructor D. K. Brown asking him how American
carriers with smaller hangars than their British counterparts could operate so many aircraft (he sent a letter
tabulating relative hangar areas). The British and the Japanese were both forced into using multiple han-
gars (stacked one atop the other) on carriers, with all the attendant operational problems of complicated
elevators. The British became aware of U.S. practices during the 1930s, but by that time the Royal Navy
found attempts to increase the rated number of aircraft per carrier blocked by the Royal Air Force. When
he designed the two-hangar Ark Royal in 1934, the director of naval construction pointed out that, using
U.S. practices, she could have stowed all her aircraft on deck, considerably simplifying the design. Ark Royal
was fitted with a barrier, but she did not employ U.S. deck-parking practices. They were introduced in other
British carriers during World War II. The Japanese never employed deck parks, for all that they claimed they
were following U.S. examples in carrier operation.
36
Reeves was still commander of Air Squadrons, Battle Fleet. In his August 1929 address to the Naval
War College, Captain Halligan (commanding officer of Saratoga) quoted Reeves “whose work can truly be
described as brilliant:” Saratoga had been unopposed even though she had had to approach the target much
more closely than she would have in wartime. In war, she would have launched 220 miles from her objective,
and thus probably would have been unopposed and undetected due to her high speed. In the exercise, Sarato-
ga was escorted only by the cruiser Omaha. That suggested to Reeves that attack groups should be organized
around a carrier, cruisers (armed with 8-inch guns), and destroyers. They could avoid detection thanks to
their speed, and they could punch through an enemy’s scouting forces to “arrive at a point from which [they]
could launch an air attack which could be stopped only with the greatest difficulty.” Reeves pointed out that
such employment demanded more U.S. cruisers, a category in which the U.S. Navy was deficient.
37
Flight-deck cruisers vanished from war games after Admiral Laning left the War College in 1933 (which
coincided with the departure of their great sponsors, Admiral Pratt as CNO and Admiral Moffett at BuAer

224 Winning a Future War

[due to his death]). Probably, Pratt’s departure in June 1933 was the key event. The concept survived as late
as 1940.
Memo from Admiral Mark L. Bristol, who headed the General Board, dated 20 July 1931, in General
38

Board file 420-7 on characteristics of the new carrier.

39
The War College letter dated 30 July 1932 is in RG 80 CNO/SecNav file (SC) CV; it replies to a CNO
letter dated 12 May 1932 enclosing General Board letter 420-11(SC) CV of 7 May 1932. The Naval War
College letter refers further to an August 1931 War College letter NC-3(1930-168) discussing aircraft char-
acteristics. A General Board (SC) CV 420-11 letter dated 20 July 1932, which is in the SecNav/CNO CV
file (classified correspondence), connects the War College letter with a proposed alteration which would give
Lexington and Saratoga greater bomb and torpedo stowage. The War College had failed to answer specific
questions about how much such stowage was needed, and the question was urgent because the modification
was to be included in the coming year’s budget. The questions were also circulated within the fleet. The
surviving 420-11 file includes correspondence on bomb and torpedo stowage dated July 1932, beginning
with the statement that stowage in Lexington and Saratoga was clearly inadequate—a conclusion that could
certainly be drawn from the War College letter, but which the ships’ commanding officers had reached the
previous year after the Fleet Problem.
40
The War College letter went on to describe lessons of Blue-Red games after observing that “the chances
of war appear most remote between Great Britain and the United States.” Flight-deck cruisers and cruiser
aircraft had proved very valuable in scouting along distant British trade routes. In games against a Red force
trying to establish a base near Nova Scotia or in the Caribbean, Blue carriers with both shore and carri-
er-based aircraft were valuable in reducing or destroying the Red air force, “upon which so much depends for
Red’s service of information. Just as in the Blue-Orange problems, a great need has been shown for aircraft
forces in strategical operations in addition to the demands of the major engagement.”
41
The games showed that one side or the other generally gained control of the air over the engaged battle lines
and thus denied air spotting (i.e., long-range fire control) to the other; but the games also showed that ships
had to close to less than 20,000 yards to secure decisive results before they exhausted their ammunition. At
such ranges, air spotting was not needed. In some cases battleship spotting aircraft were launched prema-
turely, either due to a misunderstanding of the situation or a deliberate enemy feint intended to weaken Blue
spotting and long-range fire. Without flight decks nearby, the spotters could not refuel and be serviced; the
battleships certainly could not stop to recover them. They had to be sacrificed. “On the other hand, it has
been shown that, with our spotting planes on carriers with flight decks damaged before an engagement, the
battle line opens fire under a severe gunnery handicap at extreme ranges.” War game rules offered consider-
able incentive to open at favorable ranges defined by the relative vulnerabilities of the two battle lines, a factor
always included in tabulations of relative capability. A 1937 analysis cited elsewhere in this book emphasized
the unreality of this “range band theory,” because it was based on very incomplete information and ignored
many significant combat factors.
42
Although in principle a technical bureau like the Bureau of Ordnance, at this time BuAer was also respon-
sible for pilot training and for air tactics (the relevant arm OPNAV arm, Op-05, was not created until 1943).
Hence its more strategic view of carrier operations, which seems to have taken into account typical war plans
affected by War College gaming.
43
These comments were made in connection with a newly proposed 25,000 ton cruiser-carrier armed with
8-inch guns. BuAer pointed out that within the available tonnage only two could be built, costing one flight
deck and aircraft capacity. Two Lexingtons and five Rangers would have supported 635 aircraft; the new
distribution offered 553. With two 25,000-ton carriers in place of the three now planned, the number would
fall again, to 429; and instead of seven carriers there would be only five. The fleet air force would be reduced
by a third compared to the original plan. The current London Naval Treaty of 1930 allowed for construction
of as many as eight flight-deck cruisers.
Wasp had an armor deck but no torpedo protection; she was designed for later addition of side armor, if and
44

when the treaty limit was abandoned. That was not done. She was slower than the Yorktowns, but faster than
Ranger. Given the fixed total carrier tonnage available, weight growth in Ranger while she was being built

Notes 225
squeezed down the tonnage available for Wasp. Wasp was not included in the FY34 program, but rather in
FY35 (she was launched only in 1939). Another year’s delay would have carried her clear of the treaty limit
on total U.S. carrier tonnage.
27 December 1927 memo by Admiral Reeves in the CNO/SecNav classified files (NARA RG 80), Folder
45

CV 1926–31.
46
Essex-class machinery was designed to sustain 20 knots astern to support landings over the bow, into the
forward arresting gear. It is not clear whether the Yorktowns had a similar capability. The hangar-deck cata-
pult was eliminated when additional anti-aircraft guns were installed blocking its opening, but before that,
it was onboard several Essex-class carriers. Pre-war game rules eventually counted the usual two flight deck
catapults, but not any hangar deck installation. The Yorktown-class hangar was partly open so that aircraft
could be fueled in it. Partial closure was considered desirable to deal with high wind and foul weather.
47
The C&R letter referred to Comairbatfor letter S1-1/FF2-3(1775) of 5 May 1932. Unfortunately, this
letter has not survived in the CNO/SecNavy correspondence file in RG 80, folder CV 1932. The detailed
letters from Air Squadrons Battle Force that have survived from this period do not mention survivability.
48
BuAer seems not to have appreciated the implications of quick repair, because in a 30 July 1934 note to
the Secretary of the Navy (Aer-E-34-SCP CV5&6/S12 in CNO/SecNavy classified files in NARA RG 80
File CV July–Dec 1934) on a proposed further lengthening of flight decks, it referred to the need to be able
to operate with part of the flight deck shot away (paragraph 6), a possibility that explained the need for two-
way arresting gear and the ability to steam astern at high speed (19.6 knots). “Under this damage condition
the length of the remaining intact deck is a measure of the carrier’s residual fighting power which further
emphasizes the importance attached to obtaining the maximum possible flight deck length....” The General
Board approved an additional 65 feet, but no more because additional length added weight, which had to be
subtracted from the third, smaller, carrier (Wasp).
49
The design of the British “armored deck” carriers reflected the utterly different British operating practice.
These ships were conceived when the British faced the prospect of attack from land bases in the Mediterra-
nean and in the North Sea. Up to that point, they had had the same idea as the U.S. Navy: They could gain
air superiority by hitting the enemy’s carriers first. Unfortunately, as the U.S. Navy already well understood,
land bases could not easily be neutralized by air attack. The British idea was that the hangar could be ar-
mored so that a ship’s aircraft could survive an air attack inside it, then emerge to operate (the sides of the
hangar were armored against cruiser fire). The ends of the flight deck were not armored. The cost of such
protection in terms of numbers of aircraft was high: On about the same tonnage, Illustrious had only about
half as many aircraft as the unarmored-deck Ark Royal although she displaced slightly more. Armor also
made for a shorter flight deck, which during World War II turned out to be more dangerous when the ship
operated high-performance U.S. aircraft. Moreover, despite the considerable ingenuity involved in the Brit-
ish design, it could withstand only the 500-pound bombs contemporary British dive bombers could deliver.
The armor was defeated by German Stuka dive bombers dropping much heavier bombs. It would also have
been defeated by contemporary U.S. dive bombers, which normally carried 1,000-pound bombs.
50
Under the last of the inter-war treaties (the London Naval Treaty of 1936), the allowable tonnage of indi-
vidual carriers was reduced from 27,000 to 23,000 tons. The design displacement of the Essex was 27,500
tons. This provided enough weight to add an additional armored deck on the hangar deck. It protected the
ship’s hull from the effect of a bomb passing through the flight deck (the Yorktowns had only a single armor
deck, over their machinery). This arrangement was chosen in preference to a proposal to armor the flight
deck—not so much to protect the flight deck as to protect the ship underneath it. A June 1939 study of a
modified Yorktown showed that flight-deck armor would have weighed 1,460 tons, reducing aircraft comple-
ment by at least two thirds to about 25. Even then, numerous openings in the deck, for example for elevators,
would compromise any protection it offered. An increase in the thickness of the protective deck deep in the
ship to resist a 1,000-pound bomb dropped from 10,000 feet was considered more practicable; it would cost
600 tons. The next step, once war broke out and tonnage limits were abandoned, was the two-deck solution.
51
General Board 420-7 of that date, in NARA RG 80 General Board files.

226 Winning a Future War

52
Wasp was delayed to FY35. Construction of the two Yorktowns was partly funded under the National
Industrial Recovery Act rather than the usual Navy appropriation.
53
General Board documents list Hornet as a single carrier in the second 1939 Deficiency Bill, and in the
spring of 1938 this ship was described as delayed. Hornet was actually ordered on 30 March 1939, the York-
town design being adopted to avoid delay.
54
GB 420-2 Serial 1803. In an accompanying paper (GB 420-2 Serial 1790), written at about the same time,
the General Board pointed out that the two Yorktowns and Wasp were about to enter service, offering the
possibility of comparison between different carrier sizes and also “opportunity for better determination of
the requisite fleet carrier strength. The Board feels that this experience and information should be at hand
before determination of future carrier building, and believes it of sufficient value to warrant omitting carriers
from the 1940 program. The Board at present contemplates including two 20,000 ton carriers in the 1941
program.”
55
General Board study of the FY40 program dated 25 April 1938, responding to the Secretary’s request for
guidance, 5 January 1938. There was certainly a prospect that Navy fighters were about to become fully com-
petitive with land-based aircraft. The Grumman F4F Wildcat prototype had already flown and the Corsair,
for a time the fastest fighter in the world, was about to win the 1938 fighter design competition. Basic policy
was to maintain the force ratios embodied in the earlier treaties, presumably because the 5:3 ratio over Japan
had strategic significance. War gaming showed, however, that the ratio that mattered was the ratio between
total Japanese and American air arms, not the numbers of carriers. At this time, Japan was believed to have
at least five, but possibly more, carriers. That was correct: Japan had the huge Akagi and Kaga, the small
prototype Hosho, the small Ryujo, and the recently completed Soryu (a near-sister was under construction). It
was apparently not known that the Japanese had built auxiliaries and subsidized merchant ships specifically
for wartime conversion into carriers.
56
Enclosure to GB 420-2 dated 5 April 1938 based on the Vinson-Trammell Act and the 1938 supple-
mentary Vinson-Trammell Act; it included 4 rather than 2 carriers because some would be over-age (hence
replaceable) during the ten years. The plan was based almost entirely on shipyard capacity. The Bureau of
Aeronautics was not represented; the plan was submitted by Construction and Repair, Engineering, and
Ordnance. A 6 April 1938 cover letter by the Secretary of the Navy approved the program as a basis for
budgetary planning and particularly the procurement of shipbuilding facilities. In this sense, battleships
and carriers were almost interchangeable. An attached memo from the General Board states that “in view of
the arising questioned value of aircraft carriers relative to increasing improvement in characteristics of land
planes, it is believed that if more carriers are desired they should be built as soon as possible,” hence in FY41.
The four other carriers replaced Lexington and Saratoga in tonnage terms. No tactical rationale was given.
A General Board memo on the program was dated 25 July 1938 (GB 420-2 Serial 1802). It listed the 1939
carrier (Hornet) only because construction had been authorized; “the Board is not in favor of constructing
this carrier at this time, and has stated in previous correspondence that it contemplates including a carrier
in the 1941 program. If a carrier is to be built at this time, the Board believes that it should be …a repeat
design of the Enterprise-Yorktown as it appears extremely doubtful that an efficient all-purpose carrier can be
constructed on less than approximately 20,000 tons.” This would seem to undercut the roughly contempo-
rary statement that carrier construction was being suspended to await experience with two sizes of carrier.
57
GB 420-2 Serial 1860 dated 30 June 1939. Yorktown displaced about 20,000 tons. This paper gave crude
initial characteristics, including armament (not aircraft capacity, however) and speed (35 knots).
58
GB 420-2 file 1939. The supporting analysis is in a paper signed by R.S. Crenshaw, dated 25 September
1939, which refers to earlier “Two-Ocean Navy” programs. Numbers were based on the number of battle-
ships in the two oceans needed to provide at the minimum a parity of battle-line strength at any time in the
area of operation [emphasis in the original]. Numbers of other types were set by the need to support the
battle line and then by the need to carry out other necessary operations. It was assumed that Japan would
ultimately have 12 modern battleships. Germany and Italy together might have 16, although they would
maintain only 8 active ones. That gave a requirement for 16 battleships in the Atlantic and 20 in the Pacific,
in each case to maintain a 5:3 ratio—a total of 36 U.S. battleships. The corresponding figure for carriers

Notes 227
was 18. Note that the initial “Two-Ocean Navy” paper in this package was dated 2 May 1939, before the
outbreak of war, but after it must have been obvious that something was about to happen. It set out various
ratios of superiority which would provide security in various parts of the Atlantic and Pacific.

5. The War College and Cruisers

1
The General Board argued that cruisers were, if anything, more important now that battleship numbers
had been cut drastically by the Washington Treaty. Battleships were now too valuable to risk in many sec-
ondary roles, which in future would be filled by cruisers. This was not a bad prediction; during World War II,
cruisers often acted, in effect, as junior battleships. Cruiser roles were varied enough that it was impossible
to set appropriate numbers. Rather than any particular operational or tactical rationale, the General Board
proposed as a policy that the 5:5:3 ratio in capital ships should apply to all other classes. Quite aside from its
simplicity, this type of calculation was easiest to justify politically, as it did not invoke any particular naval
scenario. During the Washington Conference, the General Board recommended that the lifetime of cruisers
be set at 15 years, which made all existing U.S. cruisers over-age and hence due for replacement (the ten 1916
“scouts” were then under construction). The British adopted the same figure in 1925, and it was embodied
in the 1930 London Naval Treaty. In a 7 April 1923 summary of its preferred future building program
(General Board 420-2, no serial number given), the General Board gave the cruiser strength of the “strongest
power” (Great Britain) and “the second strongest power” (Japan) and stated that 20 10,000-ton cruisers
were a bare minimum for the United States. If more such ships were laid down abroad, the weakness of the
U.S. Navy would worsen. In a follow-on submission dated 18 April 1924 (Serial 1205), the board called for
the construction of 16 10,000-ton cruisers, given British and Japanese programs. Congress authorized eight
ships in 1924, construction of which was to be spread over four years. On 3 April 1926 (420-2, Serial 1271),
the General Board reported again on cruiser policy, pointing out that the United States currently had 11
large and 11 small obsolete cruisers. It did not point to any particular required number, but asked simply
that they be replaced. In a 1927 submission, the General Board called for parity with the British; as of 1928,
they would have a total of 333,380 tons of under-age cruisers, including those under construction. The U.S.
cruiser figure was 95,000 tons (including the first two U.S. 10,000-ton cruisers), leaving a deficit of 238,380
tons—equivalent to 24 of the new cruisers. This seems to have been the origin of the call for 24 cruisers,
including 6 left over from the 1924 authorization. Similar reasoning applied to Japan led to a call for 22 new
cruisers. These figures were produced as part of an attempt to frame a 20-year building program. The Gen-
eral Board file includes a 27 September 1926 letter from the president of the War College (Admiral Pratt) to
Rear Admiral A. T. Long of the General Board, but it is concerned mainly with the advantages of describing
the program as replacements rather than as increases in the Navy. That led to the creation of a list of the old
cruisers, which would be replaced by new ones. The issue, as stated by the college, was whether the obsolete
ships should be retained for their value in wartime or retained until replaced, or scrapped to clear the list of
obsolete tonnage. The old cruisers might retain value as convoy escorts and as flagships for base forces. In
an 11 October 1926 memo for the president of the Naval War College, Captain H. E. Yarnell pointed out
that the old cruisers were too vulnerable to underwater damage to be effective escorts in wartime, and that
to retain them as possible flagships of base or subsidiary forces was hardly justified. They might have some
value as arguments for their replacement, if legislators could be informed as to their condition and their
limited remaining value (otherwise “to many people, including Congressmen, a ship is a ship regardless of
age).” Pratt doubted that they would count in a future armament conference, since he thought that tonnage
ratios would be determined more by the needs of the countries involved than by anything else. Pratt wrote
across the bottom that the memo had been “discussed with the Staff and with the President [himself] and
is agreed to.” A 27 September 1927 General Board submission to the Secretary of the Navy (General Board
420-2 Serial 1358) recommended a five-year program whose size was based on that of the British cruiser
force, to provide 13 8-inch gun cruisers by 1932 with another 15 building, and a total of 33 such cruisers by
1936. This paper responded to a 20 September 1927 request by Secretary Curtis D. Wilbur for a five-year
program based on the 20-year program he had already approved. The only reference in this file to war gam-
ing is a statement, in a December 1927 version of the program, that “every war game, whether played at the
War College or carried out in practice on the high seas [i.e., in Fleet Problems], emphasizes the need for an
increased number of vessels of the cruiser type. The number called for is the result of the experiences of war

