Contents Winning Chess Strategies:

Exploiting Weaknesses
Title Page 1
About the Author 3 Cyrus Lakdawala
Introduction 6

1) The Principle of Two (or More)

Weaknesses 8
2) Strong Pawns and Weak Pawns 63
3) A Vulnerable or Dysfunctional Piece
99
4) Exploiting Holes 177 Copyright © 2025 Cyrus Lakdawala
5) Exploiting Weak Color Complexes
207
Winning Chess Strategies: Exploiting
Epilogue 294 Weaknesses
Index of Games and Studies 295

All rights reserved. No part of this

publication may be reproduced, stored in
a retrieval system or transmitted in any
form or by any means, electronic,
electrostatic, magnetic tape,
photocopying, recording or otherwise,
without prior permission of the
publisher.

ISBN (hardback) 978-1-83684-045-9

ISBN (paperback) 978-1-83684-046-6
ISBN (ebook) 978-1-83684-047-3

All sales or enquiries should be directed

to
Popular Chess
Suite 442
Baltic Chambers About the Author
50 Wellington Street,
Glasgow G2 6HJ Cyrus Lakdawala is an International
United Kingdom Master, a former National Open and
American Open Champion, and a six-
e-mail: info@[Link] time State Champion. He has been
website: [Link] teaching chess for over 40 years, and
coaches some of the top junior players
Distributed in North America by in the US.
National Book Network

Distributed in Rest of the World by the Also by the Author:

New In Chess Group through Sunrise
1...b6: Move by Move
Handicrafts, ul. Szarugi 61, 21-002
1...d6: Move by Move
Marysin, Poland
A Ferocious Opening Repertoire
Anti-Sicilians: Move by Move
Popular Chess Series Bird’s Opening: Move by Move
Carlsen: Move by Move
Commissioning editor: Byron Jacobs
Caruana: Move by Move
Book production by First Rank
Publishing, Brighton Fischer: Move by Move
Korchnoi: Move by Move
Cover by Kallia Kleisarchaki
Kramnik: Move by Move
Printed in Estonia by Tallinna Opening Repertoire: ...c6
Raamatutrükikoja LLC Opening Repertoire: Modern Defence
Opening Repertoire: The Slav
Opening Repertoire: The Sveshnikov
Petroff Defence: Move by Move
Play the London System
The Alekhine Defence: Move by Move
The Caro-Kann: Move by Move
The Classical French: Move by Move
The Colle: Move by Move
The Modern Defence: Move by Move
The Nimzo-Larsen Attack: Move by Introduction
Move
The Scandinavian: Move by Move
Old guys like me always believe that
The Slav: Move by Move “ago” is so much better than “today”. I
remember when chess engines and
databases became available in the 1990s,
and I foolishly thought: “This great leap
forward will usher in a Golden Age of
easy study!” I was wrong. Today, the
misguided tendency is for club-level
players to study data in bulk rather than
focusing on easy-to-grasp concepts,
which leave an imprint on the mind and
whose patterns we are capable of
remembering.
The aim of this book – and this series
in general – is to master chess themes
efficiently by understanding concepts
rather than dealing with concrete details.
The key focus of this book will be
technical endgames.
Technical chess is a job that many of
my students believe is beneath them.
They always want to win with brilliant
attacks, but those, of course, don’t
happen that often. It is not a coincidence
that the mastery of technical positions is
the bread and butter of many
professional players. Having good
technique means scoring excellently
against club-level opposition.
In large part, we will try to clarify an
ever-confusing question: what exactly is
a weakness in chess? Weaknesses
generally arise as unintended
consequences from either
overestimation of our position or a lack defender. To do this, we must anticipate
of vision of how the optimistic present the opponent’s plans and come up with
will soon become the unpleasant future. survivable counters.
That is obviously not a fully satisfactory 4) How to convert our opponent’s
answer, but that’s the catch: there isn’t a weaknesses into a full point
simple one. If we gathered all our shed tears after we
Correctly identifying weaknesses is a threw away winning positions –
somewhat advanced job in chess, and it allowing opponents to draw and even
requires practice and experience. So, win – they would equal the water of the
breaking down the aims of this book, we Earth’s combined oceans. One goal of
could note: the book is to train ourselves in the most
1) How to identify a weakness difficult thing in the world: efficiently
Sometimes we can believe a square, exploiting the opponent’s weaknesses
pawn, or piece is weak, yet if there is no and winning a won game.
concrete way to take advantage of it,
then it’s not a real weakness and we Cyrus Lakdawala,
misidentified it. San Diego,
2) How to formulate a plan after we June 2025
correctly identify weaknesses
Chess grants us all the freedom of
religion when it comes to style and our
choice of openings. However, when it
comes to plans in critical positions, only
one religion is allowed: the correct one.
There is no greater danger to our
position than a partially logical plan,
which can be just as bad as no plan at
all.
3) How to defend our own weaknesses
We will also learn how to triage
problems in order of precedence, from
the most dire to the least. When
defending our weaknesses, we learn to
endure all hardships – hunger, thirst, and
poverty – and hunker down with the
stubborn, survivalist mindset of the
 Garry Kasparov – Etienne Bacrot
Sarajevo 2000
Black’s position is a portrait of pain,
Chapter One
for the following reasons:
The Principle of Two (or 1) Black’s dysfunctional bad bishop is
More) Weaknesses a sullen presence, since it is blocked by
its own pawns fixed on the same color.
You’ve managed to create a single In the meantime, White’s bishop
weak pawn in the enemy camp. The radiates power, gazing in both
trouble is, you attack it three times, and directions.
the opponent defends it three times. 2) Black suffers from weak dark
How do you make progress? squares. Why? Because Black’s bishop
The way forward is to try and create a and most of his pawns control only light
second (or even third) weakness, which squares.
then distracts the opponent from the
first. When we manage to create
multiple weaknesses in the enemy 8
*+*T*+*T
position, our opponent doesn’t know 7
OvOl+*+*
which direction we come from, and the
defense stretches to breaking point.
6
*+o+*+o+
An analogy would be the bad hombre
5
+*PoPo+o
in an old black-and-white Western
4 *+*+*P*P
movie, who walks into a bar and then 3 +p+bR*+*
begins to bully a random, mild-
mannered patron by shooting bullets just
2 p+*+*+p+
inches from the patron’s feet. Then the
1 +*+*K*+r
bad hombre announces: “Dance!” That a b c d e f g h
is what it feels like. The defender
doesn’t know in which direction you
will shoot, but you do! White to move
What we attempt to learn in this
chapter is how to destroy the opponent’s 3) White’s protected, passed e-pawn is
defensive harmony and coordination by an asset, while Black’s passed d-pawn is
switching back and forth between unable to move forward without getting
targets. lost.
Game 1 4) White’s king has the option to
travel up the board to d4, while Black’s
is stuck in a defensive posture. 8
*+*T*+*+
5) Black is saddled with multiple 7
OvOl+*+*
pawn targets, the main ones being g6
and a7. You may be asking: “How is a7
6
*+o+*+oT
a target?” Well, White has access to the
5
+*PoPo+o
dangerous plan of b3-b4 followed by 4 *P*+*P*P
Ra3. Black may be forced to push the 3 +*+b+*R*
p+*+*+p+
pawn to a6, after which Black’s bishop 2
becomes an even more miserable
creature than it is now.
1 +*+*K*+r
6) Black is stuck with a hole on g5. a b c d e f g h
Why is this important? Because White
can post his rooks on g5 and g3, after
which Black must watch out for bishop Black to move
sacrifices from White on either f5 or h5.
Conclusion: Black is busted. Still, Please meet Odin, the most powerful
Kasparov’s win won’t be so easy, since deity of the chess pantheon. Kasparov
Black will try to defend all his understood chess a bit less than Caissa,
weaknesses and try to erect a fortress. but a bit more than all other humans of
his time, with the exception of Anatoly
24.Rg3 Rh6! Karpov. The idea behind this powerful
move is to clear the path for a future
Those who are hunted crave darkness.
Ra3. The timing is also important, as
This isn’t exactly the ideal post for a
against anything else, Black would have
rook to lounge in, yet it’s Black’s
the option of playing ...a7-a5.
optimal defensive idea.
On 24...Rhg8? Bacrot probably saw 25...Ke6
the direct refutation: 25.Rg5 Bc8 26.g4!
(crashing through; 26.Rh3 looks logical, 25...a5 is met with 26.b5 cxb5
but after 26...Ke7 27.Rhg3 Kf7 Black 27.Bxb5+ c6 (unfortunately, this is
breaks the pin and, for the moment, forced; the more natural 27...Bc6 loses
keeps his position together) 26...hxg4 to 28.e6+, breaking the connection
27.h5! when Black’s kingside structure between the black king and bishop)
collapses. 28.Bd3.

25.b4!
 27...a6
8
*+*T*+*+
+v+l+*+*
7 The black bishop is in a sorry state,
since all but one pawn sits fixed on the
6
*+o+*+oT wrong color.
5
O*PoPo+o 27...a5 28.b5 is equally hopeless.
4 *+*+*P*P 28.Rb3!
3 +*+b+*R*
2 p+*+*+p+ Destination: a3.
1 +*+*K*+r 28...Kf7 29.Ra3
a b c d e f g h

8
t+*+*+*+
Black to move
7
+vO*+l+*
6
o+o+*+oT
+*PoPo+o
The opening of the queenside benefits 5
the side that doesn’t have a rook on h6,
and Black is completely lost: for
4 *P*+*P*P
example, 28...Rb8 29.Kd2 Bc8 30.Rb1. 3 R*+b+*R*
After a swap of rooks, White’s king can 2 p+*K*+p+
+*+*+*+*
walk over to a4 and then pick off 1
Black’s weak a-pawn. Or White can
simply move his g3-rook to a3, with a b c d e f g h
similar results. Black cannot hold on to
two weaknesses that are so far apart.
Black to move
26.Kd2 Ra8
Black is a beaten-down nation, yet
Of course, this is just for show, since
there remains reasonable hope for a
Black isn’t really threatening to play
fortress draw since, for now, White’s
...a7-a5.
pieces lack a clear method of entry.
27.Rb1! However:
1) Bacrot’s pieces are tied down to
The g3-rook is perfectly placed, so the defending the weaknesses on a6 and g6.
other rook starts a journey towards a3. 2) There is potential for sacrificing a
piece on f5.
 3) If all else fails, White can play for a
pawn break on b5. 8
t+*+*+*T
29...Rhh8 30.Rg5!
7
+vO*+l+*
6
o+o+*+o+
This way Black must watch out for 5
R*PoPoRo
*P*K*P*P
potential piece sacrifices on f5 or h5. 4
30...Rh6 31.Kc3 3 +*+*+*+*
When our opponent is doomed to
2 p+*+b+p+
passivity, we should make all the small
1 +*+*+*+*
available improvements before a b c d e f g h
delivering a blow. Why not put the king
on the more active d4-square?

31...Rb8 32.Ra5! Black to move

Threatening Bxa6. Clearing the third rank. This is

White’s best possible set-up. The king is
32...Ra8 33.Kd4 Rhh8 34.Bc2 on the best available square, the bishop
targets both flanks, and the third rank is
Kasparov shuffles. This is partly cleared, meaning that both rooks can
seeking to reach the time control on swiftly move from one wing to the
move 40, which then adds a full hour to other.
his clock and gives him time to leisurely
work out the details. But it is also a good 36...Rhb8
strategy on its own, as one black
inaccuracy is all White needs to pounce. This doesn’t make much sense, but it’s
not like anything does.
34...Rab8 35.Bd3 Ra8 36.Be2! On 36...Rh6 White switches his focus
to the queenside with 37.Rg3!. Our car’s
rearview mirror announces the dire
warning: “Objects are closer than they
appear!” Black’s a-pawn is doomed
when the g3-rook swings to a3.

37.Ra3!

Now that all of Black’s forces are

concentrated on the queenside, 41...Kf6 42.Rxg6+!? Rxg6 43.e7! Rg8
Kasparov switches back to the kingside, 44.e8=Q Rxe8 45.Bxe8 the endgame is
preparing Rag3. obviously hopeless for Black.

37...Rh8 8
*+*+*+tT
A sad necessity. 7
+vO*+l+*
If 37...Bc8 then the threat to the b4- 6
o+o+*+o+
pawn can be ignored with 38.Rag3!. The 5
+*PoPoRo
*P*K*P*P
point is that after 38...Rxb4+ 39.Kd3 4
Rxf4 40.Rxg6 Black must give up an
exchange by putting the rook on g4, or
3 +*+*+*R*
else White has a winning attack. 2 p+*+b+p+
38.Rag3 Rag8
1 +*+*+*+*
a b c d e f g h
38...Rh6 also loses, to 39.Rxg6! Rxg6
40.Bxh5 Rag8.
White to move
8
*+*+*+t+
7
+vO*+l+* Exercise: There will be no happy
6
o+o+*+t+ fortress draw ending for Black. Do
you see White’s combinational
5
+*PoPo+b breakthrough idea?
4 *P*K*P*P
3 +*+*+*R* Answer:
2 p+*+*+p+ Sacrifice/overloading/simplification.
1 +*+*+*+* 39.Bxh5! 1-0
a b c d e f g h
Bacrot did not see any reason to
continue. Black’s problem is that taking
the bishop either way allows White the
White to move
same combination: 39...Rxh5 loses to
41.e6+! is a classic case of 40.Rxh5 gxh5 41.e6+! Kf8 42.e7+!
overloading the defender. Black’s king when Black collapses, and 39...gxh5
is tethered to the g6-rook and after 40.Rxg8 leads to the same thing.
Instead, trying to hold on to the g6-pawn
with 39...Rh6 loses in many ways,
among them 40.e6+! Kf6 41.Rxg6+!, 8
t+vWlV*T
transposing to the note on 38...Rh6. 7
+o+*OoOo
Pay close attention to how Kasparov
sent the defenders scurrying back and
6
o+m+*M*+
forth, between attacking the kingside
5
+*+o+*+*
weaknesses on g6, f5, and h5, and the 4 *+*P*B*+
a6-pawn. Eventually, Black’s 3 +*NbP*+*
pP*+*PpP
coordination got messed up and allowed 2
White a winning breakthrough
combination
1 R*+qK*Nr
The big difference between the two a b c d e f g h
camps was the number of weaknesses,
and that’s also exactly what allowed
White to be so much faster in traveling Black to move
from one side of the board to the other.
When you don’t have to be tethered to 7...g6?!
your own weaknesses, you experience a
completely different level of freedom. Fianchettoing the bishop is
positionally questionable when White
In the following game, Black’s pieces has a stable pawn on d4.
lost all their mobility and were pretty
8.Nf3 Bg4
much frozen in place on the queenside.
Even so, this wasn’t enough for White to Black doesn’t mind giving up the
win until he began operations on the bishop pair, since the game, for now, is
kingside, in search of a second closed. However, as we learned
weakness. elsewhere in this series in Exploiting the
Game 2 Bishop Pair, the structure remains
Alexander Ipatov – Jaime Cuartas somewhat fluid and despite the current
Barbera del Valles 2008 symmetry, it will at some point change
when the bishops will get to prove their
1.d4 d5 2.c4 c6 3.cxd5 cxd5 4.Bf4 Nc6 value.
5.e3 Nf6 6.Nc3 a6 7.Bd3
9.h3 Bxf3 10.Qxf3 Bg7 11.0-0 e6

Black usually castles, yet ...e7-e6

generally needs to be tossed in at some
point. squares is hard for some people. If we
think logically about it, the queen wasn’t
12.Rac1 doing much on f3 and will be put to
more efficient use on the queenside.
8
t+*Wl+*T 14...Nf6
7
+o+*+oVo
o+m+oMo+
6 Admitting that the little excursion to
h5 was a waste of time. 14...e5? only
5
+*+o+*+* gets Black a doomed isolani: 15.dxe5
4 *+*P*B*+ Bxe5 16.Bxe5 Nxe5 17.Bc2!. The
3 +*NbPq+p bishop goes to b3, and Black is unable
to keep the d5-pawn. The only try would
2 pP*+*Pp+ be 17...Nf6 18.Bb3 Nc4, but then
1 +*R*+rK* 19.Nxd5! is possible anyway.
a b c d e f g h
8
t+*W*Tl+
7
+o+*+oVo
Black to move 6
o+m+oMo+
12...Nh5?!
5
+*+o+*+*
4 *+*P*+*+
Principle: A knight on the rim is dim
(or grim!). This doesn’t make sense,
3 +*NbP*+p
since White’s bishop has a place to hide
2 pP*+*PpB
on h2. 1 +*Rq+rK*
Sometimes it’s best to just walk away a b c d e f g h
from confrontation, saying “This is not
my fight.” Indeed, Black is better off
with 12...Nd7!, aiming to castle. If
White to move
13.Bd6 then Black can play 13...Bf8.
Even the immediate 12...0-0 is a better 15.Na4!
approach, despite the somewhat
annoying pin with 13.Bg5. Both c5 and b6 are weak.

13.Bh2 0-0 14.Qd1 15...Nd7

Returning pieces to their initial Practically forced. If the white knight

is allowed to reach c5, then the black Principle: Eliminate your opponent’s
pawns on b7 and a6 come under serious bishop pair by swapping one of them
pressure. off. This is the only move not to lose.
Without the inclusion of 16.Bd6 Re8,
16.Bd6 this wouldn’t be available, and Black
wouldn’t have a remedy to the threat of
The immediate 16.Qb3! is a bit more
Bc7, picking up the knight on a5. That’s
accurate and forces 16...Ra7. The
why 16.Qb3! couldn’t be answered with
downside of the game move is
16...Na5. However, one could argue that
highlighted by Black’s 18th move.
it wasn’t necessary to spot this tactical
16...Re8 17.Qb3 detail in order to understand that ...Rf8-
e8 is more useful for Black than Bh2-d6
Attacking b7. is for White. The black rook, anyway,
has to move from f8, as the bishop on g7
17...Na5 is terrible and needs to be rerouted to the
a3-f8 diagonal. You could even claim
17...Ra7 is a not-so-tempting option.
that, if this is the case, then Black
shouldn’t fianchetto the bishop in the
18.Qb4
first place, and, well, you’d be right.
I’ve already said that 7...g6?! isn’t the
8
t+*Wt+l+ most logical approach.
7
+o+m+oVo 19.Bxf8 Rxf8 20.Nc5 Nxc5 21.Rxc5
6
o+*Bo+o+
5
M*+o+*+* 8
t+*W*Tl+
4 nQ*P*+*+ 7
+o+*+o+o
3 +*+bP*+p 6
o+*+o+o+
2 pP*+*Pp+ 5
M*Ro+*+*
1 +*R*+rK* 4 *Q*P*+*+
a b c d e f g h
3 +*+bP*+p
2 pP*+*Pp+
Black to move
1 +*+*+rK*
a b c d e f g h
18...Bf8!
 Black to move where Black lost the game.
The correct solution was 22...Nc6!,
21...Qd6!
but after 23.Qxb7 Black needs to find
Your new refrigerator is on the fritz, in more strong moves to survive.
a state of broken/not broken, but it is
under warranty. When you call for the
8
t+*+*Tl+
repair person to come over, the company 7
+q+*+o+o
agrees but adds: “We can’t guarantee
either the date or the time.” This is what
6
o+mWo+o+
it feels like when we defend in that
5
+*Ro+*+*
unpleasant twilight zone of OK/not OK, 4 *+*P*+*+
as Black does here. Black is defending 3 P*+bP*+p
well and has decent chances to hold the
game. The knight, which hangs two
2 *P*+*Pp+
different ways, doesn’t really hang, and
1 +*+*+rK*
Black gets to develop the queen, which a b c d e f g h
previously seemed impossible.

22.a3!
Black to move
White protects the queen, dealing
The correct defence runs: 23...Ne5!
effectively with ...b7-b6 and threatening
24.Be2 Rfb8! 25.Qc7 Qxc7 26.Rxc7
to take on a5.
Rxb2 27.Re1 Nc4!. Forced moves can
The exchange sacrifice involving
still be strong ones. Principle: Rook
22.Qxa5? b6 is possible, but completely
endings can be drawish, even when one
unnecessary.
side is down a pawn, as long as the
22...b6?! material-down side’s rook or rooks are
active. After 28.Bxc4 dxc4 29.Rxc4 Ra2
One of the only two possible moves, 30.Ra4 h5! 31.Rc1 Rb8 32.Rxa6 Rbb2
but unfortunately a mistake. With one 33.Rf1 Rb3 Black will reach a 5 versus 4
move, Black creates three weaknesses single-rook endgame, which offers
on a6, b6, and c6. When pressed, we decent practical chances to hold.
must hang in there, since there is no
other place to hang. The engine would 23.Rc3!
still hold the resulting position, but in
practical terms, this may be the spot
 24.Rfc1 Kg7 25.Qxd6 Rxd6 26.b4 Nb7
8
t+*+*Tl+
+*+*+o+o
7 Richard Reti warned us never to
fianchetto our knight! Black would be in
6
oO*Wo+o+ great shape if he had time for ...b5,
5
M*+o+*+* ...Rd7, ...Nd6, and ...Nc4, but, of course,
4 *Q*P*+*+ that’s not happening any time soon.
3 P*RbP*+p 26...Nc4? 27.Bxc4 dxc4 28.Rxc4 is
completely lost for Black, since in
2 *P*+*Pp+ addition to the extra pawn, White has
1 +*+*+rK* absolute control over the only open file.
a b c d e f g h
27.Rc7 Nd8 28.a4

Black to move 8
t+*M*+*+
Now that the black b-pawn is on b6
7
+*R*+oLo
instead of b7, Black cannot clog the c-
6
oO*To+o+
file by placing a knight on c6, which 5
+*+o+*+*
means that White will gain complete 4 pP*P*+*+
control of the only open file. In general,
that is one of the most significant types
3 +*+bP*+p
of positional advantage anyone could
2 *+*+*Pp+
ever hope for. 1 +*R*+*K*
a b c d e f g h
23...Rfd8

After 23...Qxb4?! 24.axb4 Nb7

25.Rc6 Black is also in trouble. Black to move
The engine suggests 23...Nb7!
24.Qxd6 Nxd6 25.Rc6 Nc8, which Let’s assess:
looks awful, but it is at least hard to 1) The bad news for White is that
break down, and Black can hope to hold Black’s position is quite solid, as the f7-
after slowly improving his solid position pawn is safely guarded by Black’s
with ...Rd8 and ...Kg8-f8-e7. Still, knight and king, and the a6- and b6-
White’s advantage is indisputable. pawns cannot be attacked more times
than they are defended.
 2) Black’s position is super passive, 29.b5!
with a sad, immobile knight versus
White’s powerful bishop. Black’s rooks There are two effects from this move:
are also pretty much frozen in place. 1) The b6-pawn is a fixed target.
Only the black king can move. 2) Black is forever deprived of either
...Rc6 or ...Nc6 should one white rook
28...a5?! move off the c-file, as any slim chances
of finding counterplay through these
Black wants to get his a6-pawn away
resources are eliminated.
from its current light square, where it
serves as a target for the d3-bishop. 29...Kf8
Forcing White to put more pawns on the
color of his bishop is usually a good
idea, isn’t it? In this case, it’s a definite 8
t+*M*L*+
no. The most important ingredient in a
successful defensive effort is the ability
7
+*R*+o+o
to create counterplay, and after this,
6
*O*To+o+
Black is doomed to passivity.
5
Op+o+*+*
28...Kf8?! can be met with 29.b5!? 4 p+*P*+*+
when denying the exchange with ...a6-a5 3 +*+bP*+p
*+*+*Pp+
would lead to the game structure, and 2
29...axb5? 30.Bxb5 is even worse, as the
bishop from b5 has the ability to go to
1 +*R*+*K*
e8, meaning that the final piece Black a b c d e f g h
could move, the king, is also frozen.
The only way to create counterplay
was 28...b5!!, fixing a target on the face White to move
of the pawn on b4. White’s initiative has
a harder time gaining traction after this OK, Black is doomed to passivity, but
trick. The b-pawn isn’t really hanging how is White going to break through? Is
since Black will regain it by piling up the weakness of the b6-pawn enough?
rooks on the b-file. White needs to play No, it’s not.
29.a5!, but then Black has better chances
to save the game, since if either of 30.g4!
White’s rooks move off the c-file, Black
White begins the kingside expansion
can play either ...Rc6 or ...Nc6.
project. He needs to create a
confrontation that leads to a second
weakness on the kingside.
8
*T*M*+*+
+*R*+o+*
30...h6 31.h4 g5 32.h5! 7
The h6-pawn becomes a fixed 6
*O*ToL*O
weakness on the kingside. 5
Op+o+*Op
32...Kg7
4 p+*P*Pp+
3 +*RbPk+*
8
t+*M*+*+
2 *+*+*+*+
7
+*R*+oL*
1 +*+*+*+*
a b c d e f g h
6
*O*To+*O
5
Op+o+*Op
4 p+*P*+p+ White to move
3 +*+bP*+*
2 *+*+*P*+ 37.Ra7!
1 +*R*+*K* With the scary intention to double
a b c d e f g h rooks on the seventh rank. White needs
to trade a pair of rooks to make
progress.

White to move 37...Rb7

33.f4! 37...Kg7 38.Rcc7 is also hopeless for

Black, who faces the newly created
Creating as much confrontation on the
threat of Bg6.
kingside as possible. Black’s army,
stuck on the queenside, is strikingly 38.Rxb7 Nxb7 39.Rc7 Nd8 40.Bc2!
unable to participate in the fight for the Kg7
kingside.
40...gxf4 makes the win easier for
33...Kf6 34.Kf2 Rb8 35.Kf3 Ra8 White, as after 41.exf4 Black’s king
must move to g7 all the same, and White
35...Rb7 36.Rxb7 Nxb7 37.Rc7
just gained the possibility to create a
transposes to the game.
passed pawn and simultaneously create
36.R1c3 Rb8 a path for his king to infiltrate the black
camp: 41...Kg7 42.g5 hxg5 43.fxg5 exact opposite of the truth when it
Kg8 44.Kf4! (threatening Ke5, trapping comes to zugzwang. Now, Black’s king
Black’s rook) 44...e5+ 45.Kxe5 Re6+ is forced to back down to g7.
46.Kxd5 with an easy win for White.
42...Kg7
41.Bb3 Kf6
Step 2: A pawn breakthrough.

8
*+*M*+*+ 43.e4!
7
+*R*+o+* White threatens to trap Black’s rook
6
*O*ToL*O with e4-e5, so exchanging is forced.
5
Op+o+*Op 43...dxe4+ 44.Kxe4
4 p+*P*Pp+
3 +b+*Pk+* Do you see Black’s problem? There is

*+*+*+*+
2 no remedy to the coming Ke5. That’s
why GM Ipatov, despite being 15 years
1 +*+*+*+* old when this game was played, had the
a b c d e f g h prophetic idea to put his bishop on the
a2-g8 diagonal before playing e3-e4 –
the d5-square is under control.
White to move 44...gxf4 45.Ke5

Exercise: Come up with a plan for That’s it. Black must resign since there
White. is no place for the rook to go.

Answer: Step 1: Black is in a semi- 45...Nb7

zugzwang, and we need the king to
Hey, I distinctly remember saying,
move to g7, so we waste a tempo by
“Black must resign!” Our move choices
shuffling the bishop to a2.
grow incredulous when we are out of
42.Ba2! reasonable options. Of course, this gives
away a full piece.
To refuse to decide upon a decision is, After 45...Rd5+ 46.Bxd5 exd5 White
in a weird way, a decision. We are wins, so long as he doesn’t fall for the
programmed to believe that waiting is knight fork cheapo on e6.
synonymous with failure, which is the
46.Rxb7 Rd8 47.Bxe6 1-0
8
tMv+lV*T
+o+*+oO*
OK, that’s enough. 7
The single weakness on b6 wasn’t
enough for White to win. He only made
6
oW*OoM*O
progress when he opened a second front 5
+*+*+*+*
on the kingside with 30.g4!, which 4 *+*NpP*B
eventually led to the creation of a
weakness on the other side of the board
3 +*N*+*+*
in the shape of the pawn on h6. In the
2 pPp+*+pP
end, Black’s defense was stretched to 1 R*+qKb+r
breaking point, and the previously a b c d e f g h
passive pieces on d6 and d8 were the
ones that paid the price.

White to move
The following game doesn’t neatly fall
into the sphere of a single chapter, since MVL’s specialty used to be this
it involves holes, bad minor pieces, a delayed version, with ...h7-h6 and Bg5-
weakness in a color complex and, h4 tossed in, until 2018. Then, he
finally, the most important factor, the migrated to the immediate 7...Qb6.
principle of multiple pawn weaknesses.
Is it better to play many different
Game 3 openings and be a Jack of all trades,
Sergey Karjakin – Maxime who is difficult for opponents to prepare
Vachier-Lagrave for? Or should we cultivate a
London 2017 narrow/predictable opening repertoire,
which we know incredibly deeply, yet
1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 our opponents will expect?
Nf6 5.Nc3 a6 6.Bg5 e6 7.f4 h6 8.Bh4
I don’t know the answer to this
Qb6!?
question. I have personally always
switched between different opening
lines to keep my opponents guessing.
MVL is the poster child of the other
method. He is highly predictable in his
opening choices, but understands the
resulting positions to an immense depth.
To me, it’s a scary thought that my
opponent, a mortal, would have pre- of fighting for the g1-a7 diagonal is the
existing knowledge of the position that main reason the approach with a2-a3 is
would arise on the board. To imagine more popular than sacrificing the b2-
giving that advantage to the likes of pawn when ...h7-h6 and Bg5-h4 are
Magnus Carlsen and Fabiano Caruana is included.
madness. It takes immense confidence to
take the MVL opening preparation 10...Qc7 11.Qf3 Nbd7 12.0-0-0 b5
route, possibly even more so at the 13.g4
highest level.

9.a3
8
t+v+l+*T
7
+*WmVoO*
o+*OoM*O
It shows the level of respect MVL’s 6
opponents show him when they back off
from gambiting their b-pawn. In this
5
+o+*+*+*
version, with ...h7-h6 and Bg5-h4 4 *+*NpPp+
included, this is actually the most
popular move, and has been tried by the
3 P*N*+q+*
most respected of opening theoreticians.
2 *Pp+*B*P
Is the old, sacrificial model bankrupt?
1 +*Kr+b+r
Actually, no. Fully playable is the high- a b c d e f g h
risk gamble 9.Qd2!?, allowing 9...Qxb2.
It is a theoretical hornet’s nest, and
nobody knows the lines better than
Black to move
MVL. Enter such a line only if you are
willing to cram a massive quantity of Black to move
data into your overloaded, bewildered
brain. We have a typical Najdorf position
before us, except that White’s bishop is
9...Be7 on f2, rather than its normal square on
e3. Does this make a difference?
Players rated in the 2800 range don’t Actually, it does, since it makes it easier
fall for 9...Qxb2??, which gets Black’s for Black to play ...g7-g5, which results
queen trapped after 10.Na4. in Black seizing control over the central
and kingside dark squares, and creates
10.Bf2
an outpost on e5.
This threatens Nxe6. The possibility
13...Bb7!
 It’s too soon for the break on g5. MVL Nxd5 16.Nf5! when we see the idea
beat Giri with the premature 13...g5?!, behind posting the bishop on g2. White
although Giri got the better of it in the stands better.
opening after 14.h4! gxf4 15.Be2!. This
had already been played in a few 8
t+*+l+*T
correspondence games. White’s idea is
that g4-g5 will send the f6-knight back
7
+vWmVo+*
to g8, after which Black is in danger.
6
o+*OoM*O
15...Rg8 was Giri-Vachier-Lagrave, 5
+o+*+*O*
Stavanger 2016. According to MVL, 4 *+*NpPp+
White could have gotten a powerful
attack by playing 16.g5! hxg5 17.hxg5.
3 P*N*+q+*
Black’s position is in deep trouble,
2 *Pp+*BbP
whether he accepts the g5-pawn or not. 1 +*Kr+*+r
a b c d e f g h
14.Bg2

This is a more logical square for the

bishop than d3, since White wants to White to move
suppress the ...d6-d5 break.
14.h4!? is the move to play if White 15.h4!
wants to discourage Black from playing
White does not back down from the
...g7-g5. MVL faced this move five
fight for the kingside.
times against strong opponents, and all
five games were drawn. Black’s main 15.f5?! e5! is OK for Black. With the
reaction is centered on the weakness of kingside locked, the king on e8 feels
the e4-square: 14...Nc5 15.Bd3 d5 16.e5 safer than ever. Of course, the worst
Nfe4, with a complete mess in which thing White can do from a positional
Black appears to be fine. standpoint is to take on g5: after
15.fxg5? hxg5 the h8-rook is activated, a
14...g5! knight is coming to e5, and the g4-pawn
is weak. Black has the better structure,
White focused on controlling the light the more active pieces, and the safer
squares, so Black implements his other king. It’s a perfect Najdorf.
thematic break, trying to carve
weaknesses on the central dark squares 15...gxf4 16.g5!
instead.
Principle: Open the game and create
14...d5?! can be met with 15.exd5
confrontation when leading in would take Black, for the following
development. reasons:
1) Black has smoothly completed
16...Ne5 development and his king is safe on the
queenside.
This way d7 is cleared for the f6-
knight. 2) Black holds a big outpost on e5,
currently occupied by a knight.
17.Qxf4 hxg5! 3) White’s e4- and g5-pawns could
later turn into targets, especially if the
The text move is an improvement over queens are removed from the board.
the previously played 17...Nfg4?!,
4) White has nothing to attack in
which allows 18.g6! with an advantage
Black’s position.
for White.
Conclusion: As in most cases, we
18.hxg5 Rxh1 19.Rxh1 Nfd7 20.Kb1 cannot question the objective merits of
the engine evaluation. However, we can
20.Rh8+ doesn’t bother Black after question its practical relevance. In this
20...Nf8. case, sure, maybe an engine finds it
“easier” to play with White (what does
20...0-0-0 ‘easy’ really mean for an engine?), but
for humans, it is much easier to play
Black, since White’s position is full of
8
*+lT*+*+ weaknesses while Black is both active
7
+vWmVo+* and solid.
6
o+*Oo+*+ 21.Rh3
5
+o+*M*P*
4 *+*NpQ*+ Maybe Karjakin is looking for tricks
3 P*N*+*+* involving a future Rc3.
2 *Pp+*Bb+ 21.g6!? leads to crazy complications:
21...Rg8 22.gxf7 Rxg2 23.Nxe6 Qc4
1 +k+*+*+r 24.Rh8+ Nf8 25.Nxf8 Qf1+ 26.Nd1!
a b c d e f g h Qxd1+. The engine works this out to a
forced draw.

21...Kb8 22.Be3 Rg8 23.Rg3 Rg7

White to move
Black reinforces his only real
The engine slightly prefers White. I
weakness on f7, trying to create options
for the knight on e5. 8
*Lw+*+*+
7
+v+*Vo+t
8
*L*+*+*+ 6
o+*Oo+*+
7
+vWmVoT* 5
+oM*M*P*
6
o+*Oo+*+ 4 *+*Np+*+
5
+o+*M*P* 3 P*N*B*R*
4 *+*NpQ*+ 2 *Pp+*Qb+
3 P*N*B*R* 1 +k+*+*+*
2 *Pp+*+b+ a b c d e f g h
1 +k+*+*+*
a b c d e f g h
White to move

The engine continues to slightly prefer

White to move
White, but I don’t believe in this
24.Bh3!? assessment for the following reasons:
1) With his last move, Black enables
Karjakin is playing for undermining his queen to move either to g8,
tricks like g5-g6, as well as ideas like increasing the pressure on g5, or h8,
Bxe6. seizing full control over the h-file.
2) Simultaneously, the queen’s retreat
24...Rh7!
cleared the path for the dark-squared
This way Black evades all Bxe6 ideas, bishop to reroute to the more active b6-
since there will be no rook on g7 for square.
White’s knight to fork. 3) White’s e4- and g5-pawns remain
targets.
25.Qf2 Nc5 4) Black’s e5-knight continues to
occupy the hole on e5.
Reminding White that he must worry
about the defense of the e4-pawn. 5) Black’s knight on c5 is, in effect,
occupying another hole, since White
26.Bg2 Qc8! can’t afford the weakening to his
position by ejecting the knight with b2-
b4.
 6) The option of ...Ne5-c4 looms over
White unpleasantly. And if White tosses 8
*Lw+*+*+
in b2-b3, then the c3-knight is somewhat 7
+v+*+o+t
oV*Oo+*+
undermined and the white king’s cover 6
becomes a bit loose.
5
+oM*M*P*
27.Bc1 Bd8! 4 *+*Np+*+
The b6-square beckons for the bishop.
3 P*N*B*R*
28.Nf3 Ng6!
2 *Pp+*Qb+
1 +k+*+*+*
On g6, blockading the g5-pawn, the a b c d e f g h
black knight is still better than its white
counterpart.

29.Nd4 Ne5 30.Nf3 Ng6 31.Nd4 White to move

A silent draw offer. The knight is aiming to jump to c4

next, harassing the e3-bishop.
31...Bb6!
33.b3
No draw! MVL rightly believes that
he should play on, despite the engine This is a sign of degeneration in
now assessing the position as equal. White’s position, but anything else
would have allowed the knight to access
32.Be3 Ne5 c4.
33.Nf3?! Nc4 is better for Black, who
is about to eliminate the all-important
white dark-squared bishop. More
prudent is 33.Nb3, but after something
like 33...Qc7 34.Rh3 Rxh3 35.Bxh3
Ncd7 Black’s position remains the
easier of the two.

33...Rh4!

The pressure on e4 increases.

34.Nde2 Qc7 35.Bd4!

 desperate. This swap only helps Black
8
*L*+*+*+ reach a superior endgame.
7
+vW*+o+* White was better off complicating
6
oV*Oo+*+ matters with a pawn sacrifice: 36.g6!?
fxg6 37.Bh3!. At least now the white
5
+oM*M*P* pieces have some targets to work with.
4 *+*Bp+*T
3 PpN*+*R* 36...Qxb6 37.Qxb6
2 *+p+nQb+ White doesn’t want an endgame, but
1 +k+*+*+* backing off from the trade would cede a
lot of control over the central dark
a b c d e f g h
squares.

37...Nxb6 38.Nd4
Black to move

35...Ncd7! 8
*L*+*+*+
This coming swap of dark-squared
7
+v+*+o+*
bishops will leave White with a bad
6
oM*Oo+*+
remaining bishop. 5
+o+*M*P*
35...Nxe4? is tempting yet incorrect: 4 *+*Np+*T
36.Bxe4 Bxd4 37.Nxd4 Bxe4 38.Nxe6!
and White stands better, since
3 PpN*+*R*
38...fxe6?? is met with 39.Nxe4 Rxe4
2 *+p+*+b+
40.g6! when White’s passed g-pawn is 1 +k+*+*+*
actually worth more than Black’s extra a b c d e f g h
piece. If Black returns the piece, his
king becomes too exposed to survive.
The engine calls this position dead lost
Black to move
for Black.
38...Ng6!
36.Bxb6?!
White is held hostage to his
The last thing we want when on the
opponent’s sinister agenda:
defensive is the opponent stabilizing,
since chaos is the last refuge of the 1) White’s bishop is not a good one,
since it is hemmed in by its own pawn
on e4. White’s bad bishop is the mark of a
2) White’s e4-pawn is weak. flexible mind. Yes, White’s bishop was
3) White’s rook is tied down to indeed a bad one. On the other hand, the
babysitting duty for the g5-pawn. bishop protected White’s weak pawn on
Black’s inactive b6-knight will head for e4, and Black only had one superb
the hole on e5 via d7, when both black outpost for his knights, on e5. With the
stallions will be placed excellently, on bishop off the board, e4 becomes
the squares in front of isolated enemy difficult for White to hold.
pawns.
42.Rxg2 Rh3+! 43.Kd2
4) White’s c3-knight is passively
placed. White’s king must give ground.
5) Black has seized control over f4 and 43.Nf3?? isn’t possible due to 43...Ne5
now White must watch out for both 44.Rf2 Ng4+.
...Rf4 and ...Nf4 ideas.
43...Ne5
39.Kc1!

Principle: Use your king as a 8

*L*+*+*+
functional piece in the ending, rather
than hide him away in a corner.
7
+v+*+o+*
6
o+*Oo+*+
39...Nd7 40.Kd2 Nf4! 5
+o+*M*P*
Threatening ...Nxg2, followed by
4 *+*Np+*+
...Bxe4.
3 PpN*+*+t
41.Ke3
2 *+pK*+r+
1 +*+*+*+*
41.Bf3 is met with 41...Ne5. a b c d e f g h

41...Nxg2+!?