228 Winning a Future War

and annual post-war combined fleet exercises and represents the conclusions of commanders-in-chief afloat
and the General Board.” However, it seems clear from the other General Board papers that the figures used
were based on the political argument of parity with the British (and the 5:3 ratio against the Japanese). They
were what the General Board thought Congress might be convinced to provide, not the larger number the
fleet needed. Military considerations entered in the choice that only the largest (10,000-ton) cruisers were
acceptable. In January 1929, when the 15-cruiser bill (HR 11526) was going through Congress, the General
Board attacked a proposed amendment that would have cut it to five cruisers (“inadequate even for training”)
and would have reduced the ships to 7,000 tons (which would have precluded mounting 8-inch guns). It char-
acterized the latter as “a step backward of ten years” leaving a ship “inferior in some respects to the converted
merchant vessel.” This was the 4th Endorsement of GB 420-2 Serial 1400. The proposed amendment also
called for the President to convene a further naval arms conference to be held in Washington on or before
1 January 1930. The 15-cruiser act approved on 13 February 1929 called for construction of five cruisers in
each of FY29–31.
2
They seem to have kept changing the terms of any agreement, so that no British offer was ever acceptable.
Prior to the Geneva Conference, the U.S. Navy’s General Board reviewed U.S. naval policy on the basis of
parity with the Royal Navy; it called for construction of a total of 24 8-inch cruisers. As a first step, the Gen-
eral Board wanted Congress to provide money to build the eight cruisers authorized in 1924. For the role of
the U.S. officers at Geneva, see Stephen Roskill, Naval Policy Between the Wars (London: Collins, 1968), I,
505–13. The British were never able to extract an explanation of why the U.S. Navy wanted large cruisers,
because it was politically impossible for the U.S. Navy to point out that it was designed specifically to fight
Japan. Ironically, the Royal Navy was also being designed to fight Japan, but its representatives apparently
could not tell the U.S. officers that. Because this point could not be discussed openly, the British were un-
able to offer a compromise that would allow the U.S. Navy its large cruisers. British internal naval papers
emphasize the desperate need the British felt to cut the cost of naval construction. A British compilation
of American press material gathered during the run-up to the 1930 conference (ADM 116/2686) reveals
claims that the Geneva conference failed because of effective lobbying by William S. Shearer, who was hired
by the major naval shipbuilders. President Hoover was portrayed as outraged and determined to overcome
the “big Navy” lobby in the interest of peace (the articles include his proclamation of the Kellogg-Briand
Pact, which “outlawed” war). In all of these articles, the British and not the Japanese figure as possible future
maritime enemies, even though all contemporary internal U.S. naval papers emphasize the Japanese as the
only likely future enemy.
3
The one naval officer who was allowed to attend as a delegate, Admiral of the Fleet Sir John Jellicoe, did so
as governor-general of New Zealand. He was the one who pointed out that under the terms of the treaty,
the British would be accepting a reduction below the 70 cruisers he considered essential to safeguard the sea
traffic of the empire (it is not clear how he reached this figure). Ultimately, the treaty called for a reduction to
50 by the time it expired in 1936, although some cruisers scheduled for scrapping were actually retained. In
the case of cruisers, parity was difficult to define because individual cruisers differed so much in armament
and in other qualities; the U.S. tried to create an appropriate “yardstick.”
4
Pratt was supported by a naval technical staff of twelve, including Rear Admiral Moffett, Rear Admiral J. R.
P. Pringle, Rear Admiral H. E. Yarnell, and Rear Admiral A. J. Hepburn, as well as four captains. They in-
cluded Captain A. H. Van Keuren of the Bureau of Construction and Repair, who produced sketch designs
to test the practicality of various treaty proposals (copies of these designs survive). Pringle was president of
the Naval War College. Moffett was the long-serving chief of the Bureau of Aeronautics. Hepburn was chief
of staff of the U.S. Fleet. Yarnell was chief of the Bureau of Engineering (responsible for engines), but in 1928
he had commanded the carrier Saratoga during her surprise attack on the Panama Canal during the Fleet
Problem, and in 1932 he would execute a similar attack by both fast carriers against Hawaii during Fleet
Problem XIII (at that time he would be serving as CinC Asiatic Fleet).
5
See, for example, Stephen Roskill, Naval Policy Between the Wars II (London: Collins, 1976), 67.
6
General Board treaty papers (GB 438 series) include a discussion of “parity” in naval air power, GB 438-1
(Serial No. 1464). It was based on a proposal by Rear Admiral Moffett, chief of the Bureau of Aeronau-
tics, that comparisons should be based on the flight-deck areas of the various fleets rather than on carrier

Notes 229
displacements. In this connection, he considered the huge Lexington and Saratoga poor bargains, as two
carriers of half their displacement offered real advantages. Moffett proposed moving these two ships to the
“experimental” category so that they could quickly be replaced by smaller units. The General Board pointed
out that the United States had consistently maintained that tonnage was the best and most equitable basis
for limitation. Moffet’s comment presumably reflected the new carrier operating practices that Reeves had
developed. The board added that allotting different nations specific carrier tonnages offered equality of op-
portunity, not outcome, in the way the tonnage was used. Looking forward to the London Conference (the
General Board letter was dated 6 January 1930), the General Board added that if the conference sought to
reduce total carrier tonnage below 135,000, the United States should insist on moving the two big carriers
into the experimental (i.e., instantly replaceable) category—but not otherwise. The two carriers were ac-
ceptable only if enough tonnage was left over to buy a sufficient number of carriers. Moffett did not know
(or preferred not to acknowledge) that the British equated carrier capacity to hangar size, so he produced a
table crediting British carriers with far more aircraft than they ever carried—for example, the Furious and
her two half-sisters were given 72 aircraft, whereas their rated capacities were 36 for Furious and 54 for Cou-
rageous and Glorious—and even that was considered extreme. His British figure totaled 348, compared to
216 (including 36 for the tiny Langley) for the United States. Because most of the British carriers fell under
the “experimental,” hence instantly replaceable, category, the British could in theory build seven 13,800-ton
carriers compared to five for the United States, giving them a total of 669 versus 565 aircraft, a 20 percent
advantage. Moffet’s ideas had no impact on the initial U.S. treaty proposal, which was drafted by the State
Department without reference to the General Board, but he certainly had an impact at the conference, where
he was a member of the staff working with Admiral Pratt.
7
The 6-inch gun was generally considered the largest whose shell, weighing about 100 pounds, could be
handled by one man. Although specially built warships would have power ammunition handling, a converted
merchant ship would not. In theory, a 6-inch gun would weigh about 42 percent as much as an 8-incher, and
a mounting would be correspondingly lighter. Mountings would not be quite as much lighter, but the British
designed a 7,000 ton 6-inch cruiser (Leander class) they considered satisfactory. It was generally assumed
that the cost of a ship was proportional to her displacement. The treaty referred to 6.1- rather than 6-inch
guns, because the French used that caliber. However, in the end the French rejected the treaty because it
would have prevented them from building what they considered sufficient numbers of large destroyers (con-
tre-torpilleurs) armed with 5.5-inch guns. These ships would have come under the cruiser category.
8
Cruisers were assigned a lifetime; under the London Naval Treaty of 1930 they could not be replaced until
they were over-age.
9
As a dramatic gesture supporting the conference, both Hoover and McDonald suspended or cancelled new
cruisers. Hoover was attacked for violating the law passed by Congress. Although McDonald’s cancellations
were not entirely reversed (three of the five British cruisers were built), Hoover found himself ordering all
five of the projected FY29 ships (Cruisers 32 through 36). All 15 of the 1929 cruisers were eventually built.
10
The 6-inch gun tonnage was adjusted during the conference, which ultimately allowed the United States to
build three more 8-inch gun cruisers, but required it to delay laying them down. The initial U.S. position is
described in a British memo dated 5 February 1930, in the collection of confidential conference memoranda
in British files (ADM 116/2746).
11
The General Board 420-2 file (naval policy, including building policy) includes a 28 May 1930 letter (coded
Aer-GB CV) from Rear Admiral Moffet, chief of the Bureau of Aeronautics, pressing for more carriers.
Moffett claimed that a U.S. proposal that landing decks be allowed on all combatant ships was defeated,
leading to the proviso that landing decks could be placed only on a quarter of total cruiser tonnage (amount-
ing to 80,850 tons of cruisers), provided that they were not designed primarily as carriers. Moffett also
claimed that it was only with the greatest difficulty that an attempt to cut the carrier allowance (of which
the United States had used only 66,000 of the 135,000 tons) was headed off. He expected further attempts
at any future conference, hence wanted the Navy to build as many air-capable ships as possible before any
further naval arms control conference could meet. That included five 13,800-ton carriers, to be laid down
in 1931–34.

230 Winning a Future War

12
General Board file 420-8, folder on flight-deck cruisers dated 1934 includes notes from Admiral Moffett’s
diary and papers dated 16 June 1933 and initialed G.W.D. On January 25, 1930, Moffet agreed to the ten-
tative U.S. negotiating position except that he pressed for authority to use battleship and cruiser tonnage
for aircraft. He furnished Secretary of the Navy Adams a copy of secret papers regarding aviation, “espe-
cially urging an option be obtained to use cruiser tonnage for aircraft.” At a 30 January 1930 conference of
delegates, he read a paper arguing that the United States was at a disadvantage because the huge Lexington
and Saratoga had consumed so much of the allowed 135,000 tons of carriers; the British would have more
flight decks. He opposed any reduction in overall carrier tonnage. He was willing to stop the construc-
tion of 10,000-ton carriers, but wanted five for the United States; the total 135,000-ton allowance was not
enough. Moffett also pointed out that the British, but not the United States, had merchant ships suitable
for conversion to carriers. A practical way out would be to allow transfer of tonnage from other categories
to carriers. When a delegate suggested simply avoiding the question of carriers altogether, Moffett said that
the conference was intended to extend parity (with the British) across the board, and that under current
conditions the United States would not have parity. Moffett suggested that each country have the option to
use cruiser tonnage, or at least some of it, for 10,000-ton carriers, with guns of 6.1-inch or smaller caliber.
Another possibility would be to allow each nation to build carriers instead of battleships; Moffett “said that
battleships are obsolete vessels, that neither the Congress nor the people would authorize building any more
of them, but that they would authorize building of carriers and therefore authority to transfer battleship
tonnage to carrier tonnage would result in more carriers.” On 1 February, Moffett proposed that the right
be obtained to transfer up to a sixth of allowed tonnage between the categories of carriers, surface craft,
destroyers, and submarines, but this idea was rejected. On 3 February, he opposed the proposal to charge all
carriers of 10,000 tons or less against cruiser tonnage: “neither the Delegation nor the advisors appreciated
the value of aircraft except Admiral Yarnell, and that all hands were thinking in terms of surface ships. He
said that the surface ships we had would last until the next war and then that everything would be aircraft.”
On 6 February, he remarked that the delegates seemed to have come to the conference determined to secure
an agreement regardless of what sacrifices that might entail. The next day, Secretary Adams said that he
was right: It was necessary to get an agreement—“we could not go back without one.” Adams did say that he
would prevent any cuts in the 135,000-ton carrier tonnage allowed in the Washington Treaty (British Prime
Minister MacDonald had proposed to cut it to 100,000 tons, and there was also a British proposal to cut it
to 80,000 tons). Moffett said that he had done all he could to get more carriers, but would have to be satis-
fied with holding the line at 135,000 tons. Adams also said during the conference that in future the United
States might want to build carriers instead of cruisers. By 18 February, Moffett was frustrated. Not only
the delegates, but also Admiral Pratt rejected all attempts to allow battleship or cruiser tonnage to be used
instead for carriers, and the draft treaty also prohibited 10,000-ton carriers. “My only consolation is that I
have done my best.” Yet another possibility was to give the Japanese their desired larger (than 5:3) ratio in
cruisers in exchange for more U.S. carriers. On 11 March, Pratt announced that he had recommended that
aircraft carriers under 10,000 tons be charged to the cruiser category (he stated that all of the naval advisors
had agreed to this before leaving Washington). The British and French agreed (the Japanese were doubtful).
On 24 March, Moffett wrote in his diary that “aircraft will settle next war. Don’t care how many surface
vessels we build. The more the other nations build the less money they will have for aircraft.” On 21 March,
he had asked Captain van Keuren, the ship designer, to produce a sketch of a cruiser with 6-inch guns and
a landing deck; the tentative design was finished the next day by Mr. Bates, the chief civilian ship designer.
Moffett remarked that such a ship would violate the principle of purity of type and the principle that aircraft
carriers should remain out of range, but “I think, however, we may meet the opposition to aircraft carriers
and especially the argument that aircraft carriers must be accompanied by a cruiser. It might be useful in the
transition period between aircraft carriers and cruisers.” He told Admiral Pratt that such a ship might be
produced outside the carrier tonnage—that it would help solve the problem of insufficient carrier tonnage.
Secretary Adams was unenthusiastic, and Yarnell said that guns would interfere with carrier functions.
Pratt was enthusiastic. On 1 April, Moffett was designated to return home, leaving Pratt to handle the possi-
bility of transfer of tonnage. According to his diary, the U.S. delegates wanted to limit aviation at sea as much
as possible. Pratt failed to get the 10,000-ton carrier but promised to push for the right to put flying decks on
all types of ships should the United States wish to do so. Moffett said that it was not what he had wanted, but
might be satisfactory. In a later note, Moffett wrote that the proposal had met with great opposition, leading
to the compromise allowing 25 percent of cruisers to be fitted with flight decks.

Notes 231
13
General Board (NARA RG 80) file 420-7, file 1925–1931, letter coded SC138-7.
14
The Bureau of Aeronautics disagreed vehemently, in a 22 March 1926 letter in the same General Board
file. It saw the Construction and Repair studies differently; according to the letter, it appeared that a fairly
satisfactory ship could certainly be built, to carry 24 torpedo bomber/scouts or 64 fighters. This file includes
the Bureau of Construction and Repair report on alternative carrier designs. Langley (12,700 tons) did not
count in the U.S. total at this time because she was experimental.
15
Letter from RADM C. S. Williams, writing as a member of the General Board, to the Senior Member
of that board, 23 April 1925, in General Board 420-7 file 1926–1931. Williams was President of the War
College at the time.
16
Letter dated 19 September 1930, NWC RG 8 file UNC-1929–1935 (Box 37, folder 8).
17
No such calculation appears anywhere in U.S. records; the point is that the stated demand for about 24
large cruisers was well below what might have been calculated.
Letters quoted in the report on the analysis of torpedoes in cruisers and destroyers, file XTYB (1930-170).
18

The letters themselves have not surfaced either at Newport or in the National Archives in Washington.
19
GB 420-8 of 11 December 1930 in War College file ‘Determination of Characteristics of Cruisers,’ file
XTYB 1930-157 and 1931-38 NWC RG 8, Box 113. The duties were (a) screening; (b) supporting and
opposing attacks by light forces including aircraft; (c) scouting; (d) operating offensively on lines of commu-
nication; and (e) operating defensively on lines of communication. The ships should be evaluated in night
action, day action, and low-visibility day action. This file includes a 25 November 1930 letter from Rear
Admiral Rock (chief of C&R) to Admiral Laning (War College) enclosing a statement to the General Board,
and commenting that the cruiser the Navy would ultimately want would be one that could “stand up and
fight and not be in imminent danger of being put out of action in the first attack, due to absence of necessary
protection.”
20
At this time, the new-construction program was CL 32–38, with the nature of CL-39 not yet established.
CL-32, -34, -36, and -37 were immediately reordered to the new CL-38 design. The new design was so attrac-
tive that it was adopted for three of the five ships of the FY29 program (cruisers 32–36): the cruisers New Or-
leans (CA-32), Astoria (CA-34), and Minneapolis (CA-36) as well as the heavy cruisers of the FY30 program
(CA 37-39). The other two FY29 ships, Portland (CA-33) and Indianapolis (CA-35) had been ordered from
private yards, hence any radical design change would have been very expensive. Their designs were modified
to considerably improve protection. In 1930–31, these changes had not yet been made, so the new design was
referred to as CL- (or CA-) 38 (CL-37 was to be ordered from a private yard). Plans were drawn up to build
CL-39 as a flight-deck cruiser, but that did not happen. In effect, the War College argued that the United
States should abandon building 8-inch cruisers assuming that the flight-deck cruiser was adopted. The decision
not to adopt the flight-deck cruiser thus guaranteed that the United States would build the three permitted
8-inch gun cruisers (CA-39, -44, and -45).
21
The initial War College reply to the question, posed by the Secretary of the Navy, was dated 6 Janu-
ary 1931; it was filed with XTYB (1930-167). Six alternatives were suggested: 10,000 tons (12 to 15 guns,
protected against 8-inch fire between 12,000 and 24,000 yards); 9,000 tons (12 guns, similar protection),
8,000 tons (ten guns, protection against 6-inch guns at the stated ranges), 7,000 tons (nine guns, 6-inch
protection), 6,500 tons (nine guns, slightly better than 5-inch protection), and 6,000 tons (nine guns, 5-inch
protection). All alternative ships were credited with the same speed (32.5 knots) and with the usual torpedo
armament (six tubes), but with different numbers of guns and with different levels of protection. Radius at
15 knots was given as 10,000 nautical miles for the 8,000- to 10,000-ton ships, 9,500 nautical miles for the
7,000-ton ship, and 9,000 nautical miles for the two smallest ships.
22
“Blue Cruiser Requirements,” War College file XTYB/1931-53. The destroyer section included British
destroyers, but only after the Japanese entries.