Some people, being mean to our White to move

game’s funkier piece, claim that the best
knight is still worse than the most 44.Nde2
terrible bishop. Taking on g2 is
White is running out of useful moves
unnecessary, though it isn’t bad.
and, for example, 44.Rf2? loses to
Swapping the powerful knight for
44...Rg3 when the g5-pawn falls.
Instead, 44.Rg1?! is possible, but
inadvisable. White is completely 8
*+*+*+*+
paralyzed – nothing can move. Black 7
+vL*+o+*
o+*Oo+m+
will keep improving his position, first by 6
centralizing his king.
The way for White to hold was 44.g6!,
5
+o+*+*P*
undermining the defense of the e6-pawn. 4 *+*+p+*+
As usual, going for counterplay is what 3 PpN*+*+t
*+pKn+*+
saves the day, not passive defense. 2
However, that’s not all there is to it.
After 44...fxg6 (44...Nxg6 45.Nxe6 is
1 +*+*+*R*
equal) White needs to find 45.Rf2!!, a b c d e f g h
which is probably the move Karjakin
missed. White will either capture the
pawn on e6 next or invade with the rook
White to move
to f6, getting just about enough
counterplay to maintain the balance. The Black’s king heads for e7, from where
point is that the hasty 45.Nxe6?? loses he protects Black’s only weakness, the
to 45...Nf3+ followed by ...Nh4, with a f7-pawn.
double threat on the rook on g2 and the
knight on c3. That’s why 45.Rf2!! is an 48.Nd4 Kd7 49.Rf1 Ke7 50.Rg1 Ne5
essential inclusion before taking on e6. 51.Rg2 Ba8 52.Nde2 Nf3+! 53.Kc1
Rh1+!
44...Nf3+! 45.Kc1
This alert insertion pushes White’s
White’s king is unable to move to the king further away from the center.
third rank, since that would blunder
material to a discovered check on e1. 54.Kb2 Nh4! 55.Rg4 Ng6

45...Nh4 46.Rg1

46.Rf2 is met with 46...Rf3.

46...Ng6 47.Kd2 Kc7

 59...d5 0-1
8
v+*+*+*+
+*+*Lo+*
7 White realizes that it is expensive to
be poor. Black will take back on d5 with
6
o+*Oo+m+ the bishop and win easily with his two
5
+o+*+*P* connected passed pawns on the e- and f-
4 *+*+p+r+ files. The technical task would not be

PpN*+*+*
3 hard for a player of MVL’s caliber, so
Karjakin decided to call it a day.
2 *Kp+n+*+ Throughout this game, MVL’s
1 +*+*+*+t technique was hypnotic.
a b c d e f g h When the world’s current best player
is in form, he becomes an unstoppable
force. In the Meltwater Tour Final of
2021, Carlsen secured clear first place
White to move
with two rounds to spare! And if
56.Nf4 Magnus Carlsen taught the chess world
one thing, it’s that simplified positions
This loses a pawn, yet the engine says don’t necessarily deplete a position of its
it’s White’s best move! creativity, subtlety, or hidden ideas.
After 56.Ng3 Rh2 57.Nf1 Rf2 the Especially astounding was the richness
black pieces are totally dominating their of his endgame ideas in the following
white counterparts. Alternatively, game, particularly when we consider
56.Rg3 loses a pawn to 56...Rh4 57.Re3 that each player started the game with
Ne5 when it’s impossible to keep the 15 minutes and a 15-second increment.
pawn on g5: for example, 58.Kc1 Rg4 Game 4
59.Rg3 Rxg3 60.Nxg3 Nf3 and the Magnus Carlsen – Hikaru
pawn drops. Nakamura
Internet (rapid) 2021
56...Nxf4 57.Rxf4 Rg1 58.Rh4 Rxg5
59.Kc1

White’s last chance was 59.Rh8 Bc6

60.Rc8 Kd7 61.Rf8 f5 when the
endgame is still probably lost for White,
but at least Black doesn’t get two
connected passed pawns.
 8
*+l+*+*T 8
*+l+*+*T
7
OoOt+oO* 7
OoOt+*O*
6
*+w+*M*O 6
*+*+oM*O
5
+*+*+*+* 5
+*+*+*+*
4 *+*+p+*B 4 *+*+p+*+
3 +q+*+*+p 3 +*+*+*Bp
2 pP*+*Pp+ 2 pP*+*Pp+
1 +*R*R*K* 1 +*R*R*K*
a b c d e f g h a b c d e f g h

Black to move Black to move

19...Qe6?! Commentators Ian Nepomniachtchi

and Peter Leko both correctly forecasted
I haven’t decided if it is a sign of that such an endless ‘+=’ ending is
bravery or madness to agree to enter an suicide against Carlsen’s machine-like
inferior endgame against Magnus technical skills. White exerts nagging
Carlsen, who may be the most dominant pressure for the following reasons:
technical endgame player in chess 1) Black has three pawn islands
history. against White’s two. Remember, fewer
19...Qb6! is a better way to offer a is better when it comes to pawn islands
queen swap. In this case White should since each island has at least one base
dodge the swap and try to attack Black’s pawn, and base pawns require
king. protection.

20.Qxe6!? 2) Black has an isolated e-pawn,

which could later become a serious
The engine says this is not optimal, weakness.
calling for 20.Qa4, yet the top human 3) White’s bishop is clearly superior to
disagrees. Black’s knight, since a bishop’s power
tends to be enhanced in endings with
20...fxe6 21.Bg3 opposing-wing pawn majorities. The
reason is that the long-range bishop can
simultaneously attack and defend, while 23.Be5!
the short-range knight can only do one
or the other, but not both. Preventing ...e6-e5. If the pawn were
4) White’s pawn majority looks faster. allowed to e5, then Black would
Pushing Black’s 3-2 queenside majority stabilize with ...Nh5-f4, solving almost
looks difficult – if not outright all of his problems, since the e8-rook
impossible – since White owns the open would be freed from its defensive duties
c-file, hampering a push to c5. and ready to play an active role. On e5,
White’s bishop annoys Black on both
21...b6 sides of the board and covers the b2-
pawn, in case Black later tries to
Nakamura wants to activate his infiltrate the second rank. Moreover, e5
majority with ...Kb7 and ...c7-c5. is hard for Black to reach, and the
bishop is surprisingly stable.
22.Rc6
23...Kb7 24.Rec1 Ree7
Stopping ...c7-c5, attacking the e6-
pawn, and preparing to double rooks on 24...Nxe4?? is a ridiculous decision,
the c-file. since it allows White to infiltrate
Black’s seventh rank with 25.Rxc7+
22...Re8 Rxc7 26.Rxc7+ Ka6 27.Rxg7. White has
won a pawn while keeping most of his

*+l+t+*+
8 positional trumps.
7
O*Ot+*O* 25.f3
6
*Or+oM*O Securing the e4-pawn.
5
+*+*+*+*
4 *+*+p+*+
3 +*+*+*Bp
2 pP*+*Pp+
1 +*+*R*K*
a b c d e f g h

White to move
 Allowing the white pawn to h5 didn’t
8
*+*+*+*+ look appealing, but in this case, the cure
7
OlOtT*O* is much worse than the disease.
6
*Or+oM*O 26...a5!?, stopping b2-b4, offered
Black better chances to save the game.
5
+*+*B*+* A less refined approach with 26...Nd6!?
4 *+*+p+*+ could also work, when Black is also
3 +*+*+p+p sneakily threatening ...Nf7 followed by

pP*+*+p+
2 ...e6-e5. That’s because after something
normal like 27.b4 Nf7! White shouldn’t
1 +*R*+*K* play 28.Bxg7??, because it runs straight
a b c d e f g h into 28...Nd8 and Black wins.

8
*+*+m+*+
Black to move 7
OlOtT*O*
25...Ne8!?
6
*Or+o+*+
5
+*+*B*+o
Nakamura goes into Maginot Line 4 *+*+p+*P
+*+*+p+*
mode, hunkering down. If any piece in 3
Black’s camp must be passive, it’s best
that the knight is the one, as that allows
2 pP*+*+p+
the stronger pieces to be activated. 1 +*R*+*K*
25...a5! is a better move order, a b c d e f g h
stopping b2-b4, but the knight should
retreat to e8 on the next move.

26.h4!? White to move

Intending to play h4-h5, gaining space Exercise: Nakamura’s last move

and fixing the weakness on g7. created a hole on g5 and weakened
Carlsen’s move is logical, but one the kingside dark squares in
could argue that his plan will always be general. Come up with a plan for
available, while b2-b4 can be stopped, White to exploit these newly
so it might be better to start with 26.b4!. created weaknesses.

26...h5? Answer: The g5-square beckons the

white king.
8
*+*+m+*+
OlO*+t+*
27.Kh2! 7
Principle: The king is a fighting piece 6
*O*+o+o+
in the endgame and should be used as 5
+*+*B*+o
*+*+p+*P
such. In fact, when acting as a fighting 4
unit, the king is often considered to be
worth four points, which is more than a
3 +*+*+pK*
minor piece. 2 pPr+*+p+
27...g6?!
1 +*+*+*+*
a b c d e f g h
There isn’t a good reason to further
weaken the kingside dark squares.
Black should have tried 27...Nd6!, White to move
with the same idea as in the previous
move. 31.b4!

28.Kg3! Rd2 Endgame Principle: When possible,

stifle and slow down the opponent’s
The obvious idea is to discourage Kf4, pawn majority. White’s intention is to
as it would leave g2 hanging, but what prevent ...c7-c5.
Black really wants is to ease the
pressure slightly by forcing White to 31...a6 32.a4!
swap a pair of rooks.
Black can’t stop the Kf4 idea with Intending to push forward to b5 to
28...Rf7?? due to 29.Rxe6, winning a secure the c7-pawn.
pawn.
32...Rf8 33.b5
29.R6c2 Rxc2 30.Rxc2 Rf7!
The consistent, human approach.
It’s critical for Black to prevent the Carlsen fixes the c7-pawn and creates an
white king from reaching f4 and g5. outpost for his rook on c6.
Indeed, 30...c5? 31.Kf4 is hopeless for
33...axb5 34.axb5
Black.
Not everything is rosy for White on
the queenside; Carlsen must always
watch out for ...Nd6, attacking the b5- ineffective excursion the rook is about to
pawn. embark on.
White’s most precise route to the goal
34...Kc8 isn’t trivial, but it involves keeping the
black pieces dominated for as long as
The king goes to d7 to safeguard the
possible: 35.Rc6! Kd7 36.Kf2! (Black
pawn on e6. 34...Nd6?? fails to
cannot escape the bind, which gives
35.Bxd6! cxd6 36.Rc6 Rd8 37.Kf4
White infinite time to improve his
when the upcoming infiltration with the
pieces; thus, the king is heading towards
king on the kingside will pick up all of
d4) 36...Rf7 37.Ke3 Rf8 38.Bf4! Rf7.
Black’s pawns. The only hope is 37...d5,
trying to create counterplay with the
passed d-pawn, but White can still pick
8
*+*+m+*+
up the pawn and win: 38.Rxe6! d4 7
+*Ol+t+*
39.Rc6 d3 40.Rc1 d2 41.Rd1. Too slow 6
*Or+o+o+
+p+*+*+o
buddy. Black’s pawn is stopped and 5
White wins.
4 *+*+pB*P
8
*+l+mT*+ 3 +*+*Kp+*
7
+*O*+*+* 2 *+*+*+p+
6
*O*+o+o+ 1 +*+*+*+*
5
+p+*B*+o a b c d e f g h
4 *+*+p+*P
3 +*+*+pK*
2 *+r+*+p+ White to move
1 +*+*+*+* Here 39.e5!! is a necessary change in
a b c d e f g h structure. It may seem like this
surrenders some light squares, but it
doesn’t – White can control the f5-
square with g2-g4, keeping the e8-
White to move knight trapped on the e8-g7 circuit.
Pushing the e5-pawn controls the d6-
35.Ra2?!
square, forever preventing ...Nd6, and it
This is not a mistake, as it doesn’t ruin also frees the e4-square for White’s
anything, but it’s the prelude to an king. After 39...Rf8 40.Bg5 Rf7 41.g4
Rf8 42.Ke4 Rf7 now that the white king,
bishop, and pawns are perfectly placed, 8
*+*+mT*+
it is time for the strongest piece to 7
+*Ol+*+*
*O*+o+o+
deliver the final blow: 43.Rc1!. The rook 6
will invade the black camp with Rd1+
followed by Rd8, and Black has nothing
5
+p+*B*+o
better than surrendering. As usual, there 4 *+*+p+*P
was a way to break through the fortress.
Principle: Rarely trust passive defense!
3 +*+*+p+*
2 *+*+*Kp+
35...Kd7 36.Ra8 Rf7 37.Ra1! 1 R*+*+*+*
a b c d e f g h
Nice catch. Carlsen admits his mistake
and goes back, preventing ...Nd6.

37...Rf8! Black to move

37...Nd6? 38.Rd1! followed by taking 38...Nd6!
on d6 leads to a winning rook endgame
for White: for example, 38...Rf8 Nakamura is one of the most tactically
39.Bxd6! cxd6 40.e5! d5 41.Ra1! and alert players of our era, and rarely
Black cannot hold on to all of his misses a chance to generate counterplay
weaknesses (b6, e6 and g6). when offered an opportunity. Black
simultaneously threatens both ...Nxb5
38.Kf2? and ...Nxe4+, exploiting the pin on the
f3-pawn.
In rapid games, brilliant ideas have a
way of being punctuated with mistakes, 39.Bxd6
since the players don’t have the luxury
of time to work out the details. In this case, 39.Rd1 leads nowhere, as
Correct was 38.Rd1+! Ke7 39.Rc1 Black is ready for 39...Ra8!, activating
Kd7 40.Rc6!, transposing to the note the rook and seizing the important a-file.
above on 35.Rc6!.
39...Kxd6

White is still slightly for choice, but

Black, having an active king that could
potentially harass the b5-pawn, should
be able to hold a comfortable draw.
 Of course, 39...cxd6?? 40.Ra7+ Ke8
41.Rb7 is dead lost for Black. 8
*+*+*T*+
40.Rd1+ Ke7?
7
+*O*L*+*
6
*O*+o+o+
This holds, but it is wrong in principle, 5
+p+*+*+o
*+*+p+*P
and the real reason behind Nakamura’s 4
loss.
Principle: When we have more
3 +*+*Kp+*
weaknesses, we should be willing to 2 *+*+*+p+
change the course of the game. Dynamic
tactical skirmishes are often better than
1 +*+r+*+*
a b c d e f g h
slow technical maneuvers. Even though
it requires some good nerves, Black
should go for 40...Kc5! and try for an
active draw. As I’ve already said, never Black to move
trust passive defense. You might find
this line a bit nervy, but it is quite easy 41...Ra8?
compared to trying to hold passively:
White’s king should not be given entry
41.Rc1+ Kxb5 42.Rxc7. At this point,
to the f4-square, as that square holds the
Black has five different moves that lead
key to the entire black kingside.
to a draw, the most logical of which, I
Do you remember earlier when I said
think, is 42...Rb8!, listening to classical
that Black had five different drawing
wisdom and putting the rook behind the
moves? At this point, he had one, and it
passed pawn.
was much harder to spot: 41...Kf6!.
41.Ke3 Who in their right mind would allow the
rook into d7? I don’t think anyone
should judge Nakamura for not finding
this – it’s his previous move that should
be frowned upon. After 42.Rd7 Ra8!
43.Rxc7 Ra5 Black miraculously
manages to both regain the sacrificed
pawn and prevent the white king from
infiltrating the kingside.

42.Rc1?
 Ugh! Good, old rapid chess, the surest White to move
path to prematurely grey hair!
45.Rf1!
42.Kf4! wins after 42...Ra2 43.Rc1!
Kd7 44.Rg1! followed by Kg5. This is even more accurate than the
also winning idea of immediately
42...Kd7?
moving the king to g5. Hey, we were
Black still holds the draw with told to never make passive moves in a
42...Kf6! 43.Rxc7 Ra5. rook ending, and yet Carlsen does just
that! He realizes this is an exception to
43.Kf4! the endgame. Principle: Don’t post your
rook passively in rook endings. In this
Carlsen does not need to be asked case though, it is fully justified because
thrice. The king enters, and Black is it covers the base f3-pawn, which in turn
busted. allows White’s king to raid Black’s
kingside via the porous dark squares. By
43...Ra2
the way, as I warned in the other
After 43...Ra5 44.Kg5! Rxb5+ volumes of this series, don’t be
45.Kxg6 Ra5 46.g4 White is too fast in surprised if you try to disobey a
the race and wins. principle and it backfires in spectacular
fashion. That’s a luxury reserved only
44.g4 Ra3 for the deities of our game – we
commoners should usually abstain.

8
*+*+*+*+ 45...Ke7 46.Kg5 Kf7 47.Kh6!
7
+*Ol+*+*
6
*O*+o+o+
5
+p+*+*+o
4 *+*+pKpP
3 T*+*+p+*
2 *+*+*+*+
1 +*R*+*+*
a b c d e f g h
 Ke8 51.Kxg6 is curtains.
8
*+*+*+*+ 50.Rc6 Re7 51.f4! Rd7 52.f5!
7
+*O*+l+*
6
*O*+o+oK The pawn breaks through, and the
5
+p+*+*+o black defense is turned into dust.
4 *+*+p+pP 52...gxf5 53.exf5 exf5 54.Rf6+ Kg8
3 T*+*+p+* 55.Rxf5 Rd6+ 56.Kxh5
2 *+*+*+*+
1 +*+*+r+* 8
*+*+*+l+
a b c d e f g h 7
+*O*+*+*
6
*O*T*+*+
Black to move
5
+p+*+rPk
4 *+*+*+*P
Carlsen takes as much space as he 3 +*+*+*+*
possibly can. The intention is g4-g5!
followed by f3-f4 and f4-f5, creating a
2 *+*+*+*+
passed g-pawn for White. Shifting the
1 +*+*+*+*
rook to c1 is also something Black must a b c d e f g h
always keep in mind.

47...Rc3
Black to move
47...Rb3 48.Rc1 is equally hopeless for
The almost sacrosanct American belief
Black.
in the right to private property is proven
48.g5! Rd3 49.Rc1! to be a delusion when it comes to the
chess board.
Carlsen applies the principle of dual
weaknesses, switching back and forth 56...c6 57.bxc6 Rxc6 58.g6! Rc8
from g6 to c7. 59.Kg5

49...Rd7 The simple idea is to make way for the

h-pawn to run up the board.
Black’s rook is forced into a state of
abject passivity, as 49...Rxf3 50.Rxc7+ 59...Rb8 60.h5!
 Carlsen confidently calculates that 1.e4 c5 2.Nf3 Nc6 3.Bb5 g6 4.0-0 Bg7
there is no need for him to play 60.Rb5, 5.Re1 e5 6.Bxc6 dxc6
which is the cleanest way for the more
anxious among us, especially in time After such exchanges on c6 in the
pressure. Rossolimo Sicilian, the players live in
distinctive ecosystems. White inflicts a
60...b5 1-0 touch of damage to Black’s structure,
while Black picks up the bishop pair,
Nakamura played his move and which could turn into something
resigned in view of 61.h6 b4 62.h7+ significant if the game opens up later on.
Kg7 63.Rf7+ Kh8 64.Kh6, threatening
g6-g7 mate. Of course, 64...Rb6 covers 7.d3 Qe7 8.Nbd2 Nf6 9.a3
one mate, while leaving open another in
65.Rf8 mate.
What an amazing technical
8
t+v+l+*T
performance. If we had to pick only one 7
Oo+*WoVo
aspect of what we just saw, particularly 6
*+o+*Mo+
+*O*O*+*
instructive is how Carlsen tormented his 5
opponent by simultaneously threatening
the backward c7-pawn combined with
4 *+*+p+*+
threats to infiltrate with his king on the 3 P*+p+n+*
kingside. 2 *PpN*PpP
In the following game, Michael
1 R*BqR*K*
a b c d e f g h
Adams, almost imperceptibly, built his
winning position upon small gains, until
he reached a position where Black was
unable to defend his multiple Black to move
weaknesses. It is a good example of how
a strong GM, with seemingly Playing for the b2-b4 counterplay is a
unremarkable upheavals – as upheavals standard idea in such structures. But you
go – quietly improves his position may be wondering, why would White
without the opponent noticing until it’s want to undouble Black’s c-pawns?
too late. 1) First of all, Black’s c5-pawn isn’t
Game 5 so bad, since it’s not weak and it allows
Michael Adams – Nikita Meskovs Black to clamp down on the d4-square.
Germany 2022 2) Secondly, if White plays b2-b4 and
Black exchanges, then essentially White 11.h3
is exchanging off a less valuable a-pawn
for Black’s more central c5-pawn. Also, This way Black either loses a tempo or
the a-file then opens for White. hands back the bishop pair.
3) If Black doesn’t take on b4, but
11...Bxf3 12.Nxf3 Rfd8?!
tries to maintain the pawn on c5, he runs
the risk of remaining with doubled Intending ...c5-c4, but more to the
isolated pawns, which would be point is 12...a5!, fighting for the
extremely weak. Most players, when queenside squares.
thinking about the downsides of having
doubled pawns, think about how these
pawns are weaker when attacked by
8
t+*T*+l+
pieces compared to healthier structures.
7
Oo+*WoVo
However, that’s not the only way 6
*+o+*Mo+
doubled pawns are weak. They are also 5
+*O*O*+*
weaker than healthy ones when it comes
to pawn breaks, and they’re also weaker
4 *P*+p+*+
when it comes to pushing them, i.e. it’s
3 P*+p+n+p
harder to create passed pawns. 2 *+p+*Pp+
9...0-0 10.b4 Bg4?!
1 R*BqR*K*
a b c d e f g h
It’s liberating to break a law, rather
than follow it – but only if we get away
with the crime. It always mystifies me White to move
why so many players on the black side
of such structures are willing to hand 13.Bd2
over their valuable asset of the bishop
pair. With one small move, White:
Grabbing the pawn with 10...cxb4 1) Protects the b4-pawn.
11.axb4 Qxb4?? runs into 12.Ba3, 2) Breaks the pin on the d3-pawn,
winning the exchange. I think relatively avoiding ...c5-c4.
best for Black is 10...Be6!, planning 3) Stops ...a7-a5.
...Nd7. The game looks even, and the
e5-pawn is taboo, as 11.Nxe5? runs into 13...Nd7! 14.Qb1
11...Ng4, exploiting the pin on the e5-
It’s wise for the queen to leave the d-
knight.
file, as Black’s rook sits on d8.
14...b5 15.a4!? More constructive would be to reroute
the awkward black knight towards e6
With what is essentially a minority with 19...Nf8!.
attack, Adams manages to create a small
initiative on the queenside. 20.Ng5! Qf6 21.cxb5!