232 Winning a Future War

23
It seemed unlikely that the Japanese would transfer tonnage from the destroyer to cruiser categories:“[O]n
account of the peculiar geographical situation of Orange … it would seem that destroyers and submarines—
all smaller types—are more valuable to Orange in a defensive role than larger types.”
24
Cruisers were considered key to trade protection, and the only way to get even second-rate ones quickly
would be to convert merchant ships. The standard for suitability was gross tonnage above 4,000 and speed
above 15 knots. Japan had 26 such ships, compared to 83 in the U.S. merchant fleet. None of these ships was
converted for cruiser duties during World War II, but converted cruisers figured in U.S. war plans. Note
that submarines, which would have access to the “safe” trade routes, did not figure in this analysis.
25
Alternatives were (1) 10,000 tons with 8-inch protection and 15 6-inch guns; (2) 9,600 tons with 12 6-inch
guns and 8-inch protection; (3) 8,350 tons with 12 6-inch guns and 6-inch protection; (4) 7,175 tons with
nine 6-inch guns and 6-inch protection; (5) 6,000 tons with nine 6-inch guns and 6-inch protection; (6)
6,000 tons with six 6-inch guns and 6-inch protection; numbers denote scheme numbers. These alternatives
were defined by the General Board and based on design studies by the Bureau of Construction and Repair.
Naval War College game rules made it possible to compare these alternatives and to conclude that (2) offered
greater fighting strength as well as advantages, important in a long overseas campaign, of seaworthiness,
sustained speed, and habitability.
26
We now know that the Japanese considered and rejected a flight-deck cruiser as an alternative to building
the carrier Soryu. The War College logic would not have predicted Japanese construction of any additional
carriers.
27
Two types were compared, one protected against 6-inch fire (24 aircraft) and one protected against 8-inch
fire (12 aircraft). The latter was considered at least a match for an 8-inch gun cruiser (actually superior at
some ranges). The ability of the flight-deck cruiser to cover large areas while scouting and the effect of dive
bombing by its aircraft made it far superior to an 8-inch gun cruiser for scouting and also for offensive
screening, commerce destruction, and convoy work. Its 6-inch guns would be most effective against destroy-
ers in a formation.
28
If the United States built 15 heavy cruisers, they would mount 137 8-inch guns against 104 Japanese. Of
the U.S. cruisers, five would be adequately protected against 8-inch fire, the rest insufficiently protected; but
none of the Japanese ships were so protected (eight were protected against 6-inch fire, four not at all). In a
fight between these two groups, the United States would have an even chance of victory and probably a slight
edge. However, in a fleet action “it is impossible to say positively that any number were adequate.”
29
The last of seven French 8-inch cruisers was Algérie, approved under the 1930 program; the Italians, their
rivals, built their last 8-inch gun cruiser Pola under their 1930–31 program.
30
It did not help that current policy was o use four 8-inch cruisers as flagships. Battleships were considered
unsuitable because they were tied down in the battle line; a flagship should be smaller and faster, with ade-
quate communications and space for an admiral and his staff. By 1939, perhaps because cruisers were needed
so badly for other roles, the battleships were again flagships (of the fleet and the Battle Force). Heavy cruisers
served as flagships of the Scouting Force and of the cruiser components of the Scouting Forceand a heavy
cruiser was flagship of the Asiatic Fleet.
31
There would be a zone of ranges at which the 6-inch gun cruiser could not return 8-inch fire. However, at
long ranges (22,000 to 27,000 yards), 8-inch accuracy would be poor, particularly since the ships were heavy
rollers, and inside 22,000 yards many existing heavy cruisers could be penetrated by 6-inch fire. The new
heavily protected 8-inch gun vessels were a different proposition: They were immune to 6-inch fire outside
8,000 yards (outside 10,000 at a 90-degree target angle). Such a ship could keep a 6-inch gun cruiser under
fire from 29,000 down to 8,000 yards before suffering penetrating hits.
32
The requirements set by war against Red were mentioned only in passing. “We rest on firm ground in
demanding parity with Red for it may be assumed that Red will maintain a safe margin of superiority over
all others.”

Notes 233
33
The U.S. analyst assumed that the British would continue to build 6,500- ton cruisers (Leanders), which
was certainly their intent as they emerged from the 1930 London Naval Conference. However, when the
U.S. Navy and the Japanese decided to build large 6-inch cruisers, the British felt compelled to follow suit
(theirs were the “Town” class). They also invoked an escalator clause to retain some cruisers due to scrapping,
citing the worsening political situation. At the 1935 London Naval Conference, the British tried to avoid
continued construction of unaffordable large cruisers by capping individual cruiser tonnage at 8,000, the
result being their “Colony” class—which turned out to be ruinously limited in growth margin. The United
States designed a class of 8,000-ton cruisers, but war broke out (and the treaty limits lapsed) before they
could be laid down.
34
After the London Conference, Admiral Pratt claimed credit for a clause that created an unlimited category
of “sloops” displacing up to 2,000 tons and armed with guns of up to 6.1-inch caliber. In theory, they could
take over the anti-raider role. The U.S. Navy built two of them, Erie and Charleston, armed with 6-inch guns.
The French, who did not sign the London Treaty (but largely abided by it) laid down ten Bougainville class
“colonial sloops” (of which they completed eight) armed with 5.5-inch guns. The British built less-powerful
sloops, conceived as anti-submarine and later as anti-aircraft escorts, in quantity.
35
Conclusions drawn in the War College paper were somewhat ambiguous. The optimum for a Blue-Orange
war was given as four flight-deck cruisers, six large (9,600 tons) and six small cruisers (6,500 tons) plus
the ten Omaha-class cruisers. For Blue-Red, it was four flight-deck cruisers, two large (9,600 tons) and 18
somewhat smaller (8,250 tons) cruisers, the Omahas being discarded. In both cases construction of 8-inch
gun cruisers ended at 15. A compromise scheme would be four flight-deck cruisers, nine large (9,600 tons)
and 12 smaller (6,500 tons) cruisers. It seemed that although 6,500 tons might be too little, a good flagship
could be built on 7,125 tons. Building four such flagships would reduce the 6,500-tonners to 12 and the
9,600- tonners to six.
36
That is, the U.S. Navy, or at least the War College, expected the Japanese to emphasize the torpedo, both
to wear down the U.S. fleet as it proceeded west through the Mandates, and in any fleet action. This was
entirely without knowledge of the primacy the Japanese were assigning torpedoes, or of their work on the
Type 93 (“Long Lance”) oxygen-powered torpedo. At this time, the U.S. Navy was working on longer-range
torpedoes for battle-line work. CinC U.S. Fleet recommended a torpedo with a range of 15,000 yards and
called for high speed (the General Board wanted at least 35 knots) in destroyers and destroyer leaders. In
August 1931, new settings gave torpedoes the desired long range at 27 knots (they could reach 7,000 yards
at 42 knots). It had to be assumed that other navies were making similar improvements. The U.S. Navy had
been interested in long-range destroyer torpedo attacks on battleships since before World War I, and its
destroyers had the heaviest torpedo batteries in the world (12 tubes in four triple mountings) and a technique
designed to allow them to fire all of their tubes in a single salvo (curved-ahead fire). The War College pointed
out that the role of large long-range torpedo attacks had changed radically since World War I. At that time,
fire-control systems required ships to maintain steady courses as they fired. Mass torpedo attacks could
force battleships to maneuver and thus ruin their fire control solutions. However, modern U.S. and foreign
fire-control systems could cancel out the effects of such maneuvers. To be effective, attacking destroyers had
to ensure that their torpedoes actually reached the enemy battle line, disrupting it either by hitting or by
forcing ships to maneuver so wildly that their fire-control systems could not cope.
37
This was based on 5.75-inch belt armor, which the General Board later reduced to five inches.
38
A slightly later War College account pointed out that World War I experience had shown the value of
protection (Admiral Jellicoe cited the loss of his battlecruisers at Jutland as proof that firepower without pro-
tection might well be pointless). However, that had been obvious since 1916, and only in 1931 was the U.S.
Navy suddenly interested in heavily protected cruisers. It could be argued that foreign navies were reaching
the same conclusions, reflected in the redesign of the French and Japanese cruisers (the British cruisers that
would have been built had the 1930 conference failed were also heavily protected). The new Japanese Moga-
mis were thought to be well-protected. Just why this was believed is not clear, as it was also claimed that the
existing (and much larger) Japanese heavy cruisers were poorly protected.

234 Winning a Future War

39
The tone of the War College memoranda suggests that the idea was very much C&R’s own initiative, but that
its chief thought he had support elsewhere in the Navy. At this time, the British were promoting the idea that
battleships should be pruned back to 25,000 tons and 12-inch guns. The U.S. Navy rejected the idea, partly on
the ground that small ships could not resist unlimited weapons such as bombs and torpedoes. Captain Thomas
Withers made the interesting point that were cruisers to become the new capital ships, the United States would
find itself with a battle line both much slower than the enemy’s and badly outranged (the U.S. Navy would have
6-inch guns on the new cruisers C&R advocated, the enemy 8-inch guns). The situation would be hopeless. At
present, the U.S. battle line was slightly slower than that of the probable enemy (presumably Japan), but it was
not outranged. Moreover, powerful and fast light U.S. forces could overcome the enemy’s slight speed advan-
tage. If the battle lines consisted of cruisers, the enemy would be as fast as the U.S. light forces, and these would
be compelled to remain within the support of the U.S. battle line, possibly even behind it. The 6-inch main bat-
tery was apparently a necessary trade-off for the sheer weight of armor C&R envisaged. A few of the Naval War
College staff did accept the idea that cruiser speed could be reduced to some extent. However, any U.S. cruiser
might have to escape battlecruisers. War College games had featured cruisers running down enemy carriers. In
1932, Orange was credited with one 28.5-knot carrier and Red had three 31-knot carriers. Fast liners suitable
for conversion to carriers were being built. All of these factors made 29 knots the effective minimum cruiser
speed. At this time, the fast-liner argument applied only to Japan, but it is not clear whether this was known.
40
C&R offered a range of possible cruisers, of which Alternative A was the 10,000-ton 8-inch gun cruiser
approved in June 1931 (CA-37–38). The bureau’s preferred design was C, with a maximum speed of 27.5
knots; F had a speed of 29.7 knots.
41
This might mean analysis requested by either the CNO or the General Board, or both. The result was
“Analysis of Use of Torpedoes in Light Cruisers and Destroyers in Major Actions on the Game Board,”
NWC Archive, RG 8, Box 113, file XTYB (1930-170). The writer was Commander A. G. Kirk, USN, later
a major World War II amphibious commander. This document begins with a compilation of staff views on
cruiser characteristics. Its main content is the War College letter to CNO dated 2 April 1931, which refers
to several earlier letters sent in January 1931.
42
Commander Raymond Spruance, then on staff, examined the speed issue in a letter to the War College
chief of staff. He thought that the game board, on which battles began with all forces in position, distorted
the perception of the value of speed that seagoing officers understood. A cruiser should be fast enough to
avoid heavier and more powerful ships and sufficiently protected to engage other cruisers and, with impu-
nity, lesser ships. Spruance rejected the 6-inch gun 10,000-ton cruiser because he doubted that all that
tonnage would offer much additional ability to stand up to capital ship gunfire, or to dive bombing, or to
torpedoes. Numbers mattered, and many cruisers would be needed to screen a large convoy and its escort
of heavy ships as it steamed across the Pacific. A cruiser had to have an ample speed margin over the capital
ships with which she was working, or against which she was fighting. A 27.5knot cruiser would be good for
about 25 knots at sea, and it was well known that the last two knots of a ship’s speed were typically obtained
only under perfect conditions and after some time had been spent working up to full power. A 25-knot
cruiser would have no real margin over British 23-knot battleships and she would be slower than the 26-knot
Japanese battlecruisers. Nor would she have a sufficient margin over 21-knot U.S. battleships. The projected
U.S. destroyers (Farragut class) were designed for 35 knots, and should be able to keep station at 30; how
could a 27.5-knot cruiser back them up?
43
Some of the points raised by the War College were based on fleet experience, as gathered by the War
College to support gaming, rather than on gaming itself. One was that because the 1.1-inch shell would
explode only on contact, pilots would not know they were under fire (most shells would miss), hence would
not be deterred. The existing 5-inch gun might be less effective, but the effect on a pilot of its bursts would
be considerable. It might suffice to break up an attack. World War II experience confirmed this perception;
the 1.1-inch gun had to be given tracer ammunition to make pilots aware they were under fire (the British
had similar problems with their pom-pom). Captain van Auken, who was responsible for analysis of game
lessons, pointed out that C&R was blithely sacrificing a known anti-aircraft weapon that might require
1.1-inch or .50-caliber guns to supplement it, when and if these latter guns were successfully developed. He
was the only one of the staff who pointed out that the 5-inch guns were needed to deal with heavy bombers,
submarines, and close-in destroyers under reduced visibility or at night.

Notes 235
 “Torpedoes on Light Cruisers,” item XTYB/1931–32 in NWC RG 8, Box 113, folder 8, a memo from the
44

President of the War College to the General Board, 16 January 1931.

45
After the Battle of Midway, the cruiser Mogami suffered serious damage from explosion of her torpedoes.
According to the War College paper, “only recently has the College learned that the war heads of unfired
torpedoes are likely to be detonated.”
46
Post-game discussions of Operations III (1932) showed that the question was generally how effectively
torpedoes could be delivered. After this game, for example, Rear Admiral Halligan, a student who had
come from the carriers, said that there was no question that torpedoes were more effective than bombs as
a way of slowing down enemy battleships, but the issue was how well they could be delivered. A destroyer
captain admitted after the same game that after all his experience, he had not the slightest idea of how to
make a successful destroyer attack, night or day, and he thought (incorrectly) that the British were thinking
of abandoning destroyers and torpedo attacks as a battle tactic. Simulated night action showed how sudden
contacts could lead to instant torpedo firing; cruisers in the Blue screen were sunk by Red destroyer torpe-
does. This particular game suggested that night destroyer torpedo attacks would be ineffective, but it was
clear that this conclusion might be nothing more than a consequence of the rules. For example, destroyers
trying to penetrate a screen were unable to get through by evasion. The Research Department, which was
responsible for updating the rules, concluded that they needed review, including comparison with the main
known case of night destroyer attack, the night after Jutland, as well as experience in fleet maneuvers and in
night anti-destroyer gun practices.
47
Typical inter-war Japanese torpedo batteries onboard destroyers were three triple or two quadruple tubes.
The first U.S. destroyers built during the inter-war period (Farragut class) had two quadruple tubes, but
that increased to three such tubes in the Mahan class, and some ships had four quadruple tubes—the heavi-
est torpedo battery in the world. The British typically used two quadruple tubes. As for cruisers, like the
Americans, the British removed torpedo tubes to gain weight for aircraft and anti-aircraft guns when they
modernized some heavy cruisers, but they retained torpedo tubes in new cruisers.
48
The games began with Operations III (November 1931 for the Class of 1932): NWC RG 4, Box 58, Folder
17. The Red Situation is in Folder 18.
49
The game began at the highest level, in which a situation developed that forced the Joint Army and Navy
Board to update the Red War Plan. That gave students a chance to consider national strategy (in the set-up
there was no U.S. government guidance). It was assumed that the war would be ignited by a clash between Red
and Blue economic interests; at the time, the favored explanation for World War I was that it had arisen out
of the clash of British and French versus German economic interests. In much the same way, it was assumed
that war might break out against Japan due to that country’s need for an exclusive market in China. Neither
assumption would probably be accepted now; the Japanese policy of ejecting Westerners from Asia seems in
retrospect to have been a form of a much earlier kind of nationalism. The description of the European situation
was typical of the description of European politics in war game set-ups until the outbreak of World War II. As
the 1930s progressed, the set-ups often interpreted appeasement as preparation for a deal in Latin America at
the expense of the United States. The set-up emphasized the mutual hatreds and suspicions left by the World
War, complicated by economic problems (the Great Depression). “For Red, with a trying situation at home,
and the influence of Purple [Russia] close at hand, the internal political condition is far from satisfactory. Gold
[France] is the historical enemy of Red, though recently an ally, but the economic advance of Gold since the war,
combined with the economic decline of Red, has led to estrangement. Gold fears that Red resents her recently
acquired importance in the European picture. Black’s [Germany] sympathy leans towards Red. Blaming Blue
for keeping victory from her grasp in the World War, she would favor the defeat of Blue as leading to an adjust-
ment of war debts and perhaps of territory. Maroon [Spain] and Portugal would line up with Red. The first is
a rising industrial power, strongly nationalistic with a natural antagonism to Blue because of Blue immigration
and Latin American policies. The defeat of Blue would give Maroon an opportunity for trade expansion and
perhaps for imperialism. Portugal has always benefitted from Red’s friendship and economic assistance and
would unhesitatingly return past favors by benevolent neutrality. Spain (sic) also will favor Red and probably
extend benevolent neutrality.” The attitude of sovereign members of the Red Commonwealth was uncertain.
Emerald (Ireland) would be sympathetic to Blue, but might be restrained by proximity to Red. Crimson (Can-