15...a6 16.Qb3 cxb4 17.Bxb4 Qe8 Adams cleverly tempts his opponent
with a piece sacrifice which shouldn’t be
17...c5 18.Ba5! is slightly in White’s accepted.
favor.
21...cxb5
18.axb5 axb5
21...Qxg5?! is in White’s favor after
After 18...cxb5!? 19.Ra5 Black’s a6- 22.bxc6. Black must return the extra
pawn is slightly weaker than White’s c2- piece and be down a pawn, as the greedy
pawn. 22...Nf8?? loses to 23.c7 Rdc8 24.Rxa8
Rxa8 25.Bxf8 Bxf8 26.Qb7, winning the
19.c4
black rook and the game.

8
t+*Tw+l+ 8
t+*T*+l+
7
+*+m+oVo 7
+*+m+oVo
6
*+o+*+o+ 6
*+*+*Wo+
5
+o+*O*+* 5
+o+*O*N*
4 *Bp+p+*+ 4 *B*+p+*+
3 +q+p+n+p 3 +q+p+*+p
2 *+*+*Pp+ 2 *+*+*Pp+
1 R*+*R*K* 1 R*+*R*K*
a b c d e f g h a b c d e f g h

Black to move White to move

19...Qe6?! 22.Nf3!

This turns out to be a slight inaccuracy The knight calmly returns to f3, and its
which Adams begins to milk. mission can be viewed as a complete
success. It might have lost two tempi in Black to move
the process, but the white queen has won
the battle for control of the important 25...Rxc2?
a2-g8 diagonal.
Black must make a concession and he
22...Qc6 23.Rac1!? chooses the wrong one. He would likely
have still held a draw with 25...Bh6!.
Adams prefers board control After 26.Rxc8 Rxc8 27.Rxc8+ Qxc8
compared to a measly pawn. The 28.Qd5 Bf8 29.Qxb5 Qc1+ 30.Be1
alternative was 23.Rxa8 Rxa8 24.Ng5 Nc5! Black generates enough play to
Qf6 25.Qd5 Nb6 26.Qxb5 Qxg5 hold, despite White’s extra pawn.
27.Qxb6 Qd8. Sure, White won a pawn.
Yet conversion won’t be so simple since 26.Rxc2 Nf8
all the remaining pawns are on the same
This is a case of the right idea, played
side of the board. Black has at least a
too late!
50% chance of holding the draw, if not
higher. 27.Bxf8?!
23...Qb7 24.Rc2 Rac8 This is incorrect, but we can
understand what Adams was going for
Black should again try and activate the
when watching the remainder of the
sleeping d7-knight with 24...Nf8!.
game. He might have been ecstatic
25.Rec1 already about being given the
opportunity to enter a position where
he’d dominate the light squares.
8
*+tT*+l+ Correct was 27.Ba5!, creating the dual
7
+w+m+oVo threats of Bxd8 and Rc7. After 27...Rc8

*+*+*+o+
6 28.Rxc8 Qxc8 29.Qxb5 White has good
winning chances, as apart from the extra
5
+o+*O*+* pawn, he also has the relatively more
4 *B*+p+*+ active pieces.
3 +q+p+n+p 27...Rxf8 28.Rc5
2 *+r+*Pp+
1 +*R*+*K*
a b c d e f g h
 42.Rxb2 Kd5. Black easily holds the
8
*+*+*Tl+ draw, due to the more active rook and
7
+w+*+oVo king.
6
*+*+*+o+ 29.Qc4!
5
+oR*O*+*
*+*+p+*+
4 Meskovs likely missed this idea,
creating the dual threats of Rb5 and Rc7.
3 +q+p+n+p
2 *+*+*Pp+ 29...Qd7
1 +*+*+*K* There is no way for Black to hold on
a b c d e f g h to his b-pawn, since 29...Rb8?? loses
instantly to 30.Rc7.

30.Rc7
Black to move
Adams gains a tempo, while seizing
Exercise: Should Black play control over the seventh rank.
28...b4, or should he back up the b-
pawn with 28...Rb8? One line likely 30...Qa4?!
holds the draw, while he is in
trouble in the other. Which one Such a move tells me that Black
would you play? underestimates White’s attacking
chances, since the queen is moved away
from the black king.
Answer: Black should follow the
Principle: Rooks belong behind passed 30...Qd8 keeps an eye out for Ng5
tricks and might be slightly better than
pawns, and play the rook to b8. The b-
the move played. Even there though,
pawn is far more important for Black
than his e-pawn. Black’s position lingers halfway
between life and death.
28...b4?

An illustrative sample line of how

28...Rb8! holds could be: 29.Nxe5 Bxe5
30.Rxe5 Qc7 31.d4 b4 32.Rc5 Qf4
33.Qe3 Qxe3 34.fxe3 b3 35.Rc1 b2
36.Rb1 f5! 37.exf5 gxf5 38.Kf2 Kf7
39.Ke2 Rb3! 40.Kd2 Ke6 41.Kc2 Rxe3
 32.Kg2 Kg7 33.Rb7 Qa3 34.Qd5!
8
*+*+*Tl+
+*R*+oVo
7 Adams once again prefers board
control to the capture of a mere pawn
6
*+*+*+o+ after 34.Rxb4?!, which is still winning
5
+*+*O*+* but inferior due to 34...Re8!. In this
4 wOq+p+*+ version, Black’s rook can be transferred

+*+p+n+p
3 to its optimal defensive post on e7.

2 *+*+*Pp+ 34...Qc3
1 +*+*+*K*
a b c d e f g h 8
*+*+*T*+
7
+r+*+oLo
6
*+*+*Vo+
+*+qO*+*
White to move 5
31.g3! 4 *O*+p+*+
There is no rush to win Black’s b-
3 +*Wp+nPp
pawn since he is tangled up in the
2 *+*+*Pk+
defense of f7. White’s last move offers 1 +*+*+*+*
his king access to g2, which is a b c d e f g h
preferable to the h2-square, as from g2
the king protects the f2-pawn and sits on
a light square.
31.Ng5? would be a mistake due to White to move
31...Qa1+ 32.Kh2 Qd4!.
Let’s first gather data. Telltale clues of
31...Bf6?! the degeneration of Black’s position are
all around us:
After this, Black never gets even half a 1) Black is completely tied down to
chance. the defense of the f7-pawn.
He should be doing everything in his 2) Black’s b-pawn isn’t going
power to swap queens with 31...Qa1+ anywhere, since White controls the b3-
32.Kg2 Qc3!, forcing a queen trade. square.
White is unable to play Qd5, as in the 3) Black is stuck with a bad bishop
game, since the rook on c7 would hang. versus a potentially good white knight.
 4) White has the only pawn break in 38.Rxf7 Be7
the position with d3-d4.
5) Black bleeds on the light squares, This just hangs the bishop. But in any
which White dominates. case, Black was lost. 38...Rxf7 allows
39.Qxd8+ Kg7 40.Ne6+ Kh6 41.Qh4
Conclusion: White has reached a
mate, while 38...Rg8 is met with 39.Qd7
classic position where he torments his
and h7 can’t be defended.
opponent on both wings, due to the
principle of multiple weaknesses. 39.Rxe7 1-0
White is clearly winning, but now
comes the question of working out the 34.Qd5!, retaining the bind and
details of how the goal is accomplished. tightening the knot further, is a powerful
Black has reached that point in the war move that separates a strong GM from
movie where the gunner yells in club-level players, many of whom
desperation: “There are too many of would be tempted to pause their
them, Sarge! I don’t know how much initiative to win a pawn with 34.Rxb4?!.
longer we can hold them off!” The lesson to learn from such a move is:
board control tends to be more important
Exercise: If you find Adams’
than material. After stretching Black’s
powerful idea, Black’s position falls
defenses to the maximum by constantly
apart.
threatening the pawns on b4, e5, and f7,
White managed to crash through. In the
Answer: Sacrifice/overloading.
end, Adams got much more than a
pawn.
35.d4! exd4 36.e5!

The precarious black construction now I’m convinced that Magnus Carlsen is
collapses. some kind of alien experiment involving
a human/machine hybrid. In this four-
36...Bd8 37.Nxd4 game final match that Magnus won 2.5-
1.5, these were his accuracy numbers:
Threatening: Ne6+. It doesn’t matter 98.9%, 99.2%, 96.5%, and 97.8% – and
that material is limited; White’s attack is these were rapid games! How is this
in full force on the light squares, and the possible? He also has an eerie habit of
black king is getting checkmated. defeating world-class opposition from
dry, bloodless, simplified positions like
37...Kh8
this one.
37...Kg8 38.e6! fxe6 39.Nxe6 is also Game 6
devastating for Black.
 Wesley So – Magnus Carlsen with a hole on g4. This pawn push has
Toronto (rapid) 2023 also created numerous weaknesses on
the light squares around the white king,
granting Black some hopes of launching
8
*+*+*+*+ an attack.
7
+*+w+oL* Conclusion: The game is judged even

*+*O*Oo+
6 by the engine, yet I slightly prefer
Black’s side, since it’s difficult to
5
OoTp+*+o envision any way for Black to lose (a
4 *+*R*P*P perpetual check should always be
3 P*Pq+*P* available), while I can see situations
where White loses due to a slight king
2 *P*+*+k+ safety deficit.
1 +*+*+*+*
a b c d e f g h 28...Rc8

Note that Magnus is in no hurry to

play 28...f5, since it blocks his queen’s
Black to move access to g4 and slightly weakens the
dark squares around his king. It also
It’s difficult to dazzle the reader with negates the outside chance of opening
one of Carlsen’s technical endings, yet the kingside with ...g6-g5.
we should indeed be dazzled, since no
other player in the world could have 29.Re4
beaten Wesley So from this position.
Let’s begin by gathering data: White would like to mobilize his
1) Material is even and reduced to a queenside majority, yet 29.c4? is
queen and rook each, which signals a premature: 29...bxc4 30.Rxc4 Qb5!
likely drawn outcome. 31.b3 Re8 leaves White’s king seriously
exposed.
2) Black’s f-pawns are doubled, which
means that White’s queenside majority 29...Rb8
is more likely to produce a passed pawn
than Black’s on the kingside. This way White will not be tempted to
3) The doubling of the f-pawns means play c3-c4, since then Black’s rook
that Black has the safer of the two kings. would attack b2.
4) White has pushed both f- and h-
30.Kh2 Qf5
pawns, which means that he has to deal
 If So is playing for a win, then this
8
*T*+*+*+ move is no more than a series of broken
7
+*+*+oL* promises. The refusal to undouble

*+*O*Oo+
6 Black’s f-pawns is a risky decision,
which later comes back to haunt him.
5
Oo+p+w+o White’s f- and h-pawns become isolated,
4 *+*+rP*P making them possible targets. Well,
3 P*Pq+*P* doesn’t that mean that Carlsen will also
be straddled with two doubled isolated f-
2 *P*+*+*K pawns? Yes, but those are unreachable
1 +*+*+*+* for the white king and rook, while
a b c d e f g h White’s weaknesses are readily
accessible.
A draw is the likeliest outcome after
the less ambitious 34.hxg5!, which
White to move
allows Black to eliminate his stagnant
31.Re3 kingside pawn majority: for example,
34...fxg5 35.fxg5 Re5 36.b3 Rxg5 37.c4
So offers to enter a drawish rook bxc4 38.bxc4 Kf8 39.Rb2 Rg4 40.Rb5
ending. Rxc4 41.Rxa5 with an inevitable draw.

31...Qxd3!? 34...gxh4 35.gxh4 Re3!

It’s difficult to find progress for Black Placing White in a semi-zugzwang.

if he retreats his queen to d7.
36.f5?!
32.Rxd3 Re8!
It was time for So to apply the
Carlsen seizes the all-important e-file. Principle: Keep your rook as active as
possible, even if it costs you a pawn. He
33.Rd2 g5!? should have played: 36.Kf2! Rh3
37.Kg2 Rxh4 38.Rd4 Rg4+ 39.Kh2 Kg6
Carlsen has shown during his career
40.c4!. White generates counterplay by
that he values piece activity over
creating a passed pawn on the queenside
everything else. This way, Black’s king
and should hold an easy draw.
has access to the center via g6 and f5.

34.Kg2?!
 save the game with 38.Rd4!.
8
*+*+*+*+
7
+*+*+oL* 8
*+*+*L*+
6
*+*O*O*+ 7
+*+*+o+*
5
Oo+p+p+o 6
*+*O*O*+
4 *+*+*+*P 5
Oo+pTp+o
3 P*P*T*+* 4 *P*+*+*P
2 *P*R*+k+ 3 P*P*+*+*
1 +*+*+*+* 2 *+*R*K*+
a b c d e f g h 1 +*+*+*+*
a b c d e f g h

Black to move
Black to move
36...Kf8!
38...axb4?
We reach the “some assembly
required” moment. Black’s king Principle: The defender tends to
activates itself via a retreat! benefit from a mutual eradication of
Carlsen’s move is far stronger than the weaknesses.
greedy 36...Re5?!, which can be met Correct though was: 38...Rxf5+!
with 37.b3!. Now after 37...Rxf5 38.c4! 39.Ke3 Re5+ 40.Kf4 a4!, keeping the a-
White has full compensation for the pawn alive. Both sides have an extra
missing pawn, partly due to the threat of pawn to lose on the queenside, but that
c4-c5. generally benefits Black, who has a head
start in every race, given that he has
37.Kf2 Re5! both an extra pawn and the more active
The correct moment, with White’s rook. In reality, White’s extra headache
king on the f-file, ...Rxf5 arrives with of potentially losing the a3-pawn
check. guarantees a decisive advantage for
Black: for example, 41.Rd4 Ke7 42.Rd3
38.b4? Kd7 43.Rd4 Re2 44.c4 Rf2+! 45.Ke3
Ra2 highlights why it was good to keep
This is not the fastest way of creating the a-pawns alive. Black wins.
a passer. White had better chances to
39.axb4 Rxf5+ 40.Ke3 Re5+ 41.Kf4 mostly symbolic, as White doesn’t have
Ke7 enough weaknesses, and Black’s
position is full of them.

8
*+*+*+*+ 42.Rd4 Kd7 43.Rd1 Re2!
7
+*+*Lo+* From e2, the rook can quickly access
6
*+*O*O*+ all three weak white pawns.
5
+o+pT*+o
*P*+*K*P
4 44.Ra1!

3 +*P*+*+* Highlighting the second reason why

2 *+*R*+*+ 38...axb4? was a mistake. The white

+*+*+*+*
1 rook suddenly gets the chance to go
active, and a player of So’s caliber
a b c d e f g h would never hesitate to abandon the
defense and go for activity.

44...Rd2 45.Ra7+ Ke8 46.Kf5!? Rf2+

White to move
47.Ke4 f5+ 48.Ke3 Rc2
Let’s take an inventory:

*+*+l+*+
1) Black has won a pawn. 8
2) All five of Black’s pawns are
isolated, which means they are
7
R*+*+o+*
devalued.
6
*+*O*+*+
3) White’s king is active yet cut off 5
+o+p+o+o
from the queenside. Hence, Black can 4 *P*+*+*P
+*P*K*+*
play according to the principle of two 3
(or more) weaknesses, by shifting
attacks on the weak white c- and h-
2 *+t+*+*+
pawns. However, White should be able 1 +*+*+*+*
to either protect his pawns or go on a a b c d e f g h
swift counterattack against one of the
many black weaknesses.
Conclusion: Black’s advantage is
White to move
undeniable, but only in the sense that he
is the one pushing. Otherwise, it is 49.Kf4!
 Active counterplay is the solution – However, in a calmer situation, So
not passive defense. As mentioned would have probably gone for 53.Kf6!,
above, a mutual eradication of trying to get to the already passed f-
weaknesses, or in other words, pawn pawn, which is far more important than
trades, almost always benefits the the pawn on h5.
defender.
After 49.Kd4? Rh2! Black takes 53...b4 54.Kh6!
another pawn and gains a winning
54.Kg5?? loses to 54...b3 followed by
advantage.
...b3-b2, when the king has no place to
49...Rxc3 50.Kxf5 Rc4 hide from the upcoming check and ...b2-
b1=Q. If he tries hiding on the h-file, the
New targets emerge on opposing f-pawn starts running.
wings: b4 and h4.
54...f5!
51.Kg5 Rxb4 52.Rb7 Rb1
Keeping the most chances alive. I just
can’t believe that Carlsen calculated
8
*+*+l+*+ everything, so it must be his incredible
7
+r+*+o+* intuition that led him to the correct path
once again.
6
*+*O*+*+ A commoner like myself would be
5
+o+p+*Ko tempted by the natural 54...b3!?, but
4 *+*+*+*P then White performs a miraculous save:
3 +*+*+*+* 55.Kg7! (otherwise, ...b3-b2 followed
by ...f7-f5 wins) 55...b2 56.Rb8+! Kd7
2 *+*+*+*+ 57.Kf8!!. The white king manages to
1 +t+*+*+* hide from the checks without
a b c d e f g h obstructing the h-pawn, and White
draws without experiencing any more
trouble.

White to move

53.Kxh5!?

It is logical for White to want to take

back one of the two missing pawns,
especially in a rapid time control.
 check in three different ways: he could
8
*+*+l+*+ hide the king behind the h-pawn,
7
+r+*+*+* preventing the check; he could go to g2

*+*O*+*K
6 or h2; he could simply approach the b-
pawn and catch it) 56.h5 b2 57.Kf6! f4
5
+*+p+o+* 58.h6 when Black needs to abandon the
4 *O*+*+*P b-pawn and take the white h-pawn,
3 +*+*+*+* leading to a complete draw.
2 *+*+*+*+ 55...f4!
1 +t+*+*+* Oops!
a b c d e f g h
So might have been counting on
55...b3 56.Kg5, transposing to 55.Kg5!,
above. However, after Carlsen’s move,
White to move the white king loses any chance of
hiding behind the f-pawn, since it’s
Exercise: Black’s last move is a actually the f-pawn that’s going to
good practical try to obstruct queen!
justice. The game is still drawn, yet
the move forces White to find the 56.Kg5 f3
correct response. Should White
move his king to g5, or should he Magnus hopes to clear up a
push his h-pawn? misunderstanding. It is he – not his
opponent – who owns the faster passed
pawn.
Answer: Moving the king to g5 is the
only way to hold, as similarly to the note 57.h6 f2
above regarding 54...b3!?, the king
needs to use the two tempi available
before Black’s b-pawn reaches b2 to
hide behind Black’s f-pawn.

55.h5??