236 Winning a Future War

ada) would be deterred by the likelihood that she would be invaded and occupied, and would probably lose her
independence. If she stayed out of the war, Blue would have to watch the border, and by her neutrality Crimson
could benefit Red by freely supplying food and essential raw materials. Newfoundland was not then part of
Canada, so it could provide Red with a base. The issue of a neutral Canada presaged an important Cold War
question: what to do in a global war if Cuba remained neutral. War might allow Orange to seek a dominant
position in the Far East, leaving Scarlet (Australia) in trouble, particularly if Red were so weakened as to be un-
able to protect her. In Ruby (India), Red already faced an independence movement; in the event of war Purple
(Russia) might try to ignite an open rebellion. Fear of this situation might be a considerable factor in Red rela-
tions with Orange. Orange would find her opportunities far too good to ignore. She would particularly want
Blue to be crippled, because if Blue emerged victorious she would gain so dominant a position that Orange
expansionism would be stopped, at the least for a very long time. Orange expansionism would be opposed by
Yellow (China) and Purple and Scarlet, but Yellow (China) and Purple were both too weak and disorganized
to be effective. Blue faced hostility in Central and South America, albeit somewhat veiled; assistance would
not be available. Venezuela in particular was sympathetic to Red and would probably assist her in establishing
the Red fleet in the Caribbean. All of these ideas translated to statements of relative Red and Blue strengths
and weaknesses, the standard prelude to decision making in a war game. Elements of Red strength included
her “astute diplomacy” and her control (through her empire) of some of the principal sources of raw materials.
The linkage between a European scenario and the Far East made this Red-Blue war much more complex than
a Blue-Orange war; there was a real question of how much of the Blue fleet would have to be retained in the
Pacific. It is not clear how seriously any of this should be taken; any observer would know how difficult it was
to write a realistic scenario for a war with Great Britain. U.S. naval strategists knew as much.
50
NWC RG 4, Box 57, Folder 33.
51
To simulate poor visibility, only umpires and movers were permitted in the game room. Every effort was
made to expedite action, moves being made rapidly without scoring gunfire. All involved thought that the
game was a good simulation of a night destroyer attack.
52
In all later games, bomb hits had less effect.
53
NWC RG 4, Box 61. The corresponding Tactical III game is in NWC RG 4, Box 62, Folder 27.
54
This was the last strategic game to use the Akron for scouting, as she was lost on 4 April 1933 with great
loss of life, including Rear Admiral Moffett. Her sister Macon, which entered service in June 1933, figured in
a 1935 strategic game (she was lost that year). U.S. Navy interest in such dirigibles persisted as late as 1940.
In the 1934 game, Blue had two rigid airships (Akron and Macon—ZRs), each with five fighters onboard.
They were assigned to the Bissagos Base Group, to warn of any approaching enemy forces. That effectively
kept them out of the game. One sent to observe the Cape Verde Islands found nothing.
55
The Research Department considered Red’s situation analogous to that of British Admiral Torrington in
1690. He was in the Channel facing a superior French fleet, and an important British convoy was homeward
bound from Gibraltar. The enemy had landed in Ireland, where a British army was fighting. To safeguard
communications with the army in Ireland, a detachment of the British fleet was placed in the Irish Sea. Tor-
rington chose to hover near the enemy on the defensive, fighting only when he had a fair chance of winning.
He considered that if he lost he would lose sea control permanently; he was powerful enough that the French
could not chance an attack on him. At Beachy Head, Torrington refused to give the French the chance of a
real decision by disengaging when the wind dropped. He retired slowly to the east, drawing the French to
Dover and making it possible for the convoy to reach port. He was also drawing the French fleet away from
the vital line to Ireland. The French gained almost nothing before they had to retire to Brest. Torrington was
court-martialed but acquitted, and eventually his conduct was approved. In 1779, another inferior British
fleet found itself in a similar position, and again the decision was to hover near the superior French force
without accepting a decisive battle. On the basis of this experience, Red’s choice should be to take the active
defensive, ready to attack whenever presented with a local advantage, using his air and submarines offensively
against the Blue battle line at once, but preserving his own battle line strength. He should also withdraw to
pull Blue away from the critical Red trade routes (and toward reinforcements heading for Red). “The Red
plan should be characterized by finesse, the Blue by determined attack and follow up.”

Notes 237
56
NWC RG 4, Box 66, Folder 3. The discussion at the critique is in NWC RG 4, Box 66, Folder 18.
57
This balance reflected real carrier operating practices, as British carrier capacity had been published. Red
had 54 observation aircraft on carriers (all U.S. observation aircraft were on battleship and cruiser cat-
apults). Blue had 162 fighters versus 120 for Red, 144 versus 84 heavy dive bombers, 96 versus 8 scouts,
90 versus 84 torpedo bombers. Heavy dive bombers could deliver 1,000-pound bombs, the type used at
Midway. Blue had two very large carriers (Lexington and Saratoga), four smaller more modern ones (like
Ranger), and the small, slow Langley. Langley was assigned to the Train, the others to the Covering Force at a
distance from the Train. Two carriers were each assigned to task groups: (i) Cape Verde Raiding Group, (ii)
Carrier Scouting Group, and (iii) Striking Group. Those in the Cape Verde Raiding Group were to join the
Striking Group (initially the two very large carriers) after completing operations in the Cape Verde Islands.
That made four carriers assigned offensive roles. The Research Department critique noted that of these the
carriers in the Cape Verde Raiding Group “did nothing which could not have been done by the cruisers’
planes. They scouted where it was intended to make landings.” During the early part of Blue’s advance toward
the Azores, the Scouting Group carriers maintained a patrol around the Blue Covering Group. Later, they
were ordered to join the Covering Group, which thus included all six fast Blue carriers. This order aborted a
planned attack by these two carriers on the Red battlecruisers. Red had eight carriers: four of about 22,500
tons (54 aircraft each), one of about 14,500 tons (30 aircraft), two of about 11,000 tons (36 aircraft each),
and one of 11,000 tons (30 aircraft). The four large carriers corresponded to the three converted ships and,
presumably, Eagle (which carried considerably fewer aircraft in reality). The 14,500-tonner was presumably
Argus. One of the 11,000-tonners was presumably Hermes. The number of Red carriers made it possible to
split them up: three in the Red Main Body, three in the Raiding Force, and two in the Control (shipping pro-
tection) force. In effect, the Main Body carriers were reserves. The Control Force carriers patrolled the trade
routes between Red and Gibraltar. Red patrol planes out of Madeira found the Blue forces, and they were
tracked by observation aircraft from the Red carriers. Red launched strikes against the Blue Covering Force
and the Blue Main Body. After these strikes, the Red carriers were assigned to resist a large Blue air attack.
58
A separate critique by the Research Department, dated May 1934, is in NWC RG 4, Box 66, Folder 20.
59
A discussion at the critique brought out the War College’s optimism regarding aircraft operations. The
War College rules based scouting range on fuel capacity, but in practice it depended on how well pilots could
navigate. Captain J. O. Richardson (later CinC of the Pacific Fleet) said that no cruiser plane had ever flown
farther than 75 miles and returned by its own navigation, but someone pointed out that in Fleet Problem
XIII, the carriers Lexington and Saratoga maintained a search pattern in all directions out to 125 miles in
daylight, and recovered all their aircraft. These figures were considerably less than the search ranges used at
the War College.
There is a separate tabulation (January 1934) of ‘Casualties from First Air Attacks.’ (NWC RG 4, Box 66,
60

Folder 17). The summary of “outstanding facts and conclusions” is in RG 4, Box 66 Folder 19.
61
The General Board 420-8 file on Flight Deck Cruisers dated 20 February 1934 included as its first item a
précis of General Board papers on the subject dated 26 January 1934. According to this paper, on 18 Decem-
ber 1930 CNO Pratt directed the War College to compare 8-inch and 6-inch flight-deck cruisers. The War
College reported favorably on 2 January 1931, and on 16 January the Bureau of Construction and Repair
provided the General Board with “spring styles” (preliminary sketches) of flight-deck cruisers as a guide
to writing characteristics, the broad requirements reflecting a particular spring style, from which a design
would be developed. A conference on the subject was convened on 19 January, attended by the Assistant
Secretary of the Navy for Aviation, the chiefs of the relevant bureaus, and the director of War Plans. On
20 January, the General Board disseminated a paper comparing various types of cruisers and concluding
that, following the lead of the War College, an experimental flight-deck cruiser should be built. Two days
later, CNO commented on General Board characteristics for conventional cruisers. He wrote that the War
College studies showed that a flight-deck cruiser would be a better all-round cruiser than any conventional
one, armed with either 6- or 8-inch guns. CNO offered his idea of characteristics in a 22 February paper.
On 28 January 1931, Secretary of the Navy Adams wrote that the present need was to develop plans for a
flight-deck cruiser “since Congress has indicated its intention to authorize this type of ship, and it is doubtful
if authorization bill will extend to pure 6-inch cruisers.” Two days later, Adams formally approved the ten-

238 Winning a Future War

tative characteristics laid down by the General Board. At the end of June 1931, the General Board formally
approved including one flight-deck cruiser in the FY33 program then being laid out.
62
Contract plans were detailed enough for shipyards to use them as a basis for preparing bids that would be
used to select a builder to receive a contract, hence the name (the only part not laid out in detail was the ma-
chinery spaces). This set may have been the only one produced between the world wars for a ship ultimately
not contracted for.
63
The 1934 General Board folder on flight-deck cruisers includes a paper written by Admiral Clark (General
Board) dated 4 February 1931 (headed Office of Chief of Naval Operations, not General Board) for CNO
but never signed or submitted it. It pointed to flaws in the War College studies supporting the flight-deck
cruiser. The War College based its evaluation on the result of actual dive bombing with 500-pound bombs
from 1,000 feet, assuming that one 500-pound bomb was equivalent to a 14-inch shell (an 8-inch cruiser
shell was equated to 0.4–0.5 the value of a 14-inch shell). The War College also assumed that 12 aircraft
could dive-bomb simultaneously. However, it was now generally accepted that aircraft would attack from
2,000 feet, halving the number of hits, and that attacks would be made in waves of three or two waves of six,
giving a ship’s anti-aircraft guns longer to fire. That would roughly halve the number of hits, so the overall
effect would be reduced by three quarters. Moreover, the fighter-bombers envisaged as the equipment of the
flight-deck cruisers had not yet dropped any 500-pound bombs in tests. The War College also assumed that
with any bombers still onboard a flight-deck cruiser was superior to a conventional gun cruiser, because even
if the flight-deck cruiser was sunk, the bombers would still be airborne and could still sink the gun cruiser.
Clark also pointed out that Pratt was due to retire on 28 February 1933, hence could never be held responsi-
ble for the construction of any flight-deck cruisers in the FY33 or later programs. Only with the last cruiser
design had U.S. 8-inch gun cruisers been properly armed. If the U.S. Navy did not build the last of the 18
heavy cruisers allowed under the London Treaty, the number with satisfactory armor would be reduced
from five to two (at this time it was assumed that CA-32–36 would be only lightly armored, CA-37 being the
first with satisfactory protection). Moreover, aside from the Lexington and Saratoga, future carriers would be
unarmored (Ranger was still assumed to be the standard), hence would need heavy-cruiser escorts.
64
In General Board 420-8 folder on flight deck cruisers dated 1934.
65
Letter dated 26 February 1936 in CNO/SecNav classified files, NARA RG 80 Folder CC to CF (CF was
the new designation for the CLV).

6. Downfall
1
Hattendorf et al, Sailors and Scholars, 149. They quote Rear Admiral F. B. Upham, chief of the Bureau
of Naval Personnel (which took over the War College), who argued that the War College was “primarily a
technical school for the training and education of line officers,” hence no different in principle than the Naval
Academy and the Naval Postgraduate School. Upham had apparently made this case in a 6 August 1932
letter to CNO Admiral Pratt. The October 1934 action date is confirmed by the Navy directive issued at
that time (information supplied by Curtis Utz, NHHC).
2
Based on the reference in Hattendorf et al, Sailors and Scholars; they would almost certainly have found
anything written somewhat later. It also seems clear that the remaining War College files give little sense
of the laboratory role. Hattendorf and his colleagues were very thorough in their search of the War College
archives, but they did not take it into account, particularly in the story of the College in the late 1930s. Most
of the material used in this book to demonstrate the laboratory role is to be found outside the War College.
3
This memo is coded NC3/XPOD(1934-34) E-ELM/(0). It is not in the classified SecNav/CNO file in
NARA, but it is in the General Board 438 file. It is dated 27 February 1934. It seems to be the only War
College paper in the General Board file on the London Naval Treaty of 1936, which is otherwise quite vo-
luminous.

Notes 239
4
If the British insisted on a change, McNamee offered a counter-proposal: The huge British merchant fleet
offered many candidates for conversion into wartime cruisers. He suggested that every three tons of mer-
chant ships of over 10,000 tons and 20 knots’ speed be equated to one ton of 6-inch gun cruiser tonnage.
5
In his Naval Policy Between the Wars, II, Stephen Roskill makes it clear that the British moved toward
rearmament when the Japanese in North China treated British businessmen with contempt, making it clear
their ultimate goal was to clear the Western powers from the Far East. The General Board was certainly
aware that this was the Japanese goal; it was included in the report requested by the Secretary of the Navy. It
was not in McNamee’s paper, but a Japanese policy to end the Open Door policy in China was often used as
a motivator for Japanese aggression in war games. By 1934, the British had abandoned the “Ten-Year Rule”
(the concept that defense policy could be based on the assumption that there would be no war for ten years).
The most obvious symbol of rearmament was that a committee formed to prepare a position to be used at the
(abortive) League of Nations Disarmament Conference at Geneva (1932–33) was converted into the Defence
Requirements Committee to determine how to fill holes in British defense.
6
The General Board treaty file (438) includes an extensive memo on the conversations with the British (GB
438-1 dated 30 July 1934).
7
The main remnant of the negotiations was acceptance of a Japanese proposal that each signatory (which in
the end did not include Japan) notify the others of new construction long enough in advance to avoid surpris-
es that might trigger an arms race. After 1936, there were attempts to convince the Japanese to sign the new
treaty, the main provisions of which were limitations on the different classes of warships. In particular, the
new treaty set the maximum battleship gun caliber at 14 inches unless a non-signatory (i.e., Japan) failed to
agree. In that case, the maximum would be 16 inches. A further clause allowed for escalation in the size of
battleships in the event that a non-signatory failed to agree to the 35,000-ton limit. Japan refused to agree
to both provisions.
8
The General Board’s file on the London Conference includes a list of acceptable U.S. positions. The need
to defend the current 35,000-ton limit on battleships was emphasized. A British proposal to limit carriers
to 22,000 rather than the previous 27,000 tons was considered acceptable. As for cruisers, “as a maximum
concession and if necessary to obtain agreement, the Delegation may agree to a limitation on the number
of 10,000 ton 6-inch cruisers for the United States and the British Empire and to an approximate 7,500
ton limit on additional 6-inch gun cruiser construction unless an abnormal increase be proposed in the latter
sub-category [emphasis in original].” This was written before it became obvious that there would be no limits
on total tonnage in any category. This paper was dated 4 October 1935; it was approved by the Secretary of
the Navy on 14 October.
9
Kalbfus’s comments are in “Building Program on Expiration of Washington and London Treaties,” File
UNC/1935-26 in NWC RG 8 Series 1, Box 37, folder 8. He was responding to a CNO circular letter sent
on 1 April 1935 and to a 6 July 1935 second request, “as the Navy Department wishes to continue its study
of this question.” The original letter pointed out that there might no longer be restrictions as to types of
ships, number and size of guns, and total tonnage either by categories or global. The two addressees were
CinC U.S. Fleet and the president of the Naval War College. CNO Admiral Standley emphasized that the
fact that such a study was being made “should not go beyond those officers who are required to assist in the
preparation of the replies.” The study was very secret because at this point those in London were still talking
about ratios of tonnage between the major powers. Standley guessed correctly that the talks would collapse.
He enclosed a detailed questionnaire. Questions included how many additional carriers should be built and
also the minimum number of patrol planes that should be available at the outbreak of an Orange war, and
their characteristics. All other combatants were included, and Standley also asked whether new types of
ships were needed.
10
General Board 420-2 dated 18 March 1938. At this point, the program included four more carriers as re-
placements for Saratoga and Lexington, to be built under the FY45 and FY46 programs. The General Board
was far more interested in battleships. General Board hostility to intensive carrier construction continued
into World War II.

240 Winning a Future War

11
The Iowas benefitted from an escalator clause in the 1936 treaty, activated when the Japanese definitely
refused to be bound by its limits. No one was aware that the Japanese were about to build the huge Yamato
class (about 62,000 tons) armed with 18.1-inch guns.
12
When the bill was pending, it was simply a 20 percent increase, which would have meant an additional
27,000 tons. Including some tonnage still available, the United States could have built either one 30,000-
ton carrier or two 15,000-tonners (but the 30,000-tonner would have been illegal). Presumably the figure
actually adopted was intended to provide two more fleet carriers.
13
The British, who could hardly be considered air-minded, had six fleet carriers on order in 1939. They were
intended to supplement six existing fleet carriers. The Japanese ordered two large carriers (Shokaku and
Zuikaku) under their 1937 program, at which time they had six carriers built or building. Another carrier
(Taiho) was ordered under the 1939 program; however, the Japanese also had auxiliaries and subsidized lin-
ers intended specifically for conversion to carriers in wartime, a total of seven ships. Against that, in 1940 the
U.S. Navy had six fleet carriers with another under construction and another (Essex) planned.
14
The Germans used civilian aircraft as prototypes of bombers (to evade treaty restrictions), but there were
entirely legitimate examples of conversions. In the late 1930s, the standard U.S. Army medium bomber was
the Douglas B-18, a military version of the high-performance DC-2/3 airliner.
15
General Board file 449 (aviation) includes a 30 October 1930 memo from Commander, Carrier Division
One (Rear Admiral F. J. Horne), to Commander, Scouting Fleet, describing naval air missions, beginning
with three long-range missions: strategic scouting, patrol, and long-range heavy bombing, all of which would
be performed by large flying boats. Commander, Scouting Fleet agreed that large long-range flying boats
were needed. Horne apparently had in mind something substantially larger than existing flying boats. A
March 1931 endorsement by BuAer supported him, but cautioned that really large flying boats were too
expensive to buy in any numbers.
16
The 1933 competition is described in a history of pre-war seaplane development by Lee M. Pearson, the
BuAer historian, in the November 1952 BuAer Confidental Bulletin (NHHC Aviation History Branch).
Requirements included a 3,000-mile range at 100 miles per hour, a maximum speed of at least 150 miles per
hour, a normal stall speed of 60 miles per hour, and a maximum weight of 25,000 pounds. BuAer chose Con-
solidated’s design, which promised a maximum speed of 170.8 miles per hour and a range of 3,920 miles. The
contemporary Army Air Corps Martin B -10A was credited with a maximum speed of 178 miles per hour
(at sea level, more at altitude), and its range was far shorter. A summary of U.S. naval aircraft characteristics
produced in 1933 (in the General Board 449 file, folder for 1931–34) shows that the main advantage enjoyed
by the PBY (at that time designated P3Y-1) was its high speed: 162 miles per hour, compared to 134.7 for
the P2Y-2, which had only slightly less range (2,900 versus 3,010 miles). The same technology of powerful
air-cooled engines and lightweight structure made it possible for the Imperial Japanese Navy to develop land-
based bombers capable of reaching Pearl Harbor from Japan, in the form of the G3M (“Nell”).
17
The new bomber designation applied on 21 May 1936 seems to have reflected a change in operating concept.
It is sometimes suggested that the key difference was the advent of a precision bomb sight (the Norden bomb
sight, developed by the Bureau of Ordnance). Bombing certainly mattered at the time. The General Board
449 folder for 1934 includes a 1 August 1934 BuAer report on the comparative values of small single-engine
carrier-based aircraft and multi-engine (two- and four-engine) patrol planes. Its charts offer data on both
scouting and bombing missions. Without bombs, a carrier scout bomber could scout out to 859 miles; a
two-engine flying boat, which would cost three times as much, to 3,010 miles, and a four-engine flying boat
(four times the cost) to 3,800 miles. Cost per scouting mile was somewhat less for the flying boats. For the
maximum range (834 miles) of the carrier plane, the flying boats could carry more bombs: 5,500 pounds
for the two-engine flying boat, 8,500 for the four-engine, versus 1,000 pounds for the carrier bomber. The
flying boats offered considerably greater striking power, based on maximum bombing range and bomb load
at three-quarters speed: 4,000 pounds of bombs for the four-engine flying boat at 2,250 miles’ range; 2,000
pounds for the two-engine flying boat at 2,150 miles; and 1,000 pounds for the carrier bomber at 608 miles’
range. The charts gave no hint of relative survivability in the face of enemy fighters. This comparison covered

Notes 241
only the scouting and long-range bombing missions. The 1935 General Board aviation folder (449) recounts
the first test of torpedoes from a patrol plane, in this case a P2Y-2 (the report is dated 24 August 1935).
18
Gassing was common in war games, but was not permitted when war actually broke out.
19
However, there was interest in developing float-plane versions of standard Navy attack aircraft, which
could be tender-based. Some of the pre-war tenders were designed with catapults to launch such aircraft, and
they were tested (but never mass-produced or deployed).
20
For example, the General Board aviation file (449) includes a 31 May 1938 estimate of overall patrol plane
requirements prepared by Rear Admiral F. J. Horne. Horne did not include offensive operations by patrol
seaplanes; he listed their roles as protection of friendly commerce, overhauling of enemy commerce near
our coasts, detection of enemy raids, and strategic scouting. A war against the most likely enemy (Japan)
would be preceded by a pre-war period of tension, during which the Japanese might deploy detached units
into positions to launch air raids against important U.S. positions—not a bad forecast of Pearl Harbor. The
only defense against such deployments would be searching seaplanes, which might be ordered to strike any
enemy force they found. Numbers were defined by the likely character of a raid. The enemy carrier would
probably run into position in darkness to launch from 250 miles away, two hours before dawn, so that its
aircraft would appear over the target area at dawn, achieving surprise. As aircraft performance improved, the
carrier could stand further off. The carrier would probably run in at 25 knots for 24 hours; she would have
to be within 800 nautical miles of her objective at dawn the day before the attack. Anything that forced the
carrier to launch earlier would be advantageous, since the attackers would have to arrive during darkness and
so would probably be far less effective. The need to maintain patrols made it possible to estimate the number
of seaplanes needed. Minimum seaplane range was 3,000 nautical miles, so that they could fly between U.S.
operating bases. Because it was difficult to accelerate seaplane production, Horne wanted enough not only
for initial patrols but also for operations from the closest of the Mandates, the Marshalls, which might be
occupied 60 days after the outbreak of war. To maintain the desired force of 392 aircraft on M+60 would
require a total of 592 seaplanes, which was more than the United States actually had in 1941.
21
The two competitors were the Sikorsky PBS-1 and the Consolidated PB2Y-1, the latter being bought as
the Coronado. Each had wing cells that could accommodate eight 1,000-pound bombs, compared to four
for a PBY. The Consolidated airplane could carry another four bombs on wing racks. This load compared
to a normal capacity of 4,000 pounds for the Army’s standard heavy bomber, the B-17 (with another 4,000
pounds available on wing racks as overload). However, the B-17 was substantially faster (292 miles per hour
at 25,000 feet for the B-17B compared to 255 miles per hour at 19,000 feet for a PB2Y-2). It also had a
much higher service ceiling (36,000 versus 24,100 feet). The seaplane had much longer range: 4,275 miles
maximum (no bombs) compared to 3,600 for the B-17 (no bombs). Seaplanes sacrificed considerable range if
they carried heavy bomb loads: The PB2Y-2 was limited to 1,330 miles with its full 12,000 pounds of bombs.
Later versions of the Coronado were more heavily loaded and sacrificed performance.
22
Ray Wagner, American Combat Planes of the 20th Century (Reno: Jack Bacon, 2004) associates the new
requirement with the advent of the R-3350 engine, which initially offered 2,000 horsepower (it was the en-
gine adopted for the B-29). The competitors were the Consolidated PB3Y and the Martin PB2M, of which
Consolidated asked for cancellation soon after its airplane was ordered (it was working on the B-36 at the
time). In 1938, BuAer described these airplanes as a strong peacetime deterrent, with an expected endurance
of 8,000 nautical miles. Horne’s paper showed a range of 2,500 nautical miles at 140 knots with a bomb
load of 40,000 pounds, or 7,800 nautical miles at 140 knots with 4,000 pounds of bombs. Maximum speed
was given as 240 knots (276 miles per hour). If expectations were fulfilled, this airplane should be capable
of carrying 20,000 pounds of bombs 3,800 nautical miles and returning. In wartime, such aircraft would be
able to strike deep into enemy territory, and would present a constant threat to all enemy operations. Horne
estimated that 48 could be “advantageously employed as distant overseas striking forces and as distant stra-
tegic scouting forces” (six per tender). By the time the PB2M prototype flew, its equipment was considered
obsolete, and it was employed instead as a long-range transport designated JRM-1 Mars. Its significance here
is as evidence of continued strong interest in long-range attack using tender-based seaplanes.

242 Winning a Future War

23
This change is reflected in a revision of the “Naval Aeronautic Organization,” copies of which (for various
fiscal years) are in General Board 449 folder for 1937–39. A 4 September 1937 letter reorganized the Base
Force patrol squadrons into five patrol wings, each a separate administrative command attached to its own
aircraft tender. A further letter dated 24 September 1937 assigned the aircraft to the Scouting Force. A
patrol wing consisted of two or three 12-airplane patrol squadrons (VP). At this time there were only two
large tenders, Wright and the ex-carrier Langley, the other patrol wings using converted World War I mine-
sweepers. At the least this organization required three more large tenders, of which Curtiss and Albemarle
were included in pre-1939 programs. Beginning in 1938, obsolescent destroyers were converted into small
seaplane tenders, in effect superseding the converted minesweepers (AVD, later reclassified as AVP). New
small seaplane tenders (AVP) were ordered, as well as merchant ships for conversion into large seaplane
tenders. The seaplane force expanded rapidly, to 20 squadrons in 1939 and 25 in 1941.
24
The General Board file (420-8 of 1935–36) does not include the fighting power comparison, but it is men-
tioned in two of the enclosed papers (one of which, written in January 1935, thanks Admiral Kalbfus for
his material) and seems to have been decisive. The opposing view was that the fleet needed at least some
smaller cruisers for fleet work. The study itself, dated 10 January 1935, is in the Classified CNO/SecNav
Correspondence file (RG 80) in NARA, Box 261, Folder NC (the War College was coded NC3 in the Navy
filing system) together with a summary produced by the director of Fleet Training. War College Enclosure
H indicated the best way to utilize a given cruiser tonnage; the director cautioned that “these academic
studies represent an estimate of the average expectation for a large number of future engagements [hence]
cannot be relied upon as assuring success in any one.” The study covered ships displacing 8,000 to 10,000
tons, all with the same speed. The director of Fleet Training simplified the War College results somewhat.
Fighting strength would be roughly proportional to some power of the ship’s dimensions: Gun target would
be proportional to the square and life to the cube. Number of guns and protection were taken as varying with
the fourth power of the ship’s dimensions (this factor did not appear in War College estimates). All of this
led to an estimate that fighting power was proportional to the cube of a ship’s displacement. On a fixed total
displacement, the aggregate fighting power of a group of cruisers would be proportional to the square of the
displacement of each of them, which gave the 10,000-ton cruiser a considerable edge over anything smaller.
The War College reached that conclusion in a much more detailed way, including analysis of Orange cruisers.
It pointed out that the best cruiser was roughly the type it had proposed in 1931.
25
Letter from U.S. Fleet Cruisers, Scouting Force, 10 June 1936 in General Board 420-8 file, folder for
1935–36. A sore point at this time was that the torpedo tubes were supposed to have been removed from
eight “soft” heavy cruisers in order to provide space and weight to double their anti-aircraft batteries, but that
had not been done, and there was little prospect that it could be done within limits set by available money
and equipment. The letter (and supporting letters) was therefore as much about possible re-installation of
torpedoes in the eight cruisers as about future policy. This file does not include any rejoinder from the War
College.
26
In 1939, five of the small cruisers were still assigned to Cruisers, Battle Force, alongside eight large 6-inch
cruisers. Another was flagship of Destroyers, Battle Force (and of one of its two destroyer flotillas); a second
was flagship of the other flotilla. A third was flagship of Aircraft, Scouting Force, and a fourth was flagship
of the Submarine Force of the U.S. Fleet. The remaining small cruiser was assigned to the Asiatic Fleet
alongside a larger 8-inch gun cruiser serving as flagship (at that time Augusta).
27
Studies of light cruisers (and gunboats) were requested by the General Board on 22 July 1936, shortly after
the new London Naval Treaty came into force (General Board 420-8 Serial 1718-X, in File 420-8 for 1937-
38). The addressees were the Chief of Naval Operations, the CinC U.S. Fleet, and the president of the Naval
War College (who seems to have been the first to receive the General Board letter, stamped 22 October
1936). Because of the Vinson-Trammell Act, the letter was written as though the U.S. Navy was still bound
by the tonnage limits of the old London Treaty of 1930 (as escalated, to match the British, in 1936), with
no further cruiser replacements available before 1940. The cruisers of the FY38 and FY39 programs were
projected under the Second Vinson Act of 1938, the 20 percent increase in allowed modern tonnage. The
board, then headed by the same Rear Admiral F. B. Upham who had successfully taken the War College out
from under OPNAV, asked how suitable the 10,000-ton cruisers were, and whether more should be built;
the purposes and employment of other cruisers and gunboats; the types considered suitable; and the number

Notes 243
of each type needed. On 5 January 1937, the General Board formally asked for studies of cruisers ranging
from 3,000 tons (as destroyer flotilla leaders) up, based on earlier design work. Unfortunately, the General
Board file does not include the War College’s reply to the General Board’s request.
28
The original letter from the Secretary of the Navy, dated 11 March 1938, simply asks that the Bureau of
Construction and Repair “without delay” prepare a preliminary design for submission to the General Board.
No justification is given. It is even possible that the project originated with the President; early in 1938, there
was interest on his part in fast heavily armed ships. Thus, this project may actually share the same origin as
the fast Iowa-class battleship. The Secretary’s letter was copied to the various material bureaus, but not to
the Naval War College.
29
This paper concludes with a table of two possible designs for ships armed with nine 12-inch/50-caliber
guns, dated 4 April 1938; the paper is marked “for file—Copy of Memorandum left with Board by Capt
Chantry.” A similar memorandum titled “Estimate of Situation: High Speed, Armored Cruisers” is the next
in the file (and shares the language of the first). It is dated 19 April 1938, with the initials “A.P.F.” These
papers are in the General Board 420-8 file for 1937–38. This second memo provides more alternative sets
of characteristics.
30
At this time, President Franklin D. Roosevelt himself was apparently interested in building fast big-gun
ships; he may have been ultimately responsible for the Iowa class. In other cases, files include explicit refer-
ences to the President’s views; the files on the super-cruiser do not. However, the Secretary’s urgent request
suggests presidential interest.
31
Chantry’s Class of 1936 played just such a game, Operations I (December 1935, in NWC RG 4, Box 75,
Folder 11. Blue raided Orange shipping around a focal area off Singapore. The raiders included submarines.
A follow-on version played in October 1936 by the Class of 1937 (in NWC RG 4, Box 80, Folder 2, and
also in War Plans Division files) came closer to Chantry’s ideas. As in the 1936 game, focal areas were the
most profitable in which to operate. In this case, Orange made them too hot using land-based aircraft and
submarines. Blue’s raiding force consisted of heavy cruisers, converted cruisers, and submarines. Lessons of
the game included the inability of submarines to attack commerce if they followed the usual rules of visit
and search—no surprise, given World War I experience. By the time this game was played, Captain Chantry
was back at the Bureau of Construction and Repair; it is not clear whether he was aware of the 1936 game.
The success of Japanese submarines and aircraft in that game would have brought his rationale for Japanese
construction of super-cruisers into question, particularly as it was believed at the time that the invasion of
China was badly straining Japanese finances. The Blue raiding game was played again as Operations III of
the Class of 1938 (November 1937), Operations III of the Class of 1939 (October 1938, and Operations III
of the Class of 1940 (November 1939, in NWC RG 4, Box 91, Folder 22). No corresponding game seems to
have been played by the Class of 1941.
32
The NC (Navy Schools) file in the formerly classified papers of CNO/SecNav for 1927–39 (NARA RG
80) contains the 1935 War College paper on cruiser fighting power (as well as a 1932 paper on cruiser fight-
ing capacity), but no later papers on warship-building policy. The 1937 papers are discussions of the charge
that the game rules unrealistically favored certain fleet tactics (fighting in particular range bands). This is
obviously negative evidence, and as such it can be questioned. There is, however, also no reference to War
College input in General Board cruiser files (420-8) after 1935. Nor do War College files appear to include
studies relevant to the super-cruiser idea.
33
By that time, commanders afloat enthusiastically supported the big cruiser that was generally designated
CA2 to distinguish it from a conventional heavy cruiser (CA1). It was not universally supported, however.
In its review of the FY41 program (22 April 1940, in file 420-2 Serial 1944) the chairman of the General
Board wrote the Secretary of the Navy that “the value of heavy cruisers of 26,000 tons has become increas-
ingly questionable in view of the number of high-speed capital ships built and building by foreign powers;
furthermore, from a building-time and cost standpoint, it is extremely doubtful whether their value to our
Navy as compared to other types urgently needed would be sufficient to justify their construction at this
time.” The same memo complained that the projected heavy and light cruisers were not good enough. The
big cruisers were included in some versions of the projected ten-year program, but they were clearly not

244 Winning a Future War

badly wanted. However, six of them were included in the big expansion program approved in July 1940 (the
70 percent expansion “Two Ocean Navy” Act). On 12 July 1940, CNO Admiral Stark circulated a memo
listing ships as directed by the President, specifically including six of the big cruisers (most ships were types
to be agreed upon later, and there were no specific figures for battleships and carriers). The big cruiser was
ordered in preference to an oversized heavy cruiser armed with 8-inch guns. Any evaluation of choices made
at the time is complicated by the more general reality that program planning proceeded in two stages. The
first was an attempt to produce the needed numbers of cruisers as quickly as possible, which meant deriving
any new designs from existing ones. With the London Naval Treaty dead, the redesigned ships could be
somewhat larger. Thus, the wartime Baltimore-class heavy cruiser was derived from the last pre-war ship
of that type (Wichita), and the standard wartime Cleveland-class light cruiser was an evolved Brooklyn. The
only question was how many to build, and that may have been determined largely by yard capacity. At the
same time, entirely new designs could be developed, such as the Montana-class battleships and the Alaskas.
In March 1940, President Roosevelt pressed for construction of a cruiser-destroyer like the French super-de-
stroyers, but he was brushed off. It seems interesting that although Roosevelt’s naval aide asked how well
destroyer-cruisers would fare against larger light cruisers, the War College was not asked for its views (using
its fire-effect methodology). That was left to the Bureau of Ordnance.
34
They were conceived as part of the Night Attack Force that was part of a fleet concept intended specifically
to fight the decisive battle against the U.S. fleet. Although the War College never envisaged the Japanese fleet
battle strategy (a night-attack force using torpedoes the night before a day battle), the Japanese plan to build
ships for that battle was far more the sort of thing the War College would have envisioned than was Chant-
ry’s argument based on beating off U.S. surface raiders. The Japanese projected two of their super-cruisers,
Hulls 795–796, as part of a program settled in the spring of 1941. Construction was postponed in Novem-
ber 1941, when war seemed imminent. Design work began in 1939.
35
GB 420-8 file 1937–38, in a jacket marked “Medium Displacement Cruiser” 20 February 1939. Given the
initial argument favoring large cruisers, the title is somewhat odd; however, the General Board ended up
endorsing an 8,000-ton cruiser and avoiding further use of the escalation clause in the London Naval Treaty.
The memo is dated 26 May 1938 with the initials A.P.E., and also dated 20 February 1939. It is marked for
Admiral Reeves, who had been acting chairman for a time in 1936, but was now retired.