The only way to do it was: 55.Kg5! b3

(55...f4 is trying too hard to be smart;
after 56.Kxf4 b3 White can deal with
Black’s threat of ...b3-b2 followed by
 which went after White’s pawn targets
8
*+*+l+*+ on c3, d5, and h4. As we’ve said, it’s the
7
+r+*+*+* defender who benefits from a mutual

*+*O*+*P
6 eradication of weaknesses, so in most
cases, the attacking side should try to
5
+*+p+*K* keep on the board as many pawns as
4 *O*+*+*+ possible.
3 +*+*+*+*
*+*+*O*+
2 Some of us don’t like living in the real
world, which is why I love endgame
1 +t+*+*+* studies so much. Let’s end the chapter
a b c d e f g h with a brain-twisting portion of a study,
where the position comes at us with
sharp contours and angles.
White to move Game 7
End of a Study by Otto Blathy
58.Rb8+ The Chess Amateur, 1928

58.h7 Rg1+ is similar to the game.

58...Kd7 59.Rb7+
8
*+*+*+*+
7
+*+*+*+o
59.h7 Rg1+ 60.Kh4 Rh1+ 61.Kg5
f1=Q wins.
6
*+*+*+*P
5
+oL*+*O*
59...Kc8 60.h7 Rg1+! 0-1 4 o+o+o+p+
White’s king shuffles off dejectedly,
3 P*Po+oP*
like a dismissed game show contestant
2 *R*P*P*+
who botched the answer. If White’s king 1 KnB*+*+n
moves over to the f-file, Black promotes a b c d e f g h
with check, while if White’s king moves
to the h-file, then Black gives a rook
check on h1, stopping h7-h8=Q.
Even though Carlsen wasn’t winning White to move
White to play and win
for a great portion of this game, his win
came from constant harassment of the We must identify Black’s weakest
opponent’s weaknesses with the rook,
pawns. Then we should find a way to maneuver.
make Black’s king run back and forth to Exercise: This is a completely
defend the weak pawns, all the while conceptual study. You can solve it
plotting to throw him out of sync by only if you identify which pawn in
following the principle of two (or more) Black’s camp is the weakest. We
weaknesses. You may be thinking the must then find a way to turn this
diagrammed position is some kind of pawn into a potential sacrificial
prank to troll the reader. I assure you it target.
isn’t. Let’s consider:
1) How can the side that is up a rook
Answer: There exists a single winning
and three minor pieces not win? While
plan, based upon Black’s only pawn
White is up a mountain of material, the
weakness in the position: the g5-pawn,
claustrophobic pieces set a world record
which is the true face of the black
for inefficiency. We quickly understand
position’s failure. White will:
that freedom is a feature distinctly
1) Transfer the rook to h5.
missing from White’s position. In fact,
White’s dysfunctional pieces resemble 2) Weave White’s king over to h2 or
victims, trapped in the twisted wreckage h3.
of a terrible car accident. 3) Sacrifice the rook on g5.
2) Neither knight has access to a single 4) Force entry with White’s king to
legal move, since unwanted white pawns h4, after which the h1-knight emerges
gum up the exit squares. with g2-g4 and Ng3.
3) White’s sad-looking bishop has a While the technique is purgatorially
single move to b2 and then maybe to a1, lengthy, it’s also relatively clear.
but so what? The bishop is still trapped
6.Rb4!
within its cage.
4) Sacrificing the rook for the c4-, b5- Step 1: White’s rook must clear the
or a4-pawn fails miserably, since the b2-square to transfer the king to d1.
remainder of White’s pieces are unable
to arise and remain uselessly immobile, 6...Kc6 7.Ka2!
as Black achieves a fortress draw.
5) Black has three potential Step 2: The bishop must be shifted
over to a1 to clear White’s king a path to
weaknesses: the b5-, e4- and g5-pawns.
the kingside via b2 and c1.
We must find the correct pawn for
which to sacrifice our rook, which is the 7...Kc5 8.Bb2! Kc6 9.Ba1! Kc5
only piece with which we are able to
10.Kb2 Kc6 11.Kc1 Kc5 12.Kd1
 The king has completed the first step
of his arduous journey. 8
*+*+*+*+
7
+*+*+*+o
8
*+*+*+*+ 6
*+*+*+*P
7
+*+*+*+o 5
+o+l+*O*
6
*+*+*+*P 4 o+o+o+p+
5
+oL*+*O* 3 P*Po+oP*
4 oRo+o+p+ 2 *+*P*P*+
3 P*Po+oP* 1 BnRk+*+n
2 *+*P*P*+ a b c d e f g h
1 Bn+k+*+n
a b c d e f g h
White to move

16.Ke1!
Black to move

12...Kc6 13.Rb2! Step 4: Transfer the king to h2.

Step 3: Transfer the rook to c1 via b2 16...Kd6 17.Kf1 Kd5 18.Kg1 Kd6
and c2.
When I posted this study on my
13...Kc5 14.Rc2! Kd6 Facebook group, Chess Endgame
Studies and Compositions, GM Nigel
14...dxc2+? unblocks the d-pawn, and Short pointed out that on 18...Ke5!
after 15.Kxc2 Kd5 White can play White should play 19.Re1!, since if
16.d4! when the game is over, since Black’s king is allowed to take the h6-
both the bishop on a1 and the knight on pawn and then return to f6, the position
b1 get to escape. becomes a fortress draw.

15.Rc1! Kd5 19.Kh2 Kd5

 8
*+*+*+*+ 8
*+*+*+*+
7
+*+*+*+o 7
+*+*+*L*
6
*+*+*+*P 6
*+*+*+*O
5
+o+l+*O* 5
+o+*+*Or
4 o+o+o+p+ 4 o+o+o+p+
3 P*Po+oP* 3 P*Po+oP*
2 *+*P*P*K 2 *+*P*P*+
1 BnR*+*+n 1 Bn+*+*Kn
a b c d e f g h a b c d e f g h

White to move White to move

20.Re1! The position is simultaneously funny

and not-so-funny. It reminds us of that
A critical detail: Black’s king must be sad situation where the boss retells an
forced to defend the e4-pawn. This way, unfunny joke for the 100th time, and
it is prevented from walking over to g6 you’re dutifully forced to laugh, even
and capturing the h-pawn. though you would actually prefer to
White could still blow it: vomit. The position undergoes a
20.Kh3?! Kd6 21.Rg1?! Ke5 22.Rg2 shocking reappraisal, and Black
Kf6 23.Kh2? (It wasn’t too late to play achieves a fortress draw.
23.Rg1! followed by Re1, switching
back to the correct plan) 23...Kg6 20...Ke5 21.Bb2 Kd5
24.Kg1 Kxh6 25.Rh2+ Kg7 26.Rh5 h6.
 Black’s single, undefendable weakness:
8
*+*+*+*+ the g5-pawn. There is no defense to the
7
+*+*+*+o plan of Kh2, Rxg5! and Kh3.
6
*+*+*+*P 26...Kg6 27.Kh2 Kf6
5
+o+l+*O*
4 o+o+o+p+ 8
*+*+*+*+
3 P*Po+oP* 7
+*+*+*+o
2 *B*P*P*K 6
*+*+*L*P
1 +n+*R*+n 5
+o+*+*Or
a b c d e f g h 4 o+o+o+p+
3 P*Po+oP*
White to move
2 *B*P*P*K
1 +n+*+*+n
22.Rg1! a b c d e f g h
Step 5: The rook is shifted over to the
h-file. Note that this shift must be
perfectly timed, as when on d5 Black’s White to move
king is too slow to walk over and
capture the h6-pawn. Now we break down the wall.

22...Ke5 23.Rg2! Kf6 28.Rxg5!

23...fxg2 24.Kxg2 Ke6 25.f3 wins for Step 6: An annihilation of the

White since the h1-knight emerges into defensive barrier, and Black’s dream of
broad daylight. a fortress draw is destroyed.

24.Kg1 Kg6 25.Rh2 28...Kxg5 29.Kh3 Kxh6 30.Kh4

Just in time. White’s h6-pawn is The h1-knight emerges after we push

secured. both g-pawns, with an easy win. For
example:
25...Kf6 26.Rh5!
30...Kg6 31.g5 Kf5 32.g4+ Kg6
At last, we identified and targeted 33.Ng3
 Game over. White’s winning plan
came after the following realizations:
1) Black’s only targetable weakness in
the initial position is the g5-pawn.
Chapter Two
2) Before we get to g5, we need to
Strong Pawns and Weak
throw the black king out of sync by Pawns
forcing him to defend e4.
3) Details matter. We should go for Pawns, when separated from other
the transfer of our rook to the h-file only pawns, can turn into liabilities, requiring
when Black’s king is on d5. continual support from the pieces. This
effect can be even more dire than the
Conclusion actual loss of the weak pawn, since even
The principle of multiple weaknesses a single weak pawn which requires
has a special place in the ranks of game- defense can act as an energy drain on
winning techniques, since it has the our piece activity. In this chapter, we
habit of throwing the defense out of shall look at the dynamics of what it
sync, as our opponent scrambles to means to attack or own a weak pawn or
protect one weak point after another. pawns.
This chapter can be seen as a huge The following game is one I wanted to
reminder: don’t fixate on a single task. If show to demonstrate that sometimes
you can’t break your opponent’s pawns can look weak, yet their dynamic
defenses by piling up the pressure on potential often overrides their prospects
one front, then it might be a good idea to of being a burden to their owners. In this
consider opening another. The most case, Kasparov demonstrated that his
common way of overcoming a tough spirited hanging pawns on the d- and c-
defense is stretching it to breaking point. files were more of a dynamic strength
than a static weakness. Larsen-Kasparov
clashes remind us of Nimzowitsch
versus Alekhine battles. The Danish pair
didn’t fare too well against their
firebrand opponents.
Game 8
Bent Larsen – Garry Kasparov
Niksic 1983

1.c4 e6 2.Nc3 d5 3.d4 c5 4.cxd5 exd5

5.Nf3 Nc6 6.g3 Nf6 7.Bg2 Be7 8.0-0 0- dynamic style.
0 9.b3 The isolani route with 12...cxd4 is also
played here. After 13.Nxd4 Bxd4
This line is today considered a bit too 14.Bxd4 Bf5 Black has an IQP, but also
milquetoast to offer White an edge, if active pieces and good central control.
you ask theory or the engines. Yet when
looking at the database, White scores 13.dxc5 Bxb2 14.Nxb2 bxc5 15.Na4
slightly above average.

8
t+vW*Tl+
8
t+vWt+l+
7
Oo+*VoOo
7
O*+*+oOo
6
*+m+*M*+
6
*+m+*+*+
5
+*Oo+*+*
5
+*Oo+*+*
4 *+*P*+*+ 4 n+*+m+*+
3 +pN*+nP* 3 +p+*+nP*
2 p+*+pPbP 2 p+*+pPbP
1 R*Bq+rK* 1 +*Rq+rK*
a b c d e f g h a b c d e f g h

Black to move Black to move

9...Ne4! 15...Ba6!

This aggressive hop has been known Are Black’s central pawns weak or
to equalize since 1932. The idea is to strong? The answer is yes, and yes! The
clear the path for ...Bf6. pawns do indeed have the potential to be
weak, but at the moment, they exert
10.Bb2 Bf6! 11.Na4 pressure by occupying the center. Note
that the c5-pawn is tactically protected
11.e3 Bg4 is about even. due to the latent attack on e2.

11...Re8 12.Rc1 b6 16.Re1

Kasparov is OK with hanging pawns, The e2-pawn is protected, which

since the structure perfectly suits his means that the Nxc5 threat is renewed.
16...c4! d8-square for the a8-rook. All of Black’s
pieces are more active than their
Active play is the only solution when counterparts, and Kasparov is already
our structure has more weaknesses than nursing an annoying initiative.
our opponent’[Link] to hold on to the
pawns for too long with something like 18.Nf5
16...Qd6 can backfire in many ways,
including 17.Nxc5!? Nxc5 18.Qc2. 18.bxc4 would be met with 18...Rad8!
when the engine somehow keeps the
17.Nh4?! white position together, but a human
would rightly be trembling with fear.
The knight goes to the rim, reducing
White’s activity even further. 18...g6!?
17.Nd2 is more prudent, trying to
trade off some of Black’s active pieces, Kasparov is not the type of player who
but 17.Nd4 runs into the annoying tolerates intruders within his territorial
17...Qf6!, with a double threat on d4 and boundaries.
f2. Also good was the immediate
18...Rad8!.
8
t+*Wt+l+ 19.Nd4
7
O*+*+oOo
6
v+m+*+*+ Larsen’s knight is a model of
inefficiency, as it took two extra moves
5
+*+o+*+* to reach d4 when it could have done so
4 n+o+m+*N immediately on move 17.
3 +p+*+*P* 19...Rac8
2 p+*+pPbP
1 +*RqR*K*
a b c d e f g h

Black to move

17...Qa5!

Activating the queen and clearing the

 Now ...Ng4 is in the air.
8
*+t+t+l+ 21.Bh3 Rc7 22.Nc2?
7
O*+*+o+o
6
v+m+*+o+ Reducing the energy of White’s army.
5
W*+o+*+* The engine is in panic mode and calls

n+oNm+*+
4 for 22.Rf1.

3 +p+*+*P* 8
*+*+t+l+
2 p+*+pPbP 7
O*T*+o+o
1 +*RqR*K* 6
v+*+*+o+
W*+oM*+*
a b c d e f g h 5
4 n+o+m+*P
White to move
3 +p+*+*Pb
2 p+n+pP*+
20.h4?! 1 +*RqR*K*
a b c d e f g h
A move based on the belief: it’s better
to try risky action, fail, and die, rather
than do nothing and live with sorrow
and regret. However, the move violates Black to move
the Principle: Base your decision on the
position’s facts, not emotional Exercise: White’s last move
preference. We all know how Larsen – allowed a combination. What
sort of an early human version of should Black play?
AlphaZero – felt about rook’s pawns: he
loved to push them! In this case, the Answer: So dynamic is Black’s
push is illogical, since White isn’t position that he can actually place a
attacking, which in turn weakens the g4- knight on d3.
square and the g3-pawn, amplifying the
risk to White’s king more than Black’s. 22...cxb3
Any move that would have kept White’s
set-up intact would have been a better This still gives Black a winning
choice. advantage.
Even stronger was 22...Nd3!!, after
20...Ne5! which Black’s dynamic potential
explodes: for example, 23.exd3 cxd3 play. Who is going to guard f2?
(among many others, ...d3-d2 is a threat)
24.Re3 dxc2 25.Rxc2?! d4! and White’s 25.Rf1
rook has no place to go. 26.Rf3? is
25.f3? loses to 25...Ngf2 26.Qd4
crushed by 26...Ng5! with deadly dual
Rxc2!. It’s incredible how many tactics
threats on f3 and e1.
related to the overloaded piece theme
23.axb3 Bc8!? appear when a position is full of
weaknesses.
Kasparov is trying to gain control over
the g4-square. 23...d4! was both 25...Bd7?!
thematic and strong. The main idea is
That’s not the most active post for the
revealed in the line 24.Nxd4 Nxg3!!
bishop.
25.Rxc7 (25.fxg3 Rxc1 26.Qxc1 Nd3 is
also crushing) 25...Qxc7 26.fxg3 Ng4! The e2-pawn is a much bigger
when the white position collapses. weakness and should be targeted with
25...Ba6!. The knight on a4 is no match
24.Bg2 for the bishop on a6, so why go about
exchanging them?

8
*+v+t+l+ 26.Ra1 Bxa4?!
7
O*T*+o+o This trade only helps White repel
6
*+*+*+o+ some of the pressure. It was Mikhail Tal
5
W*+oM*+* who said, “Chess is the struggle against
4 n+*+m+*P errors.” Kasparov’s position went from
crushing to even in the space of just four
3 +p+*+*P* moves. He didn’t blunder – he simply
2 *+n+pPb+ made multiple subpar moves, which
1 +*RqR*K* were enough to sink his once crushing
evaluation.
a b c d e f g h
Keeping more pieces on the board
with 26...Qb5!? 27.Nd4 Qb8 was a
better option, keeping some advantage.
Black to move However, in this case as well, it is clear
that Larsen has left the worst behind
24...Ng4!
him.
The point of Kasparov’s previous
27.Rxa4 Qc3 threatens both the rook and the knight,
but Larsen has an ace up his sleeve.
29...h5 30.Nd4 leaves White with zero
8
*+*+t+l+ problems.
7
O*T*+o+o
6
*+*+*+o+ 30.Rxe4!
5
+*+o+*+* The hidden trick that saves White.
4 r+*+m+mP 30.Nd4? leaves Black up a pawn in
3 +pW*+*P* the ending.
2 *+n+pPb+ 30...Rxe4 31.Qd8+ Kg7 32.Qxc7 Rc4
1 +*+q+rK* Double attack. Has Larsen been
a b c d e f g h
tricked? Actually not.

33.Nd4!
White to move
The final detail.
Exercise: If you find Larsen’s exact
33...Rxc7 34.Nxb3
sequence, White can reach a
drawable ending.
8
*+*+*+*+
Answer: Exchange then stabilize. 7
O*T*+oLo
28.Bxe4!
6
*+*+*+o+
5
+*+*+*+*
*+*+*+mP
The battle cries against the white king 4
begin to grow fainter.
3 +n+*P*P*
28...dxe4 29.e3! 2 *+*+*P*+
Stopping ...e4-e3, thereby stabilizing
1 +*+*+rK*
the kingside and threatening the black a b c d e f g h
knight.

29...Qxb3
Black to move
The only ambitious approach. Black
 Wow. You don’t see Kasparov out-
calculated too often! Larsen proves that 8
*+*+*+*+
the old guard still knows how to play 7
+*+*+oLo
o+n+*+o+
chess. Black holds an edge after his next 6
move, due to his more active rook and
passed a-pawn, but the position now
5
+*+*+*+*
looks drawable. 4 *+*+p+mP
34...Rc2 35.Nd4 Ra2 36.e4!?
3 +*+*+*P*
2 *+*T*P*+
The idea behind this itchy move is f2-
f3, followed by Rf2. There are other
1 +*+*+rK*
a b c d e f g h
ways to do it, but somewhat
surprisingly, this is also the engine’s
first line. The obvious downside is that
White disconnects the e-pawn from the White to move
rest of his structure and destabilizes the
knight. 38.e5?!

36...Rd2 This complicates. It was better to stick

to the plan with 38.f3! Ne3 39.Re1! Nc4
36...a5 is met with 37.f3 and 37...Ne5 40.Rc1 Nb2 41.Rc3!?. With the knight
runs into 38.Rc1! followed by Rc5, with on b2, there are no realistic hopes for
active counterplay. Sure, the white Black to create problems.
structure is fragile, but tactics are more
important than generic remarks. 38...Re2!

37.Nc6 a6 White is in zugzwang and needs to

decide what to give away.

39.Ra1!

Trying to be clever with 39.Kg2??,

thinking that ...Ne3+ can safely be met
with Kf3, loses on the spot to: 39...Rc2!
40.Nd4 Rd2 41.Nf3 Ne3+ 42.Kg1
Nxf1. The rook on d2 is protected.

39...Rxf2 40.Rxa6 Rc2!

 There is no use going after the e5- 41.h5!?
pawn. For example, after 40...Re2
41.Kf1 Re3 42.Kg2 Nxe5 43.Nxe5 Exchanging pawns favors the
Rxe5 44.g4 the position is trivial to hold defending side, but even clearer is:
for a player of Larsen’s caliber. 41.e6! (getting rid of the weakness)
41...fxe6 42.Nd4 Rc1+ 43.Kg2 when
8
*+*+*+*+ White’s king has been let out of his
cage, so the task is much simpler. After
7
+*+*+oLo 43...e5 44.Ne6+ Kf6 45.Ng5+ Kf5
6
r+n+*+o+ 46.Ra3! White’s threats include Rf3
5
+*+*P*+* mate and the simple Nxh7. It should be
4 *+*+*+mP a stress-free affair from now onwards.
3 +*+*+*P* 41...Kh6
2 *+t+*+*+ Larsen’s tactical point was that
1 +*+*+*K* 41...gxh5 runs into 42.Nd4 Rf2 43.e6!,
a b c d e f g h with an easy draw.