. Conclusion: Games Versus Reality in the Pacific

1
ADM 205/19, First Sea Lord papers 1939–1945. This is NID LC Report No. 18 dated 21 January 1942,
“United States Naval Efficiency.” It was “intended as a guide to Flag and Commanding Officers who may
have to operate in conjunction with U.S. Navy forces.” Document courtesy of Corbin Williamson.
2
U.S. Naval Attaché Captain Russell Willson described British practices in a December 1937 report. He
was the first foreigner allowed to attend the Tactical School. It used gaming to teach officers current standard
tactics. After 1939, it was also used to develop tactics in response to requirements. However, earlier issues of
the official Progress in Tactics mention experiments on the tactical board at the school. Willson thought the
11-week course would not seem strenuous to students at the U.S. Naval War College. “In general it may be
said that the Tactical School is divorced from study of strategy, logistics, command, the broader principles
of war, technical matters, and the formal estimate of the situation (called an ‘appreciation’ by the British). It
seems that the course may be described as ‘straight tactics’ with particular reference to the employment of
tactical units in accordance with the standard instructions.... On the game board the British forces are gen-
erally handled in accordance with standard instructions ‘so the officer concerned will know what to expect
when he gets out in the fleet’—but the enemy fleet are allowed a free hand ‘because no one can tell what a Jap
might do.’” According to Willson, students received extra points for initiative even if they were taking undue
risks. In December 1937, Willson discussed Royal Navy tactical training at length with a former director
of the Tactical School, who asked him whether U.S. tactical and strategic training were handled separately.
Willson said that they were not, and gained the impression that the British were not entirely happy with
the split between tactical and strategic training, and that they were considering establishment of a research

Notes 245
staff for tactics. RG 38 ONI Series (NARA), 21144 P-11-b, papers on British Tactics and Tactical Policy by
Naval Attaché Captain Russell Willson, USN, report dated December 1937.
3
The problem in the Musashi attack was that attacks by groups of aircraft from different carriers were not co-
ordinated. The problem might be analogized to that of distributing fire properly among multiple battleships,
something that certainly was understood by the inter-war Navy. It is possible that the issue never came up in
the 1930s because it seemed so unlikely that more than one carrier would attack a target at a time. To some
extent, however, it did arise at Midway, when three U.S. carriers were involved. In the 1944 strikes, pilots
concentrated on Musashi because of her evident size and, apparently, because she had already been damaged.
There was no strike coordinator who could distribute fire among the potential targets. The problem was un-
suspected immediately after the attack because the pilots thought they had hit all the Japanese ships and had
inflicted so much damage that the Japanese force had turned back. Evidence that it was going back toward
Leyte Gulf apparently did not reach Admiral Halsey before he turned north. Games were not fought out
in so detailed a way that such issues would have been evident; there were just not enough participants. The
war games had completely failed to predict the pilots’ over-optimism, and they had never raised the question
of how to insure that a mass air attack properly distributed its fire. Fire distribution was a well understood
issue in gunnery battle, but its air equivalent seems not to have been appreciated. The attack on the battleship
Yamato in April 1945 was much better coordinated, probably as a result of the earlier experience.
4
Papers describing the pre-war discussions with the British are in the files of the OPNAV War Plans Di-
vision, NARA II RG 38 Entry 355, Box 115, “Records Relating to Anglo-American-Dutch Cooperation
1938–1944.” Item 1 is material on British-U.S. conversations in London, 1938–39. A draft “Estimate of the
Situation” for a Blue versus Red and Orange war dated 1935 is in Box 53 (item 1).
5
This was very much the view the British had taken during World War I. They found it difficult to provide
sufficient destroyers to escort convoys (or to do other anti-submarine jobs) because without destroyers to
screen it the main (Grand) fleet would be immobilized. It is often forgotten that in the absence of the British
fleet, the Germans would have been free to use their surface forces to attack convoys. Surface ships were
far more efficient than submarines in this task—if they could reach convoys. When two German cruisers
attacked a Scandinavian convoy in November 1917, they wiped it out. A much larger German anti-con-
voy operation involving most of the German surface fleet, mounted in April 1918, failed only because poor
German intelligence failed to discover that the convoy schedule had changed. These two episodes were well
known to inter-war naval officers. They made the connection between a main fleet and trade protection clear.
6
This was not an inevitable approach. The alternative would have been to see weapons not limited by treaty as
an equalizer during the decisive battle. For example, the Japanese spent heavily on torpedoes and their tactics.
A combined-arms approach to the decisive battle might have tipped the odds by using torpedo attacks to
complement heavy-ship gunnery. In 1940–41, the Japanese chose instead to begin forming a torpedo-heavy
night attack force that would attack the U.S. battleships on the eve of battle in a final attempt to even the odds
by attrition. They seem to have envisaged the decisive battle itself as a gunnery fight. It should be added that
Japanese naval aviators, who certainly did not dominate the Imperial Navy, thought that their aircraft already
could and should replace all heavy-gun ships. The evidence is that at the outset of war they did not carry much
weight in the service. Part of the evidence is the pre-war Japanese choice to invest enormous resources in the
super-battleships Yamato and Musashi and two abortive sisters rather than in further carriers beyond the sis-
ters Shokaku and Zuikaku. The Japanese did form the first large carrier task force, which struck Pearl Harbor.
It seems likely that this concentration of naval air power was conceived primarily to attack fixed land targets
rather than for fleet-on-fleet battle. Japanese tactics for fleet-on-fleet battle entailed dispersion for survival in
the face of hostile carriers, but the carriers were to be concentrated to strike land targets such as Pearl Harbor.
7
Alan D. Zimm, Attack on Pearl Harbor: Strategy, Combat, Myths, Deceptions (Havertown, PA: Casemate,
2011), 21–23. Zimm points out that Yamamoto’s view was paradoxical. As chief of the Japanese equivalent
of the U.S. Navy’s BuAer, he had opposed the construction of the Yamato. However, in planning the Pearl
Harbor attack he made battleships his principal targets, to the point that he was willing to abandon the
project if it was impossible to use battleship-killing torpedoes in the relatively shallow water of Pearl Harbor.
Even without those weapons, his dive bombers could have destroyed carriers present in the harbor. Zimm
emphasizes the Japanese view that the attack would be extremely risky. He argues that the sense of risk was

246 Winning a Future War

partly the consequence of successful Japanese development of land-based torpedo bombers like the ones that
sank the British Prince of Wales and Repulse in December 1941. In his view, the Japanese projected their
own capabilities onto the substantial number of U.S. land-based bombers present in Hawaii in December
1941. Elsewhere in his book, he points to the likelihood that an alerted garrison at Pearl Harbor might have
inflicted significant damage both on the Japanese attack force and on the Japanese carriers. This analysis is
based partly on the Naval War College game rules then current. Zimm adds that the Pearl Harbor garrison
had been on a practice alert until 6 December, when it was allowed to stand down.
8
It is only fair to point out that the Japanese had an alternative to carrier aircraft: a substantial force of long-
range, high-performance land-based naval bombers. In 1941, this was a unique asset. The U.S. Army Air
Force had high-performance bombers capable of delivering torpedoes, but their crews did not practice such
attacks and they were not of much concern to the air force leadership of the time. Only later in the war did
the U.S. Navy acquire this land-based capability. At least in theory, the Japanese could shuttle their land-
based bombers and covering aircraft to concentrate in threatened places in their chains of Mandate islands. In
reality, it turned out that U.S. carriers could strike airfields in the Mandates before they could be reinforced
in sufficient numbers. Japanese search capacity was not sufficient to support the shuttle tactic, which was vul-
nerable to surprise attack. In his book on Pearl Harbor, Zimm argues that the Japanese tended throughout
the Pacific War not to have backup plans and thus not to evaluate their plans objectively. He attributes this
lack to Japanese culture, in which failure was virtually unthinkable. The existence of a Plan B would have been
admission that the basic plan might be flawed. As in other aspects of Japanese wartime behavior, however, it
is also possible to attribute this behavior to something more concrete: a failure to adopt anything resembling
the type of education in decision making the U.S. Naval War College provided. That education began with an
estimate of enemy choices and included a test of how well that estimate had been made in the form of gaming.
9
Halsey’s choice not to detach his battleships was very controversial. When the escort carrier force called
for help, Admiral Nimitz in Pearl Harbor asked where they were. Halsey later pointed out that for him the
battleships were integral to his carriers, a point critics probably misunderstood because they did not under-
stand that under some circumstances powerful Japanese surface ships might have endangered the valuable
but fragile fast carriers. The situation was further complicated by Nimitz’s orders to Halsey that emphasized
the destruction of the surviving Japanese carriers, even if that cost considerable damage to the U.S. invasion
force in Leyte Gulf. All of these issues arose in pre-war games. The orders to Halsey seem to have been for-
gotten in the heat of the battle. It did not help that Halsey’s pilots, who had damaged the Japanese surface
attack force mainly by sinking one of its battleships (Musashi), had been grossly over-optimistic in their
evaluation of other damage they had inflicted.
10
We now know that the aim of British World War I naval strategy, as it was understood in 1914, was to force
the Germans to fight on the far side of the North Sea. The idea was that the Germans would not tolerate a
blockade mounted by the British. The Royal Navy war planner drew an analogy with the Anglo-Dutch wars
of the 17th century. In fact, the Germans never seem to have felt compelled to challenge the British blockade.
After World War I, a senior German officer attacked the sterile German strategy. The Japanese could be
expected to fight because the fleet at Singapore controlled the vital choke point (for Japanese trade) of the
Malacca Straits. They threatened trade and Japanese possessions further north. However, it does not seem
that the British ever thought through the possibility that the Japanese would fight closer to home, placing
the British fleet in a difficult position. British assumptions as to where the decisive battle would be fought
(within 24 hours of Singapore) are evident in the detailed calculations they made for required steaming
endurance of new battleships and carriers.
11
The culmination of the year-long Royal Naval College was a week-long game simulating the British Far
Eastern war plan, the results of which were fed back to British war planners. No accounts of these games seem
to have survived. Several versions of the war plan itself survive in British archives as versions of the War Mem-
orandum (Eastern); the 1933 version is ADM 116/3475 and the 1938 version is ADM 116/3673. The closest
equivalent to Miller’s War Plan Orange is Andrew Field, Royal Navy Strategy in the Far East 1919–1939: Plan-
ning for War against Japan (London: Frank Cass, 2004). The 1933 plan envisions the dispatch of the Mediter-
ranean and Home Fleets to Singapore. The combined fleet would arrive 28 days after dispatch was ordered.
If the Mediterranean Fleet did not have to combine with the Home Fleet somewhere in the Indian Ocean, it
could arrive in 22 days. This timing was considerably stretched about 1937, the reinforcement time then being

Notes 247
70 days. Meanwhile the defenses of Singapore were being built up; by 1936, the fortress was expected to be
able to survive any practical scale of Japanese attack. The 1933 memorandum pointed further to the need for
the British fleet to be able to operate from Hong Kong to deny Japan access to resources in China. An inter-
esting part of the memorandum is analysis of the likely actions of various neutrals, including the note that
both U.S. and Japanese officers expected ultimately to have to fight each other—but that the United States
could not be relied on as an ally in the event of war with Japan. Part of the rationale for expecting Singapore
to hold out was a judgment that the torpedo bombers based there would prevent any Japanese carriers from
getting close enough to make sustained attacks. The memorandum does not include any discussion of a battle
against the Japanese fleet, only the comment that the combined British fleet would clearly be dominant. The
situation began to change in 1934. Once it was accepted that the British might also face a European threat,
the requirement was not merely to exhaust Japan by blockade, but also to force a decisive battle to free the
battle fleet to return to European waters. The British signed a naval agreement with the Germans limiting the
latter to 35 percent of British strength, specifically to leave enough British naval power to face Japan. In 1935,
however, the British almost found themselves fighting the stronger Italians in the Mediterranean, and they
feared that in the event of war Japan might attack in the Far East. British intelligence claimed that the Japanese
were maintaining an amphibious attack division that could seize Singapore. The fear of simultaneous crises
east and west is evident in Admiralty papers discussing overall British strategy, in particular the nightmare of
a two-ocean or two-hemisphere war. In that case, the question was whether the Japanese might choose to force
a battle closer to their own home waters, hence within range of land-based aircraft. The British do not appear
to have taken this possibility into account. It should have been evident in discussions of requirements levied
on their own naval aircraft. The 1933 version of the War Memorandum does list possible Japanese courses of
action. The strategic situation changed radically from 1937 on, as the Japanese seized control of the Chinese
coast, dramatically reducing the wartime value to the British of a blocking force based at Hong Kong. The
ongoing war in China also increased Japanese oil consumption and thus made that country more vulnerable
to blockade. The 1937 British war-planning memorandum included several possible cases, one of which was
that the fleet would be tied down by a European crisis or war. At this point, the logic was that the Japanese
fleet would be drawn south by British attacks on Japanese trade and possessions such as Taiwan and the Pes-
cadores. Near Singapore the Japanese fleet would be subject to attack not only by the British fleet, but also by
RAF aircraft based at Singapore. If it could be attrited sufficiently—an echo of what the Japanese hoped to do
to the U.S. fleet—the British fleet could press north. In his account of British strategy, Field points out that
much of this was wishful thinking. It included the hope that no further fleet engagement would be required,
and that British cruisers would be able to strangle Japanese trade as far afield as the U.S. coast. The situation
changed further in 1940 when the French ceded control of air bases in their Indochinese colony (modern
Vietnam) to the Japanese and thus brought large Japanese air forces within range of Singapore. By that time,
too, the notion of moving the fleet to the Far East was no longer viable. Existing accounts of the “Singapore
Strategy” do not indicate any examination of Japanese fleet employment alternatives other than seeking or
shrinking from a fleet battle somewhere in the South China Sea. The expectation of such a battle features in
pre-war British discussions of how many battleships would be needed in the Far East. The British were well
aware that the Japanese might try to seize Singapore before the fleet could arrive. Detailed planning for the
defense of the base prior to the arrival of the fleet seems to have been based on plans developed locally. There
is some indication that the British thought that their large submarine force based at Singapore might hold the
Japanese back (the U.S. Navy based numerous submarines in the Philippines for much the same reason). A
local war game played to test that idea is described in an appendix to Alastair Mars, British Submarines at War
1939–1945 (Annapolis, MD: U.S. Naval Institute Press, 1971). The submarine force had to be brought back
into European waters once war broke out in Europe but not in the Far East in September 1939.
12
The British were also concerned with the potential threat posed by long-range submarines as their fleet
steamed from the Mediterranean to Singapore. For much the same reason, they fitted destroyers with high-
speed mine-sweeping gear, the theory being that the same submarines might lay mines in the fleet’s path. U.S.
sonar installations were more primitive than those of the Royal Navy, and the U.S. Navy certainly gained
considerably from interaction with the British, but the basic sets were of U.S. origin. Naval War College files
include accounts of pre-war U.S. development. The British taught the U.S. Navy to install streamlined hous-
ings that made it possible to use sonar at higher speeds, and they also introduced the chemical recorder, a
means of estimating the proper time to drop depth charges. Other wartime developments in sonar packaging
and displays were of U.S. origin, produced by the wartime coalition of civilian academic researchers. On the

248 Winning a Future War

other hand, the British introduced the forward-throwing Hedgehog mortar, a vital wartime anti-submarine
weapon. The navies that did not spend much on sonars were the French, Italian, and Japanese. All had had
access to nascent sonar technology on an inter-allied basis in 1918. In 1939, the French were impressed by
British Asdic, the implication being that they had no equivalent technology of their own. The first Japanese
active sonar was approved for service in 1933, but it appears that even after that, reliance was placed mainly
on passive devices. The post-war report of the U.S. Naval Technical Mission to Japan mentions considerable
reliance on German assistance provided in 1942, suggesting that little had been done before that.
13
NWC RG 4, Box 27, Folder 25. The game is labelled “Battle of Emerald Bank.”
14
The Japanese demonstrated their prowess in night combat during the Guadalcanal campaign, from Savo
Island (August 1942) onward. The great demonstration of British prowess was Matapan in 1941. The pre-
war U.S. Navy seems not to have appreciated a vital technical point about pre-radar night combat. Battle
planning was based on what was observed just before dark, the assumption being that the enemy would
not change course afterward. Thus, a night attack force had to get into position, get off its attack, and then
leave before the situation changed so much that pre-battle planning was no longer relevant. U.S. observers
sometimes remarked that Japanese commanders seemed to lose their nerve if their plans failed—they were
inflexible. The reality was that their further information was too scanty to be a basis for continued action;
better to retire and fight another day. This perception is evident in Admiral Morrison’s account of the Gua-
dalcanal battles. U.S. experience, for example in the November 1942 battles, was that the advent of radar
did not solve the problem; it was necessary to do a lot more to maintain awareness of what was happening at
night. U.S. lack of perception suggests that the pre-war Navy did not conduct full-scale experiments in night
combat comparable to what the Japanese and the British had done.
15
Admiral Spruance’s choice to emphasize the invasion over the possibility of destroying the Japanese fleet
was an excellent example of a major War College theme: formulation of the mission to be carried out. A di-
rective generally covered several different possible priorities. The first step in formulating a tactical plan—in
the “Estimate of the Situation”—was to decide what the mission really was. Surviving gaming papers empha-
size the need for such formulation, and the problems many students encountered in doing so.
16
Spruance’s replacement did not reflect dissatisfaction. By this time, the Navy was using its two senior fleet
commanders in tandem, one planning an operation while the other executed the previous one. The fleet and
its fast carrier task force were renumbered depending on which admiral was in charge, so that it was Third or
Fifth Fleet and Task Force 38 or 58 with, respectively, Halsey or Spruance commanding.
17
It can be argued that the sheer complexity of some Japanese tactical plans was the feature of the Pacific
War that could most clearly be associated with a Japanese mode of thinking. War gaming stressed the value
of relatively simple plans.
18
The Naval War College technique of estimating enemy courses of action would therefore presumably have
suggested something like the kamikaze offensive had the very limited number of trained Japanese pilots been
known. As it happened, by January 1944 at least, some intelligence officers were aware that Japanese pilot
training had lagged badly. That, in turn, should have affected evaluation of the “Turkey Shoot” in June 1944
(the point of the January evaluation was that Japan could not afford major pilot losses). Whether anyone at
the college or planning level would have taken the next step to assuming that the Japanese would adopt sys-
tematic rather than occasional suicide tactics is another question, but in the spring of 1945, a U.S. account of
current kamikaze tactics made the point that they were actually more economical in terms of Japanese lives
(for a given level of damage) than conventional tactics. The Pacific amphibious force’s analysis of kamikaze
attacks laid out the advantages they offered, particularly greater lethality per attacker, and other contempo-
rary U.S. documents made roughly the same point. To some extent, the invasion of Saipan had demonstrated
that Japanese troops and civilians preferred to die rather than surrender; at Saipan, the Japanese military
forced civilians to commit suicide en masse. This too, should have been an indicator.
19
A surviving example is “Strategical and Tactical Situation Developed in Recent Problems at the Naval
War College,” produced by the Department of Intelligence in November 1935, in NWC RG 4, Box 80,
folder 29. It was almost certainly written by Captain van Auken.

Notes 249
Appendixes
1
This discussion is based on Part B of the June 1929 Maneuver Rules.
2
I have been unable to find sets of 1930 or 1932 rules. The 1933 edition was rewritten to group the rules
more logically, so it is difficult to say whether some of the changes were introduced in 1932. Dates of pages
in the 1940 rules indicate that there was no significant revision of the aircraft section. Through 1929, the
rules included aircraft characteristics, but, after that, they were provided in a separate pamphlet, just as ship
characteristics were separated from the game rules. Separation made it possible to limit changes in the rules
without neglecting changes in aircraft.
3
The major initial air lesson of the Sino-Japanese War that broke out in 1937 seems to have been that un-
escorted high-performance bombers might suffer heavy casualties at the hands of defending fighters. We
now know that this experience led the Japanese to employ naval fighters as escorts for their long-range naval
bombers, and ultimately to develop a very long-range high-performance fighter, the famous “Zero.” It is not
clear whether anyone outside Japan realized how effective Chinese fighter defense was, at least at the outset.
It appears that in Spain the Germans also found that they had to escort their bombers. Again, it is not clear
how obvious this was to outsiders. The Germans in particular also found themselves adjusting their fighter
tactics. Such issues were not taken into account in framing rules for air-to-air combat.
4
This page of the 1931 rules is dated May 1930, hence comes from that year’s rules. A single airplane could
be located approximately at 20,000 yards and accurately at 11,000; a flight of 18 at 30,000 and 15,000 yards.
Ground winds would decrease ranges by 30 percent (force 4 wind).
5
NWC RG 8, file XMAR (1935-12) dated 26 February 1935, “Manuever Rules Pertaining to Aircraft and
Airplane Characteristics—Revision of.” King offered revisions, which the president of the War College for-
mally requested on 2 March. King provided a list of suggestions in a 17 May letter. He noted that the files
of the director of Fleet Training showed very little definite information on which to base a rule for air-to-air
combat. Reports of inter-squadron camera-gun practices gave relative numbers of hits, but not the number
of hits on each attack. The average ratio corresponded closely to the current War College rule, but there was
no basis for estimating the percentage of casualties on each attack. At this time, the Navy estimated the
lethality of gunfire based on experiments with aircraft on the ground. A few years later, the fleet discovered
to its horror that shell bursts previously ruled as lethal did not down drones. The text of the rules seemed to
discourage air operations because of pilot fatigue; King wanted that softened, as otherwise it “might give the
wrong impression to one unfamiliar with aircraft.” He suggested reducing the effectiveness of anti-aircraft
fire when the main battery was firing. The rules already assumed that secondary batteries were adversely
affected by main-battery fire. Anti-aircraft batteries were generally more exposed to the blast of heavy guns.
King rewrote a table showing the effect of fatigue on pilots, noting that the war allowance was 50 percent
extra pilots aboard a carrier, to allow complete rest one day out of three. He added a detailed table showing
time between attacks during air-to-air combat, based on speed advantage (if any) and relative altitude (service
ceiling of attacker or attacked). For example, with zero speed advantage an airplane could make only a single
attack. An appended note pointed out that if the attacker was no faster than the defender it could make only
a single attack if the defender did not want to fight. Even that could be carried out only if the attacker first
sighted the defender from a greater altitude, at least 3,000 feet greater. How many passes the attacker could
make depended on the speed difference. An attacker could make a pass every three minutes if it was 50 miles
per hour faster at its service ceiling. If the speed difference was 100 miles per hour, the attacker could make
a pass each minute. A plane that had to remain on station, such as a spotter, could not avoid repeat attacks;
King allowed ten minutes between them. King added that no such table could be absolutely accurate, since
time between attacks depended on too many factors to be tabulated. However, the current rules were worse.
They allowed pilots to destroy their targets instantaneously, whereas the new rules allowed for the sort of
cumulative damage (in this case, a cumulative probability of fatal damage) typical of other war game rules.
6
The most striking case of special characteristics for a foreign airplane was the British torpedo bomber–
spotter–reconnaissance airplane (TSR), the Swordfish. The college rules accorded it rather better perfor-
mance than it actually had.

250 Winning a Future War

7
The Norden precision bombsight, developed by the U.S. Navy, used gyros to measure apparent target mo-
tion. (However, such bombsights suffered greater inaccuracy because they offered less accurate initial esti-
mates.) With the advent of this bombsight, the U.S. Navy assumed that formations of bombers flying above
the effective limit of anti-aircraft fire could hit moving warships, and such assumptions can be seen both in
the rules of the late 1930s and in internal U.S. discussions of the air threat. The assumption that level bomb-
ing was newly potent also motivated interest in long-range heavy bombers, the first being the PBY Catalina.
War experience showed that this idea was totally unrealistic. It is not clear to what extent pattern bombing
using the Norden sight was ever tried against ships at sea. The rules did not distinguish between a U.S. Navy
equipped with Norden sights and other navies or air forces without them. In 1939, both the United States
and Germany had gyro-based bombsights.
8
There were also rules concerning the use of air-launched torpedoes. A 30-knot ship could evade all torpe-
does dropped more than 3,000 yards away. The assumption, which was certainly correct at the time, that
torpedo bombers had to come so close to their targets explains why they made so few hits in games. During
World War II, several navies, including the U.S. Navy, developed means of dropping torpedoes at greater
altitude and speed, which meant greater stand-off. That greatly increased their chance of hitting, the torpedo
moving at high speed until it entered the water. Pre-war rules did not reflect this possibility. In the U.S. case,
the new kind of torpedo did not enter service until 1944.
9
The table of aircraft characteristics in the 1929 rules included two-seat fighters (F8C-1s), which could dive-
bomb with 500-pound bombs, meaning that they could pull out after diving with them. In the 1930s, the
U.S. Navy introduced dive bombers that could deliver 1,000-pound bombs.
10
The page giving these figures is dated December 1930. It must have been new, since the text refers back to
the qualitative comments of past years. Dive bombers were not expected to lose effectiveness due to a target’s
maneuvers since they dropped their bombs at such low altitudes (this reasoning was not, however, explicit).
During World War II, the Japanese in particular used radical maneuvers (circling at high speed) against
dive bombers on the theory that the bomber actually committed to a particular path at considerably higher
altitude. Horizontal bombers certainly would be affected by target maneuvers. According to the table, by
maneuvering radically at high speed a large target would reduce bombing effectiveness (from 8,000 to 10,000
feet) by 20 percent. In addition, detailed tables were provided giving percentages of near-misses, which were
expected to cause underwater damage.
11
NWC RG 8, file XMAR (1936-48) dated 19 February 1936, “Aircraft Bombing—Notes for Revision
of Rules on.” The paper was written by Head of Operations Department Captain R. A. Dawes, and it was
approved by Admiral Kalbfus, the War College president.
12
This was a controversial change. Commanders of dive-bombing squadrons reported that there was an
optimum attack altitude at which to level off after dropping the bomb, which was not the lowest safe altitude
from an airplane’s structural point of view, but depended instead on anti-aircraft fire. That suggested that
the percentage of hits would not vary widely according to the altitude at which bombers pulled out, because
the greater accuracy due to a lower pull-out would be balanced by greater inaccuracy due to anti-aircraft fire.
At this time, dive bombers were subject to safety precautions: They were not to release bombs below 1,000
feet, they were not to release bombs at more than a 70 degree dive angle, and they were only to do so when the
airplane was in a steady dive without slip or skid. It was argued that in wartime pilots could gain sufficient ac-
curacy by violating the safety instructions to overcome the effect of surface fire. The decision not to consider
wartime accuracy equal to peacetime accuracy was based on remarks by aviation officers at the War College,
three of whom had recently practiced dive bombing. They found 70 degrees a practical limit because it was so
difficult to maneuver at a steeper angle. They also rejected lower-level release on the ground that the concus-
sion of the explosion would wreck the airplane before it could pull out. There was also the morale factor: It
was unlikely that every pilot in an attack would take the steepest dive and release as low as possible, especially
in repeated attacks. This paper also suggested a change in the anti-aircraft rules. Each gun or its equivalent
would be given a percentage chance of hitting based on the average percentage of sleeves (towed targets) hit
per gun in anti-aircraft practices over a period of three years, reduced like the bombing figures by 25 percent
for the “fog of war.” That had to be modified to allow for natural interference when multiple guns were firing
at the same target (gunners might be unable to tell whether their weapons were the ones coming closest to

Notes 251
the target). The problem would worsen if two or more ships were firing at the same target. Arbitrary curves
had to be constructed to allow for this factor. Basic performance in target practice showed a 3 percent hitting
rate at 1,000 feet, falling to 2.2 percent at 9,000 feet. Only later did it turn out that these figures were grossly
deceptive, partly because even hits that tore holes in a sleeve often did not stop a real airplane.
13
It is possible that the sole wartime example of hits by horizontal bombers on fast-maneuvering ships were a
few scored by formations of Japanese bombers on HMS Repulse in December 1941 (masthead-level bombing
was a different matter). Generally, aircraft bombed from excessive altitude; in many cases, ships’ command-
ers could maneuver away from bombs they saw falling. There also seem to have been few cases of bombers
maintaining the necessary rigid formations or the necessary degree of coordination. By way of contrast, dive
bombing required far less coordination, although an attack by multiple dive bombers approaching from
different directions could achieve better results. Inter-war game rules do not seem to reflect these points.
14
When, as chief of BuAer, Admiral King advised the college on revised rules for air attacks, he noted drily
that students at the college could lay out air-deployed smoke screens far more precisely than any pilot could
execute such plans. Smokers (aircraft laying such screens) were featured in pre–World War II fleet exercises,
but they seem never to have been used during the war.
That is, the bomber could set the torpedo to run at an angle rather than straight ahead to the target. This
15

was a standard capability for surface-launched torpedoes, but it was not incorporated in air-launched torpe-
does (if at all) until World War II.
At this time, some discussions of the air-to-air rules included arguments that the flexible gun onboard a
16

bomber or scout was actually superior to fixed forward-firing fighter guns, partly because the fighter pilot
had to divide his attention between flying and firing his gun.
17
This practice was abandoned by the U.S. Navy (but not by others such as the British) early in World
War II, when Commander J. S. Thach proposed his “Thach weave” to counter the extremely maneuverable
Japanese Zero fighter. His four-airplane tactics roughly paralleled those the Germans had developed after
the Spanish Civil War. It turned out that although the second fighter could help protect the first from inter-
ference, the third concentrated far too much on keeping station, and typically it was shot down. Two pairs
of fighters offered far better mutual reinforcement. The gap between exercises, in which the three-fighter
formation seemed perfectly adequate, and reality, when it decidedly was not, give a sense of the limitations
of the exercise-based game rules.
18
It is difficult to square these rules with steps the U.S. Navy was actually taking. In buying new fighters, the
Bureau of Aeronautics emphasized the need to escort strike aircraft to beat off enemy fighters. It would not
have done so had it been nearly so optimistic about the ability of its bombers to defend themselves. On the
other hand, carrier air wing composition emphasized attack over defense, either of the ship or of its strike
aircraft. Typically, a carrier embarked four squadrons: one of fighters, one of dive bombers, one of scouts
(dive bombers, but normally carrying 500- rather than 1,000-pound bombs), and one of torpedo bombers.
When the new Essex class was designed in 1939–40, a major design feature was capacity for a second fighter
squadron. Naval War College archives do not appear to include demands by the Bureau of Aeronautics or
the fleet for a very different assessment of fighter capability.
19
Heavy anti-aircraft guns fired explosive shells with time fuses. A fire-control system aimed the gun and
automatically set the fuse in order to burst the shell near the target airplane. U.S. fire-control systems were
designed to hit airplanes flying in a straight path at constant speed—horizontal bombers but definitely not
dive bombers. Before the system could fire, it had to measure target course and speed by repeated observa-
tion. Again, that suited an attack on horizontal bombers, which had to line up on their targets as they set
their sights. The dive bomber spent too little time diving for this process to be completed, and before it tipped
over into a dive it was generally too far away to hit. Moreover, fire control required observation of shell bursts
so that corrections could be applied to the firing solution that, in turn, was used to aim the guns. The dive
bomber did not provide enough time for that. By way of contrast, a fast-firing automatic gun could adjust
rapidly, its gunner typically watching tracers to correct aim.

252 Winning a Future War

B

Danlinger, Sutherland, and Charles B. Gray. War in the Pacific: A Study of Navies, Peoples,
and Battle Problems (New York: R. M. McBride & Co, 1936).
Evans, David C., and Mark R. Peattie. Kaigun: Strategy, Tactics, and Technology in the
Imperial Japanese Navy 1887–1941 (Annapolis, MD: U.S. Naval Institute Press, 1997).
Felker, Craig E. Testing American Sea Power: U.S. Navy Strategic Exercises 1923–1940
(College Station, TX: Texas A&M University Press, 2006).
Friedman, Norman. U.S. Aircraft Carriers: An Illustrated Design History (Annapolis,
MD: U.S. Naval Institute Press, 1983).
———. U.S. Cruisers: An Illustrated Design History (Annapolis, MD: U.S. Naval Insti-
tute Press, 1984).
———. Network-Centric Warfare: How Navies Learned to Fight Smarter Through Three
World Wars (Annapolis, MD: U.S. Naval Institute Press, 2009).
———. British Cruisers: Two World Wars and After (Annapolis, MD: U.S. Naval Insti-
tute Press, 2011).
———. Fighting the Great War at Sea: Strategy, Technology, and Tactics (Annapolis, MD:
U.S. Naval Institute Press, 2014).
———. British Battleships 1906–1946 (Annapolis, MD: U.S. Naval Institute Press,
2015).
——— and Thomas Hone. American and British Aircraft Carrier Development (Annapo-
lis, MD: U.S. Naval Institute Press, 1999).