42.hxg6 hxg6

White to move

This is one of those positions that is

8
*+*+*+*+
easily drawn for engines, but unpleasant
7
+*+*+o+*
for humans: 6
r+n+*+oL
1) The material is greatly reduced, 5
+*+*P*+*
which benefits White’s hopes to score a
draw.
4 *+*+*+m+
2) Black can go after the white pawn
3 +*+*+*P*
targets on e5 and g3.
2 *+t+*+*+
3) White’s light squares around his 1 +*+*+*K*
king are punctured. This means that a b c d e f g h
Black has some hope of later moving his
king up the board and slipping in,
creating a mating attack with king, rook,
White to move
and knight against White’s insufficiently
protected king. 43.Ra4
 Again, a clearer path to the draw lies
in 43.e6! following the same Principle: 8
*+*+*+*+
The defending side benefits from pawn
swaps. After 43...fxe6 44.Nd4 Rc1+
7
+*+*+o+*
45.Kg2 e5 46.Nf5+ Kg5 47.Nh4 White
6
*+*+p+o+
holds the draw.
5
+*+*+*L*
4 r+*N*+m+
+*+*+*T*
43...Kg5 44.Nd4 3
Larsen keeps making moves that 2 *+*+*+*+
should lead to a draw, yet misses the
ones that force clear draws. He could
1 +*+*+*+k
a b c d e f g h
seal a draw with 44.Ra6! followed in
most cases by e5-e6, exchanging more
pawns. In such dull, technical affairs,
especially when we are on the defending Black to move
side, it’s good to set ourselves a task (in
this case, trading pawns), so that we 46...f5!
don’t forget what we’re fighting for.
This is Black’s only prayer to win, as
44...Rc3 45.e6! it gives him two connected passed
pawns. White’s far-advanced e-pawn
The pawn is finally pushed, but now should give him enough play to hold the
it’s not as clear as it was before. draw, but at least he will have to work
for it.
45...Rxg3+ 46.Kh1
47.e7 Re3 48.Nc6 f4 49.Ra5+?!

Why drive Black’s king closer to

White’s? This could lead to a mating
net. Principle: In endgames, don’t drive
the opponent’s king into play with
useless checks.
49.Ra3! is a clearer path to the draw.

49...Kh4! 50.Ra8 Nf6

 kid-excuse: “I didn’t do it, but if I did, it
8
r+*+*+*+ was totally justified!” It really sucks to
7
+*+*P*+* get old and blunder from exhaustion in

*+n+*Mo+
6 positions you would have held in your
youth. Larsen’s last move was a
5
+*+*+*+* miscalculation.
4 *+*+*O*L 51.Rf8?? completely misses the point
3 +*+*T*+* and after 51...Kg3 White is mated.
2 *+*+*+*+ After 51.Kh2! the line we need to
examine is 51...f3. Black is threatening
1 +*+*+*+k checkmate in two starting with ...Ng4+,
a b c d e f g h yet 52.e8=Q! drives the black knight
away from the kingside: 52...Nxe8
53.Ra4+! (the knight needed to be
White to move driven to e8 first, so that it couldn’t
block this check) 53...Kh5 54.Kg3 and
We’ve all been here. It’s a long, everything is in order. The white king
exhausting game, and what is required is isn’t in trouble anymore, and White can
superhuman effort and vigilance. Then, pick off the pawn on f3 in several ways.
our treacherous brain whispers slyly: Black will keep one extra pawn, but it
“Why work so hard? Rest a while.” That shouldn’t be enough.
is when we blunder.
51...f3+!
Exercise: White’s candidate moves
are: a) 51.Kg2, b) 51.Kh2, c) Larsen’s anguish level just went up
51.Rf8. Only one of the moves about 10 notches. This move costs
allows White to draw. Which one? White a critical tempo.

Answer: In any negotiation, attention 52.Kf1

should be paid to that which is not said:
White’s king suffers as a cow entering
Black’s threat of ...Kg3! and ...Re1 mate
the slaughterhouse.
must be addressed. The only way to
52.Kh1 allows 52...Re1+ 53.Kh2 Ng4
achieve the draw is to move the king to
mate, and 52.Kf2 is easily refuted by
h2.
52...Ng4+ 53.Kf1 Nh2+ 54.Kg1 Kg3!
51.Kg2?? 55.e8=Q f2+.

This goof-up falls under the classic

 a satisfying conclusion. The threat is
8
r+*+*+*+ ...Rf1 mate.
7
+*+*P*+* 56.Rf8 Rc3 0-1
6
*+n+*Mo+
5
+*+*+*+* Pay close attention to the position after

*+*+*+*L
4 move 40. It’s an equal ending for
engines, but not so much for humans.
3 +*+*To+* The evaluation after 40...Rc2! is
2 *+*+*+*+ deceptive. The problem is that only
1 +*+*+k+* engines can hold such positions with
confidence and consistency. We humans
a b c d e f g h
botch them routinely, since we can’t
work out the exact details of the math
needed to get us to a draw. Larsen was
Black to move unable to cope, as he was dealing with
pawn weaknesses on g3 and e5, his
52...Kg3! second rank occupied by Kasparov’s
rook, and Kasparov’s king threatening to
There is no defense to the coming
walk up the board and assist in a mating
threat of ...Ng4 and ...Nh2+.
attack.
53.Nd4
In the following game, Black got
53.Rf8 Ng4 54.e8=Q Nh2+ 55.Kg1 ambushed in the opening, which led to
Rxe8! is another relevant variation. an ending where he was forced to nurse
several weak pawns. Hollywood Squares
53...Ng4 54.Nxf3 was a game show from my childhood,
where aging, wannabe pseudo-
There was no choice.
celebrities would appear after officially
54...Rxf3+ 55.Kg1 reaching the dreaded “has-been” stage
of their careers. I was nearing the end of
55.Ke1 Re3+ pops the queen-wannabe my playing career when this game was
on e7. played. My opponent, Kyron, is now an
IM with one GM norm. As usual, I got
55...Nh2! around the issue by playing him when
he was younger, just below IM strength.
The seemingly endless toing and
Game 9
froing of this knight is about to come to
 Cyrus Lakdawala – Kyron Griffith 5...a6
San Diego (rapid) 2013
Interestingly, 5...Na5 is Black’s main
1.b3 line.

What? Why is the London System guy 6.Bxc6

pushing the b-pawn?
The swap turns the game into a
1...e5 2.Bb2 Nc6 3.e3 Nf6 4.Bb5 Bd6 distorted version of the Spanish
Exchange Variation, with White’s e-
pawn on the third rank rather than the
8
t+vWl+*T fourth.
7
OoOo+oOo 6...dxc6 7.Nc4 Qe7 8.Nf3
6
*+mV*M*+
5
+b+*O*+* 8.Ne2 is White’s main line.
4 *+*+*+*+ 8...Bg4?
3 +p+*P*+*
2 pBpP*PpP Possible are both 8...Nd7 9.e4! and
8...e4 9.Nfe5.
1 Rn+qK*Nr
a b c d e f g h 8
t+*+l+*T
7
+oO*WoOo
6
o+oV*M*+
+*+*O*+*
White to move 5
Believe it or not, this bizarre, 4 *+n+*+v+
obstructionist move is all the rage
theoretically and is even more popular
3 +p+*Pn+*
than the conventional alternatives.
2 pBpP*PpP
1 R*+qK*+r
5.Na3 a b c d e f g h
As we can see, this isn’t our
grandfather’s opening repertoire. White
intends Nc4, adding heat to e5 and the White to move
d6-bishop.
Exercise: We tend to sleepwalk in
 the opening stage, since we don’t 10...Bxd1 11.Bxe7 Bxc2 12.Bxf6 gxf6
believe in the existence of a
combination that early in the game,
and our guard is down. White has
8
t+*+l+*T
access to one such combination. It 7
+oO*+o+o
doesn’t win material, but it inflicts
damage to Black’s pawn structure.
6
o+o+*O*+
What should White play?
5
+*+*+*+*
4 *+n+*+*+
Answer: Capture. 3 +p+*P*+*
9.Bxe5!
2 p+vP*PpP
1 R*+*K*+r
The shot comes with the suddenness a b c d e f g h
of bright lights being turned on in a dark
room. Black’s e-pawn is not as stable as
first believed.
White to move
Of the four database games from this
position, in two of them White missed Black is in big trouble, mainly since
the combination and settled for the his structure will be an unending source
inferior 9.h3? Bxf3 10.Qxf3, with only a of despair:
tiny edge. 1) The f7- and f6-pawns are doubled
and isolated.
9...Bxf3
2) Number one on the list
Kyron undermines a defender of my automatically means that Black also
now hanging e5-bishop. deals with weaknesses on the f4- and f5-
9...Bxe5 10.Ncxe5 leaves Black down squares.
a pawn for no compensation. 3) The mass exchanges of the previous
moves created the imbalance of bishop
10.Bxd6! versus knight. For now, Black’s bishop
looks good, but that’s until White
I prefer an ending with superior manages to stabilize and bring the
structure over winning a pawn, while knight to a good outpost – with such a
giving Black fishing chances in the line structure for Black, there are plenty of
10.Bxf6 Qxf6 11.gxf3 0-0-0. Sure,
those.
White is up a pawn, yet his king lacks a
4) White will win the fight for the hole
completely secure haven.
on d3 since he can play Rc1, Rc3, and
Nb2 should Black park his bishop on the 8
t+*+l+t+
square and castle queenside. 7
+oO*+o+o
5) White enjoys superior central
control, especially if he can play d2-d4
6
o+o+*O*+
before Black can play ...c6-c5.
5
+*+*+v+*
Conclusion: Black is in deep trouble in
4 *+n+*+*+
the endgame since, sooner or later, he 3 +p+*Pp+*
will need to confront his structural
demons.
2 p+*P*KpP
1 +*R*+*+r
13.Rc1 Bf5 a b c d e f g h

As mentioned in the list above,

13...Bd3?! is just a waste of time after
14.Nb2 0-0-0 15.Rc3 when the bishop is Black to move
ejected from the d3-hole.
15...Bd3!?
14.f3!
This looks like a waste of time from
This way the g2-pawn can be covered the bishop, who senses its irrelevance
with Kf2, should Black play ...Rg8. and desperately seeks a position of
Moreover, White controls some light authority.
squares, reducing the importance of the 15...c5 can be met with the surprising
unopposed black bishop. 16.d4!!. Other, slower ways to play are
also good, but this is just so cool. After
14...Rg8 15.Kf2 16...cxd4 17.exd4 0-0-0 18.Ne3 Bg6
19.d5! White dominates all areas of the
board and can claim a strategically
winning position.

16.Rc3!?

I’m happy to enter a double-rook

ending should he choose to exchange on
c4.
The immediate 16.Nb2! is a tad more
logical as a move order. Why deal with
the capture on c4 if I can simply avoid Upon looking through the game with
it? an engine, one stumbles upon: 17...Bf1!!
(where did this come from?) 18.Kxf1
16...0-0-0 Rxd2 19.Nd3 Rd8 20.Ne1 Rxa2 when
Black has enough compensation for the
16...Bxc4 is a better positional
sacrificed piece. Wild stuff.
decision than keeping the bishop, but the
rook endgame is also terrible for Black. 18.d4
Despite that, the text move comes with a
devilish trap. The only problem is that White is back in control, nursing a
nobody had noticed the trap existed until decisive positional advantage.
Stockfish put the position under
scrutiny! 18...Be6 19.Nd3 Rg5 20.Rhc1 Rdg8
21.Nf4
17.Nb2?!

Trying to oust the bishop from d3 and 8

*+l+*+t+
keep the superior knight. It’s funny, but 7
+oO*+o+o
o+o+vO*+
this almost automatic move misses an 6
important tactical detail.
5
+*+*+*T*
8
*+lT*+t+ 4 *+*P*N*+
7
+oO*+o+o 3 +pR*Pp+*
6
o+o+*O*+ 2 p+*+*KpP
5
+*+*+*+* 1 +*R*+*+*
4 *+*+*+*+ a b c d e f g h
3 +pRvPp+*
2 pN*P*KpP
1 +*+*+*+r Black to move
a b c d e f g h Black experiences huge trouble, with
no solution in sight:
1) White’s knight has clearly emerged
Black to move as the superior minor piece, since now it
occupies a hole on f4.
17...Bf5 2) Black’s rooks have nothing to
attack. Black to move
3) Black has no pawn breaks to disrupt
24...f5
White’s stable structure.
4) Black’s kingside pawns are all Black can’t allow Rh5. The problem is
tender and potential targets to the plan now Black’s bishop – once a loved
of Rc5 followed by h2-h4, which seizes child’s toy, now fallen out of favor and
control over the fifth rank. When this gathering dust in a state of neglect – is
happens, White’s rook can shift over to made even worse. Six of Black’s seven
h5, harassing Black’s kingside. pawns are on the same color as his
bishop. Also, this pawn push further
21...Rd8
weakens the kingside dark squares.
Alternatively, after 21...f5 22.Rc5 Kd7
25.Nd3?!
23.b4 Kd6 24.a4 R5g7 25.a5 White
continues to make progress, while Black The human move. The engine’s
can only sit and wait. suggestion is shocking: 25.Nxe6!!.
Undoubling Black’s pawns while
22.Rc5! Rd6 23.h4!
swapping a good knight for a bad bishop
Wrestling Black’s rook away from the – this is, psychologically, virtually
fifth rank. impossible for a human, since it appears
to be giving away major concessions
23...Rxc5 24.Rxc5 without gaining anything concrete in
return. However, with 25...fxe6 26.g4!
the all-knowing engine shows us the
8
*+l+*+*+ way. Black used to have a bad bishop
7
+oO*+o+o and a ruined structure. Now he is left
6
o+oTvO*+ with a bad rook and king, but is unable
to catch our outside passed pawn. The
5
+*R*+*+* disappointing part is that it takes only
4 *+*P*N*P one more move from the engine, and we
3 +p+*Pp+* can understand its point. We can

p+*+*Kp+
2 understand both the logic behind it and
how completely winning White’s
1 +*+*+*+* position turns out to be. But pushing
a b c d e f g h ourselves to look for this and even
consider taking on e6 is a completely
different story. An illustrative line could
be: 26...fxg4 27.fxg4 Rd7 28.Kg3 Rf7 ending Black’s dreams of one day
29.g5 Kd7 30.Kg4!. The white king will freeing himself with ...b7-b6 and ...c6-
go to h6, and the passed pawn about to c5.
be created will cost Black a rook.
29...Kd8
25...Kd7

8
*+vL*+*+
8
*+*+*+*+ 7
+oO*+*+o
7
+oOl+o+o 6
o+oT*O*+
6
o+oTv+*+ 5
+*+*+o+*
5
+*R*+o+* 4 *P*P*K*P
4 *+*P*+*P 3 +*RnPp+*
3 +p+nPp+* 2 p+*+*+p+
2 p+*+*Kp+ 1 +*+*+*+*
1 +*+*+*+* a b c d e f g h
a b c d e f g h

White to move
White to move
30.Nc1!
Exercise: Come up with a plan for
White to improve his position. White transfers the knight to g3, after
which Black gets twisted into a pretzel
while trying to defend the weak f5-
Answer: Black’s kingside dark
pawn.
squares are punctured, and White’s king
can travel up the board, via g3 and f4. 30...Be6 31.a3 Kd7 32.Ne2 Ke7

26.Kg3! After 32...b6 33.Ng3 Rd5 34.Nh5

Black is unable to defend f6, since
Principle: Use your king as a fighting
moving the king to e7 hangs the c6-
piece in the endgame.
pawn.
26...Ke7 27.Kf4 f6 28.Rc3 Bc8 29.b4
33.Ng3
White clamps down on c5, forever
 Property values begin to rise in thing. Yet it’s not the same as Da Vinci
White’s neighborhood. There is no refusing to paint, since Magnus still
saving the f5-pawn. plays! I watched the following game
live, and I can testify that it is one of the
33...b6 most amazing games I have ever
witnessed from a strategic standpoint.
After 33...Rd5 34.Rc5 the f5-pawn will
It’s full of geometric anomalies and
fall all the same.
contradictions, yet the math always
34.Nxf5+ Bxf5 35.Kxf5 Rd5+ 36.Kf4 seems to work out for White’s side. My
Kd7 37.g4 h6 38.Ke4 a5 39.Kd3 eyes bulged like a bullfrog’s when
Carlsen pulled off his stunning exchange
White is ready for e3-e4. sacrifice. Watch for it.
Game 10
39...axb4 40.axb4 1-0
Magnus Carlsen – Anish Giri
Wijk aan Zee 2022
Not all chess games end with
spectacular finishes. Sometimes we win
1.d4 Nf6 2.c4 e6 3.Nf3 d5 4.g3 Be7
in unnewsworthy ways. The rook ending
5.Bg2 0-0 6.0-0 dxc4 7.Na3!?
is dead lost for Black, since he is down a
pawn and saddled with pawn targets on Far safer are the lines 7.Qc2, 7.Ne5
f6 and h6. and 7.Qa4.
The game we just saw is a good
example of how a damaged structure, 7...Bxa3 8.bxa3
especially when the position is generally
stable, can single-handedly decide the Let’s gather a bit of data to orient
game. If we disregard the shocking ourselves with the position’s new
17...Bf1!!, the sad truth is that Black, reality:
after allowing the doubling of his f-
pawns, essentially stood no chance. A
factor that further enhanced White’s
advantage was the stable, good knight,
against the opponent’s bishop, which
served no real purpose other than merely
defending Black’s sickly pawns.

In 2023, Carlsen abdicated his world

championship title, which was a surreal
 8...Bd7 9.a4
8
tMvW*Tl+
OoO*+oOo
7 Black is denied any ...b7-b5 ideas, and
the a4-square is vacated for the bishop.
6
*+*+oM*+ 9.Ne5 Bc6 10.Nxc6 Nxc6 is also
5
+*+*+*+* considered dynamically equal.
4 *+oP*+*+ 9...Bc6 10.Ba3 Re8 11.Qc2 Nbd7
3 P*+*+nP*
2 p+*+pPbP
1 R*Bq+rK*
8
t+*Wt+l+
a b c d e f g h
7
OoOm+oOo
6
*+v+oM*+
5
+*+*+*+*
Black to move 4 p+oP*+*+
1) Black inflicted some damage to
3 B*+*+nP*
White’s structure by handing him
2 p+q+pPbP
doubled and isolated a-pawns. 1 R*+*+rK*
2) The a-pawns for now are not all that a b c d e f g h
weak, since Black has no easy way to
attack them.
3) White picked up the bishop pair,
White to move
whose value rises in the semi-open
position. 12.Rac1!
4) The b-file opened for White’s rook
to apply potential pressure down it. The immediate 12.Qxc4?! is less
5) Black, for now, remains up a pawn accurate and doesn’t win back the pawn
on c4, which won’t be easy to hold on since Black has 12...Nb6, with a double
to, but will also require some effort by attack on White’s queen and the a4-
White to corral and take it. pawn.
Conclusion: The damage to White’s 12...a6 13.Qxc4
pawn structure didn’t come for free, and
the engine’s assessment is dynamic The inclusion of the rook move to c1
equality. definitely helps White, while it’s not as
clear what the pawn does on a6 when
there is no pawn to support it on c4. full exchange to inflict damage on
Black’s queenside structure. The
13...Nb6 14.Qc3 Nxa4 15.Qb3 Qd5 engine’s assessment of the sacrifice is
balanced, while for a human it’s far
Giri uncorks a theoretical novelty,
more difficult to play Black, who must
which is met with a shocking response.
solve all kinds of practical problems.
Previously played was 15...Bb5 From this point on, strategic parameters
16.Rfe1 Rc8 17.e4 c6 in Ivanisevic- begin to shift in bewildering patterns,
Cvitan, Terme Catez 2022. Black has and the game becomes a pool with no
managed to stabilize his control of the shallow end.
light squares and keep an extra pawn,
but it has come at the cost of a terrible 16...Qxc6
weakening on the dark squares. After
18.e5! Nd5 19.Nd2! I like White’s The only move. 16...bxc6?? loses to
chances on the dark squares, since Ne4 17.Qxa4 and 16...Qxb3?? 17.axb3 bxc6
and Nd6 are coming. 18.bxa4 is similar. White picks up two
pieces for the rook, with a winning
8
t+*+t+l+ structure to boot.
7
+oO*+oOo 17.Ne5 Qb5 18.Qc2!
6
o+v+oM*+
+*+w+*+*
5 Quiet moves like this one are always
the hardest ones. Rb1 is coming, and
4 m+*P*+*+ Black’s queenside, the knight on a4 and
3 Bq+*+nP* queen are all under siege.
2 p+*+pPbP 18...Nd5?
1 +*R*+rK*
a b c d e f g h The position required engine-like
accuracy.
Retreating the endangered knight with
18...Nb6! was the only way to maintain
White to move the balance. After 19.Bxb7, threatening
16.Rxc6!! to take Black’s a8-rook, while also
menacing a skewer with Bc6, Black
What the...? Carlsen really should must find:
consider pausing to chew a few CBD oil
gummies to calm down. White offers a
 8
t+*+t+l+ 8
t+*+t+l+
7
+bO*+oOo 7
+oO*+oOo
6
oM*+oM*+ 6
o+*+o+*+
5
+w+*N*+* 5
W*+mN*+*
4 *+*P*+*+ 4 m+*P*+*+
3 B*+*+*P* 3 B*+*+*P*
2 p+q+pP*P 2 p+q+pPbP
1 +*+*+rK* 1 +r+*+*K*
a b c d e f g h a b c d e f g h

Black to move White to move

19...Nc4!, throwing more wood on to Now Carlsen reveals another shocker:

the fire. Using the unprotected nature of
the b7-bishop, Black manages to force 20.Bxd5!
some exchanges. After the forced
“What the...?” Volume 2. I suppose
20.Bxa8 Nxa3 White can either go for
that if Keith Richards walks into a bar,
the safer 21.Qc6 or for the riskier
we shouldn’t expect him to order a glass
21.Qxc7, but, in both cases, the
of milk. White gives up his powerful
equilibrium is not disturbed.
light-squared bishop for the knight on
19.Rb1 Qa5 d5. This is an incredible idea, claiming
that the queen cannot take back on d5,
since she is tethered to the knight on a4.
Taking with the pawn is a positional
disaster as, among other things, the
white queen is allowed access to the f5-
square.
However brilliant this may be, it was
possible to improve upon the execution,
and even stronger was to probe for
weaknesses with a zwischenzug:
20.Be4!! g6 (instead, 20...h6? allows
21.Bh7+ Kh8 22.Nxf7 mate; 20...Nf6 particular are especially instructive. One
isn’t as terrible as checkmate in two, but pawn move and the weaknesses it
after 21.Bxb7 Black is in terrible shape creates can radically alter the evaluation
anyway) 21.Bxd5!!. The same idea is of a position. Furthermore, this
even stronger, after provoking dark- technique of making a pit stop to annoy
square weakness around the black king. your opponent on the way to the final
To illustrate: 21...Nc3 (21...exd5 destination – in this case stopping on e4
22.Rxb7 is terrible for Black no matter when the destination was actually d5 –
where the g-pawn is, but it’s still even is a valuable technique to add to your
worse with the pawn on g6 instead of repertoire.
g7) 22.Bb4 Nxe2+ 23.Qxe2 Qxd5.
20...exd5?
8
t+*+t+l+ Choose wisely when you make your
7
+oO*+o+o decision. Philosophers advise us to be
6
o+*+o+o+ content with what we have now.
5
+*+wN*+* However, this isn’t how human nature
works, since we always desire more. In
4 *B*P*+*+ strategically inferior positions, it’s
3 +*+*+*P* natural to want to hang on to our extra
2 p+*+qP*P material. Not this time, though.
1 +r+*+*K* Black had to enter concession mode
and return material with: 20...Nc3!
a b c d e f g h
21.Bb4 Nxe2+ 22.Qxe2 Qxd5 23.Rc1!
(forcing some weaknesses on the dark
squares) 23...c6 24.Bc5. White has a
White to move huge advantage but isn’t winning just
yet. Imagine if we could magically
With the pawn on g7, this position is move the pawn from g7 to g6...
playable for Black. However, pawns do
not move backwards, and weaknesses 21.Rxb7
have the annoying tendency to not
protect themselves. After 24.Ng4! Black Threatening Bb4.
is busted: for example, 24...Kg7 25.Rc1
21...c5
c6 26.Rc5 Qd8 27.Bd2 and everything
joins the attack. The weaknesses on the This hangs on to the knight, but it
dark squares will decide the game. I comes at the cost of the black king!
believe that this line and 20.Be4! in
 Also lost for Black is 21...Qe1+ attack, bringing the queen back to
22.Kg2 Nc3 23.Nf3!? Qxe2 24.Qxc3. defend her king is often a good idea.
The two pieces are wildly superior when 23...Qe1+ 24.Kg2 Qxe2 loses to
compared to a black rook. 25.Qxd5!, threatening a smothered mate
with the typical mechanism Nh6+,
8
t+*+t+l+ Qg8+ and Nf7 mate. Black’s best bet is
7
+r+*+oOo 25...h6, but that’s a little depressing.
6
o+*+*+*+ After 26.Nxh6+ Kh7 27.Nf7 Nc3
Black won’t have time to save himself
5
W*OoN*+* with a queen check on e4 and 28.Qf5+
4 m+*P*+*+ Kg8 29.Bxc5 is all over.
3 B*+*+*P* 24.dxc5
2 p+q+pP*P
1 +*+*+*K* A new problem arises for Black:
a b c d e f g h White has gained a dangerous passed c-
pawn, which threatens to move up the
board with gain of tempo.

White to move 24...Qf6

22.Qf5! What else? Giri forces the queens off

the board, but it’s not enough to save
Suddenly, Black’s puny and him.
undefended king is swarmed by three
powerful white attackers. The f7-pawn 25.Qxf6 gxf6 26.Nh6+ Kh8 27.c6 Rfc8
can’t be protected, which means that 28.c7
Black’s king is doomed.

22...Rf8

22...f6?? doesn’t cut it, since White

forces mate after the simple doubling on
the seventh rank with 23.Qd7.

23.Nxf7 Qd8!

Giri finds the only move to not resign

on the spot. When facing an extreme
 technical win.
8
t+t+*+*L 33...Rxc8 34.a4! Nxc7 35.Bxd4 Ne6
7
+rP*+*+o 36.Be3 1-0
6
o+*+*O*N
5
+*+o+*+* Nothing grows for Black in this

m+*+*+*+
4 irradiated, post-apocalyptic wasteland.
White is two pawns up with a
3 B*+*+*P* dominating position, and Giri saw no
2 p+*+pP*P reason to continue.
1 +*+*+*K* There is equal practically and equal
theoretically. Sometimes those two
a b c d e f g h
concepts overlap, but they can also be in
complete disagreement. The engine says
the position is even after Carlsen’s
Black to move bewildering exchange sacrifice with
16.Rxc6!!. Sure, it’s even if Giri had
Black’s position reaches the slowly yet found around 10 perfect defensive
surely deteriorating stage. The passed c- moves in a row, which almost no human
pawn chokes the life out of Black’s could ever pull off. Essentially, the
position. White threatens Nf7+, position is lost for a human.
followed by Nd6, regaining the
Honestly, I had no idea which chapter
exchange with an easy win.
this game belonged to, since it can fit
28...Nc3 29.Bb2 d4 into virtually every chapter in the book –
it could be described in various different
Moving the knight opens up the long ways. Carlsen seamlessly switched his
diagonal for the b2-bishop who’s always attention from one wing to the other,
dreaming of Bxf6 mate. making gains on the queenside, striking
in the center, and eventually crashing
30.Nf7+ Kg7 31.Nd6 Kg6 through on the kingside, paying homage
to the principle of two (or more)
Equally hopeless is 31...Nxe2+
weaknesses.
32.Kf1 Nc3 33.Bxc3! dxc3 34.Ke2!
As far as weak pawns are concerned, I
when Black remains paralyzed.
find it quite instructive how Carlsen
32.Kf1 Nb5 33.Nxc8 allowed his structure to be completely
ruined (he had a pair of doubled,
At last, White cashes out into an easy isolated a-pawns, in addition to being a
pawn down!), but in exchange he got
plenty of dynamic counterplay. In the 8
t+vWt+l+
end, it turned out to be more than 7
OoO*+oOo
*+m+*+*+
enough compensation to counterbalance 6
his structural damage. This is, I think, in
large part, the most instructive aspect of
5
+*VmO*+*
this game, and that is why I decided to 4 *+*+*+*+
include it in this chapter. However, one
could also interpret the situation after
3 +*Np+nP*
move 14 as White playing against a
2 pP*+pPbP
weak black piece, namely the knight on 1 R*Bq+rK*
a4. The poor beast hanging there a b c d e f g h
dragged the black queen out of play,
which is the main theme behind most of
White’s blows: 16.Rxc6!! and 20.Bxd5!
White to move
are both based on the hanging knight.
More of that type of play will be 9.Ng5!?
presented in Chapter 3.
Understandably, most of us hate it Caruana is looking to pick a fight.
when the engine challenges our 9.Nxd5 is White’s main option.
expertise. Throughout the following
game, the engine continually tells us not 9...Nf6
to trust our lying eyes. White’s ugly
Exchanging on c3 is possible, yet it
tripled pawns look much weaker than
strengthens White’s center and opens
they actually are.
the b-file.
Game 11
Fabiano Caruana – Michael 10.Qb3
Adams
London 2017 Attacking f7.

1.c4 Nf6 2.Nc3 e5 3.Nf3 Nc6 4.g3 d5 10...Qe7 11.Nd5?!

5.cxd5 Nxd5 6.Bg2 Bc5 7.0-0 0-0 8.d3
This artificial move, which loses time,
A reversed Dragon Sicilian. should actually lead to an edge for
Black.
8...Re8 Like it or not, 11.Bxc6!? is White’s
most principled move in the position.
White parts with the important light- 13...Bd4!
squared bishop, weakening the safety of
his king. However, ...Nd4 is eliminated Putting the bishop on an excellent
as a problematic dynamic option, and central square exploits how e2-e3
Black’s queenside structure is blunders the knight on g5. White’s
completely ruined. The weaknesses on ephemeral initiative has hit a brick wall,
both sides should balance out. The and now everyone must retreat. This is
engine also likes 11.Qc4!?. what happens to White when
overzealously trying for an attack that
11...Nxd5 12.Bxd5 Nd8?! didn’t have the positional grounds to be
developed, and when pieces try to
Seemingly forced, but appearances on advance but do so by stepping on
the chess board are known to be squares that can easily be attacked by
deceptive. An improvement was found a the opponent’s pawns.
couple of years later with 12...Nd4! in
Kahl-Muniz Pardino, correspondence 14.Bg2 h6 15.Nf3 Nc6 16.Be3!?
2019.
This is exceedingly risky since White
13.Qc4 will soon end up with tripled e-pawns!
Understandably, most players would
Threatening to overload Black’s queen have gone for 16.Nxd4, gaining the
with tricks on f7. bishop pair and retaining White’s
structural integrity. However, after
8
t+vMt+l+ 16...Nxd4 17.Re1 Be6 18.Qc3 Bg4 it
7
OoO*WoOo becomes clear that Black’s activity
6
*+*+*+*+ easily balances out White’s long-term
positional advantages.
5
+*VbO*N*
4 *+q+*+*+ 16...Bxe3
3 +*+p+*P* 16...Be6! and playing for dynamism is
2 pP*+pP*P better, but who can blame Adams for
1 R*B*+rK* wanting to give his opponent tripled
a b c d e f g h isolated pawns?

17.fxe3 e4!

Black to move This is a completely thematic pawn

sacrifice.
18.dxe4 falls into White’s trap after 19.Qxf7+!
Kxf7 20.Ng5+ Kg8 21.Nxe4, with a
clear advantage for White, who picked
8
t+v+t+l+ up a pawn. The fact that White’s e-
7
OoO*WoO* pawns remain doubled isn’t full
6
*+m+*+*O compensation for Black, who is behind
in development and about to come under
5
+*+*+*+* serious pressure on the c-file and the
4 *+q+p+*+ queenside in general.
3 +*+*PnP* 19.a3 Ra6! 20.Rac1 Rb6 21.Rc2 Be6
2 pP*+p+bP
1 R*+*+rK* The ogress is ejected from her post on
a b c d e f g h c4.

22.Qc3 Rb3

Black to move The queen is chased even further back,

the rook is active, the b-pawn is
If I were a mob boss, I would be blocked, and the e3-pawn is under
tempted to comment to Caruana: “Nice pressure. Adams is handling the
pawns you got there on the e-file. It situation excellently.
would be a shame if something
unfortunate were to happen to them.” 23.Qd2 Rd8 24.Qc1 a4 25.Rc5
1) White is saddled with tripled pawns

*+*T*+l+
on the e-file. His structure reminds us of 8
the once pristine bay after the Exxon
Valdez ran aground and spilled 11
7
+oO*WoO*
million gallons of oil into it. 6
*+m+v+*O
2) Black is in control over the hole on 5
+*R*+*+*
e5. If this is the case, why on earth is the
engine’s assessment ‘0.00’?
4 o+*+p+*+
3 Pt+*PnP*
18...a5! 2 *P*+p+bP
Adams finds a creative way to develop
1 +*Q*+rK*
his queen’s rook. a b c d e f g h
The tempting yet incorrect 18...Qxe4?!
 Black to move of playing this slowly.

People who believe White is OK must 26...Qd8 27.g4!?

also believe that the Apollo 11 moon
landing was faked. The only difference Now Caruana starts placing his pawns
between the two cases is that people on the same color as his remaining
thinking White is fine have it right. Our bishop – it’s as if the normal principles
eyes tell us that Black stands better, yet don’t apply to him! In reality, e4-e5 is
this assessment is contradicted by the unstoppable, so the g2-bishop only
engine, which claims that the game needs the h1-a8 diagonal. Advancing the
remains in a dynamic balance. Let’s try g-pawn gains space, hints at a future g4-
to discover just why the engine believes g5, and takes away the f5-square from
something so outrageous: the black bishop.
1) Yes, White’s eyesore e-pawns are
27...g6
indeed potential weaknesses.
2) But White’s e-pawns also control This way White’s pieces – especially
multiple key central squares like f5, d5, his knight – are discouraged from
d4, and d3. moving to f5.
3) White can use the pawns to put his
pieces on strong and stable central 28.Kh1
squares: for example, to get a knight on
He’s in absolutely no hurry.
d4.
28...Kg7
25...Rd7 26.h3!?

An ambitious move, preparing a

further gain of space on the kingside by
pushing the g-pawn.
26.e5 would look more natural to most
people, clearing the long diagonal and
rushing to put the knight on d4.
However, as soon as the pawn goes to
e5, it becomes more vulnerable, and
Black gains access to some light
squares, making the position more
concrete. Caruana’s move in the game
shows that he’s actually not afraid at all
 8
*+*W*+*+ 8
*+*+*+*+
7
+oOt+oL* 7
+oO*WoL*
6
*+m+v+oO 6
*+m+v+oO
5
+*R*+*+* 5
+*RtP*+*
4 o+*+p+p+ 4 o+*+*+p+
3 Pt+*Pn+p 3 Pt+*Pn+p
2 *P*+p+b+ 2 *P*+p+b+
1 +*Q*+r+k 1 +*Q*+rK*
a b c d e f g h a b c d e f g h

White to move White to move

29.e5 33.Rc4!?

At last. The definition of a “do not try this at

home” moment. Caruana is taking
29...Bd5 30.Kg1 Be6 immense risks to keep his winning
chances alive. The rook steps on to the
This is a tacit draw offer.
line of vision of the e6-bishop, taunting
31.Kf2!? him with the claim that there is no good
discovery.
Nope! White is playing for the full I think anyone in their right mind
point. His king adds protection to e3, would have budged and gone for
and the position remains in a dynamic 33.Rxd5 Bxd5 followed by 34.Nd4 or
balance. 34.Ng5!?.

31...Qe7 32.Kg1 Rd5 33...Ra5

Safer might have been 33...Nxe5!?

34.Rxc7 Nxf3+ 35.Bxf3 Rd7, with
approximate equality.

34.Rc2 Bd5?!
 The engine says this move is OK, And the remaining doubled pawns are
while the human writer doesn’t like it, undoubled. Who would have thought!
since White soon makes his extra pawn Radical alterations have taken place:
permanent. 1) As if by a faith-healer, White’s
former tripled pawns have been made
35.Nd4! whole again.
By hitting Black on a few different 2) Black’s g3-rook looks lost,
fronts (b3, d5, c6, and introducing e5-e6 entombed, and in grave danger.
as a threat), White manages to untriple The engine says it’s even, but only
his e-pawns. because it manages to hold in
characteristic, pitch-perfect, concrete
35...Nxd4 36.exd4 Rg3 fashion. Any human would obviously
prefer White.
8
*+*+*+*+ 38...c6 39.Kh2
7
+oO*WoL* Black must hand back the exchange
6
*+*+*+oO and be down a pawn.
5
T*+vP*+*
4 o+*P*+p+ 39...Rxg2+
3 P*+*+*Tp 39...Qh4 is met with 40.Qf4, after
2 *Pr+p+b+ which Black must return the exchange

+*Q*+rK*
1 all the same.

a b c d e f g h 40.Kxg2

White to move

37.Rf3!

This exchange sacrifice is far stronger

than 37.Rf2, passively holding on to the
bishop.

37...Bxf3 38.exf3
 8
*+*+*+*+ 8
*+*+*+*+
7
+o+*WoL* 7
+o+*WoL*
6
*+o+*+oO 6
*+*+*+oO
5
T*+*P*+* 5
+*R*P*+*
4 o+*P*+p+ 4 o+*T*+p+
3 P*+*+p+p 3 P*Q*+p+p
2 *Pr+*+k+ 2 *P*+*+k+
1 +*Q*+*+* 1 +*+*+*+*
a b c d e f g h a b c d e f g h

Black to move Black to move

Despite being down a pawn, there is 43...Qd8?

still hope for a draw, since the d4-pawn
is a target and White’s king is exposed Black overestimates his attacking
and vulnerable to a potential perpetual chances.
check. He still had decent chances to hold the
game if he switched to a queen ending
40...Rd5 with 43...Rd3! 44.Qxd3 Qxc5 45.Qc3
Qb6! 46.e6+ f6 47.Qb4 Qxe6 48.Qxb7+
40...Rb5! and putting the rook on b3 is Kf8. White’s king is exposed and
a bit more accurate. vulnerable to perpetual check, and the
game should end in a draw. This line is
41.Rc4 c5!
highly instructive, as it focuses Black’s
Adams wants to open lines to get at counterplay on White’s sole weakness:
the exposed white king. his exposed king. In that regard, trading
rooks is in Black’s favor, as it makes it
42.Rxc5 Rxd4 43.Qc3 even harder for White to stop Black’s
attempts at delivering a perpetual check.
However, it is tough to play like this.
Trading pieces as the defender is
counterintuitive, and let’s be honest:
who would look at allowing e5-e6+ and
think “this looks fine”? 47.e6!

44.Rc8 Caruana finally finds a good moment

to include this, destroying the shelter of
Even stronger is 44.e6! fxe6 45.Rc7+ the black king.
when Black’s king is in serious danger. 47.Rxb7?! is met with 47...Rxg4+!
48.hxg4 Qxb7, after which White will
44...Qb6 45.Re8?!
have a tough time converting, despite
45.Kg3!! is stronger but inhuman, as his extra pawn. This is also probably
White needs to calculate all sorts of rook why Adams put his king on g8, so that
discoveries followed by ...Qg1+. there is no Qb4+, forcing a queen trade.

45...g5 46.Re7 Kg8 47...fxe6 48.Qc2! Kf8 49.Rh7 Qc6

46...Kf8! 47.Rxb7! wins a second Black is willing to do anything to get

pawn, but, funnily enough, after the queens off the board.
47...Qxb7! (47...Rxg4+ 48.hxg4 Qxb7
loses to 49.Qb4+) 48.Qxd4 Kg8! the 8
*+*+*L*+
engine says Black should hold, despite 7
+o+*+*+r
*+w+o+*O
being down two pawns. Anyway, I think 6
that’s a bit irrelevant, and humans would
lose with Black quite consistently.
5
+*+*+*O*
4 o+*T*+p+
8
*+*+*+l+ 3 P*+*+p+p
7
+o+*Ro+* 2 *Pq+*+k+
6
*W*+*+*O 1 +*+*+*+*
5
+*+*P*O* a b c d e f g h
4 o+*T*+p+
3 P*Q*+p+p
2 *P*+*+k+ White to move
1 +*+*+*+* 50.Qxc6?
a b c d e f g h
White will indeed have some chances
in the rook endgame, but swapping the
queens is a gross violation of the
White to move
Principle: The side with the safer king 55.Rd4! when the pin on the rook is
should avoid queen trades. unbreakable.
White should continue in attack mode
and play: 50.Rh8+! Ke7 51.Qh7+ Kd6 50...bxc6 51.Rxh6 Kf7 52.Kg3 Rd2
52.Rd8+ Kc5 53.Qc2+ Rc4. 53.Rh7+ Kf6 54.Rb7 Ke5! 55.h4 gxh4+
56.Kxh4 Kf4!
8
*+*R*+*+ Adams is using his king to the best of
7
+o+*+*+* his abilities. He’s already threatening
6
*+w+o+*O ...Rh2 mate, as well as ...Kxf3. Black

+*L*+*O*
5 should be holding the draw.