Bibliography 253
Gordon, Andrew. Rules of the Game: Jutland and British Naval Command (London: John
Murray, 1996).
Hattendorf, John B., B. Mitchell Simpson III, and John R. Wadleigh. Sailors and Schol-
ars: The Centennial History of the U.S. Naval War College (Newport, RI: Naval War Col-
lege Press, 1984).
Hone, Trent,and Tom Hone. Battle Line: The United States Navy 1919–1939 (Annapolis,
MD: U.S. Naval Institute Press, 2006).
Kuehn, John T.. Agents of Innovation: The General Board and the Design of the Fleet That
Defeated the Japanese Navy (Annapolis, MD: U.S. Naval Institute Press, 2008).
Laning, Harris. An Admiral’s Yarn, introduction by Mark R Shulman (Newport, RI: Na-
val War College Press, 1999—Naval War College Historical Monograph Series, no. 14).
Lillard, John M. Playing at War: Wargaming and U.S. Navy Preparations for World War II
(Dullas, VA: Potomac Books, 2016).
Maurer, John,and Christopher M. Bell. At the Crossroads Between War and Peace: The
London Naval Conference of 1930 (Annapolis, MD: U.S. Naval Institute Press, 2014).
Melhorn, Charles M. Two-Block Fox: The Rise of the Aircraft Carrier 1911–1929 (Annap-
olis, MD: U.S. Naval Institute Press, 1974).
Miller, Edward S. War Plan Orange: The U.S. Strategy to Defeat Japan 1897–1945 (An-
napolis, MD: U.S. Naval Institute Press, 1991).
Moretz,Joseph. Thinking Wisely, Planning Boldly: The Higher Education and Training of
Royal Navy Officers 1919–1939 (Solihull, UK: Helion, 2014).
Morrison, Elting E. Admiral Sims and the Modern American Navy (Boston: Houghton
Mifflin, 1942).
Morton, Louis. Strategy and Command: The First Two Years (Washington, DC: Govern-
ment Printing Office, 1961—U.S. Army in World War II: The War in the Pacific series).
Nofi, Albert A. To Train the Fleet for War: The U.S. Navy Fleet Problems, 1923–1940
(Newport, RI: Naval War College Press, 2010—Naval War College Historical Mono-
graph Series, no. 18)
Peattie, Mark R. Sunburst: The Rise of Japanese Naval Air Power 1909–1941 (Annapolis,
MD: U.S. Naval Institute Press, 2001).
Roskill, Stephen. Naval Policy Between the Wars, Vol. I: The Period of Anglo-American
Antagonism, 1919–1929 (London: Collins, 1968).

254 Winning a Future War

———. Naval Policy Between the Wars,Vol. II: The Period of Reluctant Rearmament (Lon-
don: Collins, 1976).
U.S. Navy Office of Naval Intelligence Historical Section. American Naval Planning Sec-
tion, London (Washington, DC: Government Printing Office, 1923).
U.S. Naval War College. Sound Military Decision (Annapolis, MD: U.S. Naval Institute
Press, 1994— republication of 1942 edition).
Vlahos, Michael. The Blue Sword: The Naval War College and the American Mission 1919–
1941 (Newport, RI: U.S. Naval War College Press, 1980—Naval War College Histori-
cal Monograph Series, no. 4)
Wagner, Ray. American Combat Planes of the 20th Century (Reno, NV: Jack Bacon, 2004).
Wheeler, Gerald E., Admiral William Veazie Pratt, U.S. Navy: A Sailor’s Life (Washing-
ton, DC: Naval History Division, 1974).
Zimm, Alan D. Attack on Pearl Harbor: Strategy, Combat, Myths, Deceptions (Havertown,
PA: Casemate, 2011).

Bibliography 255
I

Aircraft carriers, 4, 6–7, 13–16, 18, 20–21, 23–28, 62, 66, 68, 73–103, 106, 110, 113–16, 126, 128,
132–35, 137–41, 146ff, 149ff, 150–53, 155, 161–63
Origins, 15, 73, 170
British, 1, 15, 22ff, 68, 74, 89–90, 113, 147, 170–71, 174ff
Japanese, 1–3, 13–14, 47ff, 58, 73, 113, 147, 167
“Battle-line carriers,” 98–101
Size of air wing, 2, 78, 91, 99
Armored versus unarmored flight decks, 2–3, 79ff, 101–103
Ideal size, 12, 24, 73, 83–86, 89–91, 97–98, 147
In World War II, 1–2, 9ff, 10ff, 15, 26–27, 40ff, 49ff, 50ff, 168, 170–71, 175–77
Airships, 19, 24ff, 133, 140
Asiatic Fleet (U.S.), 18, 125
Atlantic Fleet (U.S.), 18, 23

Battlecruisers, 36, 59, 61, 137, 171, 197

Cancellation and conversion of Lexington and Saratoga, 12, 23, 73, 91, 110
U.S. consideration of in the 1930s, 148–51
In the Japanese navy, 44ff, 122, 159
Battleships, 14, 17–18, 22ff, 25–28, 31, 36, 41, 50ff, 52, 57–61, 64, 66–69, 76ff, 79–81, 84, 87ff, 88, 96, 98,
104, 107, 110–11, 121ff, 124–27, 131–37, 145–52, 156–59, 161–67, 170–73, 178
U.S. Navy orientation toward, 1, 3–4, 9ff, 11, 15, 21ff, 27, 101, 104–105
Utility vis-à-vis aircraft carriers, 14–15, 98, 100–101, 105–106, 116
Battles
Cape Matapan (1941), 62
Coral Sea (1942), 3, 134
D-Day/Normandy (1944), 1, 17

256 Winning a Future War

 Guadalcanal campaign and associated battles (1942–43), 57, 62, 163, 176, 179, 199
Jutland (1916), 45, 47–49, 134
Leyte Gulf (1944), 92ff, 129, 163, 176–77
Midway (1942), 2–3, 11, 15, 16ff, 27, 49ff, 57, 77, 84, 89ff, 96, 134, 162ff, 176, 193, 199
Pearl Harbor (1941), 15, 26–27, 47ff, 58, 164ff, 166–67, 170–71, 173, 176
Philippine Sea (1944), 11, 15, 168, 176
Tsushima (1905), 163–64
Bureau of Aeronautics (BuAer), 14, 19–20, 75, 78, 85ff, 100–101, 105–106, 129, 141, 151–54
Bureau of Construction and Repair (BuC&R), 19, 83, 85, 90, 93, 102, 115, 119, 140, 155–57
Bureau of Navigation (BuNav), 33, 142
Bureau of Ordnance (BuOrd), 19, 57, 59, 156

Cruisers, 8, 18, 25–26, 29, 45–46, 49–50, 60–62, 64, 66, 79, 101, 107, 109–39, 145–47, 155, 158–59, 168,
176
Cruisers, British, 45–46, 66, 73, 78, 89, 110, 112, 120, 125, 147
Cruisers, “flight-deck,” 15–16, 62, 92, 97, 99–101, 107, 113–18, 123–26, 129–33, 135, 139–41, 145
Cruisers, heavy (“Treaty cruisers”), 57, 84, 107, 110, 112–13, 122–27, 146, 159
Cruisers, Japanese, 122–23, 125, 158
Cruisers, light/scout, 23, 28, 80, 107, 110, 116–17

Destroyers, 10, 25–29, 35, 46, 48–50, 56, 62, 66, 78–79, 87–88, 107, 109–13, 115, 117–18, 120, 122–28,
131–32, 137-138, 145, 150, 155–56, 161, 166ff, 173–74, 177
DumanquilasBay, Philippines, 68, 143

Fleet Problems, 8–10, 13ff, 28, 38ff, 51, 53, 55, 62, 71, 97, 100, 116, 183
Fletcher, Frank Jack, ADM, USN, 15, 73
Flying boats, 70, 90ff, 104, 154
German Dornier Do-X, 24, 70, 152

General Board of the U.S. Navy, 20, 25, 28, 36–37, 41, 58, 65, 97, 100–101, 103, 105–106, 110, 112,
115–21, 124, 127–28, 140–41, 143, 147–48, 150–51, 153
Relationship to the Naval War College, 31–35, 60, 65
Responsibility for warship characteristics, 9, 15, 20, 29, 84–85, 91, 98, 105, 140, 154–55, 159, 168
Gibraltar, 66, 132, 135–36, 165
Guam, 24, 67, 79

Halsey, William F., FADM, USN, 171, 176–77

Imperial Japanese Navy, 1–5, 8–9, 20–23, 44ff, 46, 47ff, 49ff, 61–63, 84, 89ff, 96, 102, 113, 121ff, 126, 141,
151–54, 158
Exercises, 58, 62, 165
Interwar development, 12–17, 23–26, 73–74, 91, 93, 118, 147–48
Planning for World War II, 165–67
In World War II, 84, 160–80

Index 257
Jellicoe, John, Admiral of the Fleet, RN, 45, 47–48, 52

Kalbfus, Edward C. ADM, USN, 70, 143, 144ff, 148, 150

Kamikaze attacks, 2, 79ff, 161, 177–78.
King, Ernest J., FADM, USN, 75, 100, 141, 185

Laning, Harris, ADM, USN, 38ff, 40ff, 42, 55, 58, 63, 69, 83, 97–99, 115–18, 125, 140, 142, 146ff,
173–75
London Conference (1929–30), 61, 64, 93, 111, 112, 113, 117ff, 147
London Naval Treaty (1930), 25, 26ff, 45, 60, 64, 97, 98, 103, 115, 116, 125, 140, 145, 155, 173
London Naval Treaty (1936), 25, 104, 121ff, 143, 146ff, 147, 151, 156, 157, 159

MacFall, Roscoe, CPT, USN, 2, 40ff, 174, 175

Mahan, Alfred Thayer, RADM, USN, 17
Malampaya Sound, Philippines, 68,
Manila, Philippines, 21–22, 24, 66–70, 104, 123, 154
and “through ticket” strategy, 7, 22, 67–72, 104, 123, 154
McCandless, Bruce, RADM, USN, 57
Moffett, William A., RADM, USN, 78, 101, 113–14, 129, 140–41, 155, 168

Nimitz, Chester W., FADM, USN, 2, 6, 160–61, 165, 172, 175–77

Office of the Chief of Naval Operations (OPNAV), 2, 4, 19, 25, 28, 33, 36–37, 41, 58, 61, 63, 65, 107, 112,
142, 155
OPNAV War Plans Division, 4, 6, 21, 41, 63–65, 67, 69–70, 91, 101, 104, 106, 131, 141, 161, 163, 187

Pacific Fleet (U.S.), 18, 135

Pacific War, see World War II
Pearl Harbor, HI (also see Battle of Pearl Harbor), 18, 38ff, 69–70, 104, 135, 154
Philippines, 18, 24, 99, 122, 143, 146, 172
Vulnerability to Japanese attack, 20–22, 25, 53, 67, 71, 167, 173
Recapture of in USN war plans, 21, 66–67, 70, 72, 79, 123, 151, 154, 164ff
Pratt, William V., ADM, USN, 16ff, 41–42, 111–12, 118–19, 152, 168
CINCUS, 111
Chief of Naval Operations, 4, 41, 98, 100, 104,112, 114–16, 118, 127, 140, 142
War College President, 4, 41, 42, 85–93, 97, 100, 111, 141
Arms-control negotiator, 61, 93, 97, 111, 115, 117ff

Reeves, Joseph M., ADM, USN, 2, 14ff, 16ff, 38ff, 60, 83, 102, 141, 187
Innovations while captain of Langley, 93–97
Robison, Samuel S, ADM, USN, 175
Royal Navy, British, 7, 15–16, 29, 35–37, 39, 45–49, 51, 57–58, 63, 66, 68, 74, 79ff, 95–96, 109–10, 147,
149ff, 158, 161–64, 170, 175
Exercises, 35, 48,
Interwar plans, 171–73

258 Winning a Future War

 In World War I, 47, 52, 78, 160

San Diego, CA, 12ff, 18, 28ff, 95, 135, 167

San Pedro (Los Angeles), CA, 18, 111ff, 173
Sims, William S., ADM, USN, 4, 9ff, 22ff, 28–29, 33, 35–39, 41, 44ff, 46–49, 58–61
Commander, U.S. Naval Forces Operating in Europe, 29, 35–37, 111
War College president, 32ff, 35, 37–39, 54–55, 63, 75
Spruance, Raymond F., ADM, USN, 2, 7, 10, 15, 73, 162ff, 176–77
Standley, William H., ADM, USN, 142, 147, 151
Submarines, 5, 7, 10, 25, 49, 62, 75, 80, 88, 91, 96, 100–101, 106, 107, 110, 115, 126, 131, 133, 145, 150,
161, 164, 167
As scouts, 19, 24ff, 28ff, 29
British, 29, 134, 137–38
Japanese, 3, 29, 69–70, 81, 89ff, 99, 123, 169–71, 173–75
Unrestricted submarine warfare, 23, 35, 168–69

“Treaty system,” interwar, 23–27, 68, 70, 101, 105, 116, 121ff, 141, 143, 145–48, 150, 152, 158, 165, 167
Tawi Tawi, Philippines, 68

United States Fleet (interwar U.S. Navy formation), 8, 18, 111, 125, 148, 175

van Auken, Wilbur, CPT, USN, 63–64, 66, 69, 130–31, 143, 163, 178

Vessels and classes

HMS Argus, 22ff, 93
HMS Courageous, 74, 89
HMS Dreadnought, 31–33
HMS Glorious, 74, 89, 171
HMS Hood, 46ff, 57
HMS Malaya, 22ff
HMS Prince of Wales, 27
HMS Queen Elizabeth, 61
HMS Sussex, 120
HMS Vindictive, 89.
IJN Akagi, 73
IJN Hyuga, 44ff
IJN Ise, 44ff
IJN Kaga, 73
IJN Kirishima, 44ff
IJN Kongo, 44ff, 148, 159
IJN Mogami, 123,
IJN Musashi, 14–15, 27, 163

Index 259
 IJN Nagato, 44ff
IJN Ryujo, 113.
IJN Yamato, 14–15, 157
KMS Bismarck, 27, 46ff
KMS Deutschland/ “pocket battleships,” 158
USS Arizona (BB-39), 57
USS Barracuda (SS-163), 28ff
USS Bass (SF-5), 28
USS Brooklyn (CL-40), 117ff, 118, 127, 150, 155–56, 159
USS Canberra (CA-70), 166ff
USS Claxton (DD-571), 166ff
USS Cleveland (CL-55), 156
USS Dixie (AD-14), 174ff
USS Enterprise (CV-6), 76ff, 87ff, 89ff, 98, 103, 140, 150
USS Essex (CV-9), 50, 77ff, 82ff, 98, 102–104
USS Harmon (DE-678), 174ff
USS Holland (AS-3), 28ff
USS Hornet (CV-8), 25, 49ff, 104–105, 107, 146ff
USS Iowa (BB-61), 105, 151, 159
USS Killen (DD-593), 166ff
USS Langley (CV-1), 14ff, 73, 75, 78, 83, 93–96, 137
USS Lexington (CV-2), 12–13, 21ff, 24, 73–75, 83–84, 86, 92, 97, 100, 102–103, 134, 188
USS Narwhal (SC-1), 28ff
USS Nautilus (SS-168), 28ff
USS Northampton (CA-26), 119
USS North Carolina (BB-55), 166
USS North Dakota (BB-29), 33
USS Omaha (CL-4), 116, 120, 122, 141, 155, 156
USS Pensacola (CA-24), 120
USS Randolph (CV-15), 177
USS Ranger (CV-4), 74ff, 76ff, 91–92, 97–98, 101–102, 113, 115
USS San Francisco (CA-38), 57, 120
USS Saratoga (CV-3), 12–13, 24, 38ff, 73–75, 83–84, 92, 97, 116, 134, 188
USS Washington (BB-56), 166
USS Wasp (CV-7), 101–102, 150
USS Yorktown (CV-5), 76ff, 77ff, 97–107, 140, 146ff, 150, 192

War Games
“Battle of “Siargao” (Class of 1923 Tactical III), 79–84.
Class of 1932–34 cruiser games, 129–39

260 Winning a Future War

 Class of 1932, Operations II, 129–31
Class of 1933, Operations II, 64,
Class of 1933, Operations IV, 3–4, 7, 22, 29, 68–72, 142–43
Class of 1934, Operations/Tactics IV, 135–39
Evidence for high aircraft/pilot attrition, 2, 52ff, 81, 92–93, 105, 152, 176–77
Influence on USN war plans, 2, 4, 6–7, 9, 21–22, 33–35, 41–42, 55, 61, 63–72, 129–39,
142–43, 148, 161, 172–73, 178, 180
Relationship to fleet exercises, 2, 4, 6, 8–10, 17, 20, 27–28, 30, 55, 70, 160, 178

Washington Conference (1921–22), 13ff, 111, 113

Washington Treaty (1922), 25, 41, 60, 66, 74ff, 78, 84, 90, 109–111, 114–15
World War I, 1, 7, 11–12, 15–16, 17, 29, 34, 49, 52, 111
World War II, 1–7, 11–12, 23, 26, 73, 84, 160–80

Yamamoto Isoroku, Admiral, IJN, 166–67

Index 261
A  A

N F has published more than 40 books, most of them concerned with
the intersection of defense policy, strategy and tactics, and technology. They include
design histories of most U.S. warships and many of their Royal Navy equivalents, and
Network-Centric Warfare, a history of naval command-and-control and information war-
fare. Dr. Friedman’s Cold War history, The Fifty-Year War, won the Westminster Prize
awarded by the Royal United Service Institute as the best military history book of its
year; his Fighting the Great War at Sea won the Lyman Award from the North Amer-
ican Society of Oceanic Historians; and his Seapower as Strategy won the Samuel Eliot
Morrison award given by the Naval Order of the United States. Friedman’s most recent
book is Fighters over the Fleet, a history of carrier-based fighters and their role in fleet air
defense. Works written under government contract include a history of U.S. Navy air
defense missiles and a history of the joint program to develop the MRAP (mine-resistant,
ambush-protected) vehicle. Dr. Friedman spent over a decade at a prominent defense-ori-
ented think tank, another decade in the office of the Secretary of the Navy, and two years
in Marine Corps Headquarters. He writes a monthly column on “World Naval Devel-
opments” for the U.S. Naval Institute Proceedings, and has published numerous articles
in other defense journals. Dr. Friedman received his Ph.D. in theoretical physics from
Columbia University.
 From Winning a Future War:

“To win in the Pacific during World War II, the U.S. Navy had to
transform itself technically, tactically, and strategically. It had to
create a fleet capable of the unprecedented feat of fighting and winning
far from home, without existing bases, in the face of an enemy with
numerous bases fighting in his own waters. Much of the credit for
the transformation should go to the war gaming conducted at the
U.S. Naval War College. Conversely, as we face further demands for
transformation, the inter-war experience at the War College offers
valuable guidance as to what works, and why, and how.”