4 o+t+*+p+ 57.Rf7+
3 P*+*+p+p
2 *Pq+*+k+ 8
*+*+*+*+
1 +*+*+*+* 7
+*+*+r+*
a b c d e f g h 6
*+o+o+*+
5
+*+*+*+*
4 o+*+*LpK
P*+*+p+*
White to move 3
When the best players in the world 2 *P*T*+*+
make mistakes, what is most instructive
for us is to try and understand why they
1 +*+*+*+*
a b c d e f g h
made them, instead of mindlessly
criticizing them. Caruana must have
calculated up to this point and thought
that the black king gets away with Black to move
...Kb5 next, after which Black’s army is
surprisingly well-coordinated and ready Exercise: This is one of those
to start its own attack. However, the miserably difficult decisions.
cruel engine finds 54.Qd3!! when Black Should Black’s king move to e5 or
is almost in zugzwang and unable to e3?
move the king away from danger. The
main problems are that 54...Kb6 runs Answer: Moving to e5 holds the draw
into 55.Rd6 and 54...Kb5 loses to while moving to e3 loses.
57...Ke3? White’s tripled e-pawns were indeed
weak, yet they also covered key central
Now Black’s position slumps over like squares. Moreover, the magic trick that
a question mark. This is the wrong took place between moves 35 and 38
direction for the king, who needed to isn’t as rare as it may appear at first.
protect the pawn on e6. On e3, the king Damaged structures are blessed with a
also obstructs the black rook’s mobility hidden and rarely addressed advantage:
along the third rank. their restoration can be used as a threat!
57...Ke5! just about holds: for One could even go as far as to say that
example, 58.f4+ Ke4 59.g5 Rxb2 60.g6 doubled (or tripled) pawns create
Rg2 61.Kh5 Rg3 62.Kh6 Rxa3 63.g7 stronger outposts than healthy
Rh3+ 64.Kg5 Rg3+ 65.Kf6 Kxf4 structures, as taking the piece sitting on
66.Kxe6+ Ke3 with a draw. the outpost would repair the structure.
Due to the same reason, all different
58.Kg3 types of sacrifices could be considered
on such squares as, for example in this
Caissa is an angry goddess – Black is
game, the exchange sacrifice with
dead lost.
38.Rf3!.
58...Rd1
Conclusion
After 58...Rxb2 59.Re7 c5 60.Rxe6+
Unlike what you might have expected,
Kd3 61.Rd6+ Kc3 62.f4 White is faster
this chapter only contained one game
in the promotion race.
that illustrated the problems of having a
broken structure. That’s not a mistake in
59.g5
the game selection; it was my clear
Passed pawns should be pushed! intention. As shown in my game against
Kyron Griffith, yes, a better structure
59...Rg1+ 60.Kh4 Rg2 61.Rf6 e5 62.g6 can yield in itself a full point, especially
Rxb2 63.Kg5 if the situation remains rather stable.
Piling up on those weaknesses while
White is faster! restricting the opponent’s counterplay
can lead us to victory, without even
63...Rg2+ 64.Kh6 Rh2+ 65.Kg7 c5
breaking sweat.
66.Kf7 c4 67.g7 Rh7 68.Ra6 1-0
However, surviving in the modern era
This game teaches us to be careful not requires a different scale of appreciation
to be deceived by outward appearances. of the game’s dynamics, and that’s why
the rest of the games focused on
situations where having a broken
structure did not play a pivotal role.
That’s because, unlike my game against
Kyron, most positions aren’t static in
Chapter Three
nature, and even in that game, I missed A Vulnerable or
the equalizing 17...Bf1!!. Dysfunctional Piece
Handling dynamics correctly has the
power to override and erase permanent The creature from Stephen King’s It
positional defects, and we should learn
reads your mind and heart. Then It
to view structural weaknesses through
assumed the shape of your deepest fear.
that prism: if enough dynamic This is exactly how it feels to own a
compensation is available, then ruining
dysfunctional, minor piece against the
the structure shouldn’t be a deterrent.
opponent’s strong one. In this chapter,
I feel the need to issue a warning we will highlight the battle between a
before moving forward though: correctly powerful and an inefficient or
assessing the consequences of double- dysfunctional piece, showcasing how a
edged decisions is a high-level process, bad piece could also be considered a
and even the best among us will weakness.
sometimes fail in delivering the correct It is important to expand our
verdict. I’m just saying we should perception of a weakness as it could
always keep an open mind, just like refer to many more things than a pawn
Caruana did when he correctly identified or a square. Weaknesses should not be
his tripled isolated e-pawns as being viewed as pertaining only to static, long-
better than most of us would have term factors. Short-term, dynamic
thought. deficiencies in one’s position can also be
regarded as weaknesses.
As I’ve said a few times throughout
this book and its two companion
volumes, my great secret to beating kids
is to play them when they are still young
and then retire before they get older and
much stronger. In the following game
against one of my students, who today
holds a USCF rating of 2300, we arrived
at the following position.
Game 12
 Cyrus Lakdawala – Brandon Xia fortress. This means preventing White’s
San Diego (rapid) 2016 only pawn break on b5 by playing
You might claim this case is extreme, ...Be8.
and you’d probably be right. But before
29.Kf4 Be8
dealing with more nuanced examples, I
want to show a truly clear-cut case of a Black won’t survive if White achieves
minor piece being clearly superior to its a successful break on b5.
counterpart.
30.Nd3! Kf8 31.Ke5

8
t+*T*+l+
7
+o+*+*Oo
8
t+*TvL*+
6
o+o+o+v+
7
+o+*+*Oo
5
+*PoNoP*
6
o+o+o+*+
4 pP*P*+*+
5
+*PoKoP*
3 R*+*PpKp
4 pP*P*+*+
2 *+*+r+*+
3 R*+nPp+p
1 +*+*+*+*
2 *+*+r+*+
a b c d e f g h
1 +*+*+*+*
a b c d e f g h

White to move
Black to move
1) Black’s bishop knows that all are
equal under God, yet it feels much less Wilhelm Steinitz stated the Principle:
equal than the knight. In the endgame, the king is a fighting
2) White’s king has chances to piece, so don’t hide it away. Use it! I
infiltrate the central dark squares via e5, just need to be careful not to walk into
just as long as he doesn’t fall for some any silly future mating nets based on
cheap mating net. In that regard, White’s ...f5-f4 and ...Rf5 mate.
plan could be Nd3, followed by Kf4,
31...Ke7 32.Nf4 Bf7?!
Ke5, and Nf4, adding pressure to the
e6-pawn. The wrong diagonal.
3) Black’s only plan is to try to build a The bishop should be posted on d7 to
suppress White’s b5-break, i.e. 32...Bd7.
8
*+*+*+t+
8
t+*T*+*+ 7
To+vL*+o
7
+o+vL*Oo 6
o+o+o+oP
6
o+o+o+*+ 5
+*PoKoP*
5
+*PoKoP* 4 pP*P*N*+
4 pP*P*N*+ 3 +*+*Pp+*
3 R*+*Pp+p 2 *+*+*+*R
2 *+*+r+*+ 1 +*+*+*R*
1 +*+*+*+* a b c d e f g h
a b c d e f g h

White to move
White to move
With the pawn reaching h6, Black
To some, it may appear as if Black has needs to take care not to allow either
sealed the game and will draw. This is Kf6 or Nxg6, and the h7-pawn remains
an incorrect assessment. White can a fixed weakness. Now White can focus
break through on b5, either by on b4-b5: 39.Rb2 Raa8 40.Rgb1 Rgb8
sacrificing the exchange or by rerouting (any move that doesn’t protect b7, let’s
the knight to c3. However, as is always say 40...Rge8, can be met with: 41.b5!
the case with permanent advantages, cxb5 42.axb5 Bxb5 43.Rxb5! axb5
good technique dictates that everything 44.Rxb5 Reb8 45.Rb6! and the black
must be improved to the maximum position collapses) 41.Ne2! (with the
before delivering the final blow. Hence, b7-pawn pre-emptively protected,
the first order of business is to get the h- 41.b5? runs into 41...axb5 42.axb5 cxb5
pawn to h6: 33.h4 Re8 34.h5 Rg8 43.Rxb5?! Bxb5 44.Rxb5 Ra3! and
35.Ra1 Ra7 36.Rh2 Raa8 37.Rg1 Ra7 Black gains a ton of counterplay)
(37...a5? 38.Rb2 does our work for us) 41...Rf8.
38.h6 g6.
 8
t+*+*T*+ 8
t+*T*+*+
7
+o+vL*+o 7
+o+*LvOo
6
o+o+o+oP 6
o+o+o+*+
5
+*PoKoP* 5
+*PoKoP*
4 pP*P*+*+ 4 pP*P*N*+
3 +*+*Pp+* 3 R*+*Pp+p
2 *R*+n+*+ 2 *+*+r+*+
1 +r+*+*+* 1 +*+*+*+*
a b c d e f g h a b c d e f g h

White to move White to move

With 42.Kf4! (let’s not blunder any 33.h4

...f5-f4 tricks) 42...Rfb8 43.Nc3 Be8
44.b5 the break is finally executed, and Pushing the pawn to h6 is again a
Black has no hope of holding. The b7- good place to start.
pawn, combined with the worse king
33...Re8 34.h5 g6 35.h6
position and lack of space, are just too
much. This is the granddaddy of all bad
In the game, the bishop wasn’t bishops.
protecting the b5-square, so it was even Opening a file on the kingside with
easier. 35.hxg6 would also be quite sensible,
but it’s unnecessary, as we can easily
open one on the queenside. So, fixing
the weakness on h7 and gaining even
more space is a better technical decision.

35...Bg8 36.Rb3 Kd7 37.Reb2 Bf7

 8
t+*+t+*+ 8
*T*+t+*+
7
+o+l+v+o 7
+oL*+v+o
6
o+o+o+oP 6
*R*+o+oP
5
+*PoKoP* 5
+*PoKoP*
4 pP*P*N*+ 4 *+*P*N*+
3 +r+*Pp+* 3 +*+*Pp+*
2 *R*+*+*+ 2 *R*+*+*+
1 +*+*+*+* 1 +*+*+*+*
a b c d e f g h a b c d e f g h

White to move White to move

38.b5 42.c6!

The breakthrough is achieved. When Trading rooks means that the white
the two such armies come in contact, the king can safely infiltrate the black camp,
vast disparity in activity decides. embarrassing the sad bishop.
42.Nxe6+ wins, but with experience
38...axb5 39.axb5 Rab8 40.bxc6+ Kxc6 comes wisdom, and old geezers like me
41.Rb6+ Kc7 can’t help but remember the countless
times we’ve bungled our positional
“masterpieces” by going for something
more concrete than we should have.
Yes, White wins a pawn, and, yes, it’s
completely winning, but allowing Black
to part with that bishop is an
unnecessary concession.

42...bxc6 43.Rxb8 Rxb8 44.Rxb8 Kxb8

45.Kd6 Kb7 46.Ke7 1-0

Black resigned, as 46...Bg8 47.Kf8

traps the laughing-stock bishop. Then
White picks off the h7-pawn and
promotes his own h-pawn completely 8
tM*WlVmT
unobstructed. 7
Oo+*+oO*
This was an extreme example of a
good knight versus an absolutely
6
*+o+o+*+
horrible bishop. This type of advantage
5
+*+oPv+o
tends to be stable and permanent in 4 *+pP*+*P
nature. If the position wasn’t stable, how 3 +*N*+*+*
pP*+*Pp+
could the knight be superior anyway? 2
Let’s move on to something similar
but more complex. In the following
1 R*BqKbNr
example, it is highly instructive to a b c d e f g h
observe how Karpov accumulated a
series of imperceptible advantages,
slowly but surely turning an even Black to move
position into a highly unpleasant
situation for White. Not only that, but he 6...Nd7?!
will manage to do so against Ivanchuk,
who at the time held the fourth highest This has been played many times, but
rating in the world. Because Karpov is considered inaccurate by modern
makes this win look simple, it does not theory.
mean it is easy. 6...Ne7 and 6...dxc4 are better moves.
Game 13
7.cxd5!
Vasyl Ivanchuk – Anatoly Karpov
Tilburg (rapid) 1993 Ivanchuk seals d5 with a black pawn.
This way, Black is denied an occupiable
1.e4 c6 hole on d5 after a future ...d5xc4.
If you play the Caro-Kann, then it’s 7...cxd5
obligatory to study Karpov’s games.
The d7-knight is slightly embarrassed:
2.d4 d5 3.e5 Bf5 4.h4!? h5 5.c4 e6 when it went there a move ago it didn’t
6.Nc3 know c6 was going to be available.

8.Bd3

This is the main line, trading off the

strong bishop on f5. Yet there is some 2) Black’s king is slightly more
strategic cost since, after the exchange, exposed.
White’s dark-squared bishop can turn 3) On the flipside, Black increased his
into a bad piece. Most of White’s pawns central presence and, especially in an
are fixed on the same color. ending, he may be able to seize control
over the center with ...e6-e5.
8...Bxd3 9.Qxd3 Ne7 10.Bg5
12...Kf7
Ivanchuk provokes ...f7-f6, which
would riddle the black kingside with Karpov gambles that he can keep his
weaknesses, at least as far as a king safe in the middle, since his pawns
middlegame is concerned. control the key squares on e5, f5 and g5,
blocking out potential white attackers.
10...f6 11.exf6 gxf6 12.Bd2
13.Nge2 Ng6
12.Bf4! is a better diagonal for the
bishop. Aiming to exchange a pair of minor
pieces in the event of Ne2-f4.
8
t+*WlV*T 13...Nb6! is the engine suggestion,
7
Oo+mM*+* aiming to meet 14.Nf4 with 14...Bh6.
6
*+*+oO*+ 14.g3?!
5
+*+o+*+o
4 *+*P*+*P There is no reason to weaken the light
squares and place more pawns on the
3 +*Nq+*+* same – and therefore wrong – color of
2 pP*B*Pp+ White’s remaining bishop. The engine
1 R*+*K*Nr likes White after 14.Nf4! and a knight
a b c d e f g h swap. Black’s king looks precariously
placed on f7 and, after castling kingside,
White can double rooks on the e-file,
with the better chances.
Black to move
14...Bd6 15.Nb5
A few alterations have occurred:
1) White’s bishop doesn’t look so bad This doesn’t really gain time since
anymore because of the more open Black regains it with a future pawn push
center. to a6.
15...Bb8 16.Bb4!? the bishop occupies an irrelevant
diagonal.
3) In addition, five out of six
8
tV*W*+*T remaining white pawns are on the same
7
Oo+m+l+* color as the bishop, which means that
6
*+*+oOm+ Ivanchuk may later suffer on the light

+n+o+*+o
5 squares.
4 *B*P*+*P 19...Qa5+!
3 +*+q+*P* Forcing White to make a concession.
2 pP*+nP*+
1 R*+*K*+r 20.Qd2?!
a b c d e f g h
If White’s bishop is bad, then the last
thing he wants is an endgame.
20.Nc3? lands White in trouble after
Black to move 20...Rac8! when the awkwardness in the
placement of White’s pieces can be
16...a6!? highlighted in the line: 21.0-0?! b5!
22.Bd6 Qb6! and Black wins.
Karpov is not convinced that Nd6+ is
However, 20.Kf1! is a better option
a real threat. With his last move, he
than trading queens, despite the engine
disputes Tartakower’s theory that the
evaluating both at ‘0.00’. After
worst bishop is still better than the best
20...Nb6! 21.Kg2 Nc4 Black has the
knight, and opts to create an interesting
option of chopping the bishop, but the
imbalance.
opening of the b-file will allow White to
Karpov could have kept his dark-
generate counterplay. This is much
squared bishop if he wanted to start with
easier for White to handle compared to
16...Qb6.
the game.
17.Nd6+ Bxd6 18.Bxd6 Qb6 19.Ba3
20...Qxd2+ 21.Kxd2 Rac8 22.Rhc1
Is this a good or a bad bishop? My
guess is the latter:
1) The bishop is “active”, looking at
an open a3-f8 diagonal.
2) However, one could also argue that
 1) Black gets a mobile pawn center.
8
*+t+*+*T 2) Black obtains a passed d-pawn.
7
+o+m+l+* 3) The f-file is opened for Black’s
6
o+*+oOm+ rooks.
5
+*+o+*+o 4) Black’s space advantage in the
center increases, as a pawn on the fourth
4 *+*P*+*P rank (d4) is traded for a pawn on the
3 B*+*+*P* sixth rank (f6).
2 pP*KnP*+ It would feel much more natural to
1 R*R*+*+* finally put a pawn on a light square with
24.f3!? and capture on e5 only if White
a b c d e f g h
is forced to do so by Black increasing
the pressure against d4.

Black to move 24...fxe5

22...e5!
8
*+t+*+*T
Karpov immediately occupies the 7
+o+m+*+*
o+*+l+m+
center, while also clearing the e6-square 6
to centralize his king even further. It
doesn’t matter that White can hold this –
5
+*+oO*+o
it is abundantly clear which side is the 4 *+*+*+*P
one pressing. 3 Bp+*+*P*
22...Nb8!?, followed by rerouting the
knight to c6, is another good and typical
2 p+*KnP*+
idea.
1 R*R*+*+*
a b c d e f g h
23.b3 Ke6 24.dxe5!?

Sometimes principles should be

deliberately violated. Ivanchuk follows White to move
the Principle: Open the game when you
25.Rxc8
have a bishop, and your opponent has a
knight. The engine says this is OK, 25.f4!? is risky, as after 25...exf4
while your human writer doesn’t like it, 26.gxf4 White’s h-pawn is loose. Is
for several reasons:
control over the d4-square worth it for
White? The engine says yes. Principle: When pushing pawns
forward, it is usually correct to do so by
25...Rxc8 26.Rc1?! putting them on the color of the
opponent’s unopposed bishop. When
I believe this is the true point where
you think about it, this is quite logical –
Ivanchuk’s position begins to sour. He
White could easily create a blockade on
follows the Principle: In minor-piece
the dark squares, but it’s impossible to
endgames, outside passed pawns are
create one on the light squares. The
more valuable than central passed
pawns usually both want to advance one
pawns. However, White will have a hard
step forward, in our case from d5 and e5
time creating that passed pawn in the
to d4 and e4, but the principle above can
ensuing struggle.
tell us which one should advance first.
It was more important to break the Examining this particular case, Karpov’s
connection between the formidable move serves multiple purposes:
pawn duo with 26.f4!. After 26...exf4
1) Black gains space.
27.Nxf4+ Nxf4 28.Re1+! Kf5 29.gxf4
2) White is denied the squares c3 and
Nf6 30.Re7 White should be able to
e3 for his pieces.
hold.
3) The d5-square is cleared for Black’s
26...Rxc1 27.Bxc1 king, who is obviously superior to his
counterpart on d2.
In contrast, 27...e4? would create
8
*+*+*+*+ weaknesses all around the dark squares
7
+o+m+*+* in Black’s camp, and White could keep
6
o+*+l+m+ the balance in multiple ways: for

+*+oO*+o
5 example, 28.Nd4+ Kf7 29.Kc3 Nge5
30.Be3, occupying all the newly created
4 *+*+*+*P weaknesses.
3 +p+*+*P*
p+*KnP*+
2 28.f3 Kd5

1 +*B*+*+* Whenever White dares to push f3-f4

a b c d e f g h or g3-g4, the e5-pawn will advance to
e4, and the central passers will
overwhelm White.

Black to move 29.Ba3 a5 30.Kd3?!

27...d4!
 This walks straight into Karpov’s idea. 31.Kc2 b4 32.Bc1 Nc5 33.Ng1.

30...b5! 31...a4

Of course.
8
*+*+*+*+
+*+m+*+*
7 32.Bc1

6
*+*+*+m+ The desperate 32.g4?? fails miserably
5
Oo+lO*+o to 32...hxg4!? 33.h5 gxf3.
4 *+*O*+*P 32...Nf6!
3 Bp+k+pP*
p+*+n+*+
2 Cutting off g3-g4 completely, while
also contemplating ...e5-e4+.
1 +*+*+*+*
a b c d e f g h 33.Bg5

8
*+*+*+*+
White to move 7
+*+*+*+*
Karpov begins to annex queenside 6
*+*+*Mm+
space. 5
+o+lO*Bo
31.b4?
4 oP*O*+*P
3 +*+k+pP*
This is a case of the cure being more
harmful than the disease. Ivanchuk pays
2 p+*+n+*+
too high a price to keep Black’s knight
1 +*+*+*+*
out of c5. White shouldn’t have allowed a b c d e f g h
his b-pawn to be fixed on the same color
as his remaining bishop. In addition, by
pushing the pawn, he created a hole for Black to move
Karpov’s knights on c4.
It’s psychologically difficult for a 33...e4+!
strong player to adopt a passive stance.
However, this is exactly what White With this break, Karpov manages to
should have done. Ivanchuk had a far target White’s pawns on both wings, on
better chance to save the game with g3 and a2.
34.fxe4+ Nxe4 35.Nxd4
8
*+*+*+*+
+*+*+*+*
This loses by force, but, at this point, 7
there wasn’t anything better: for
example, 35.Nf4+ Nxf4+ 36.gxf4
6
*+*+*+*+
Nf2+! 37.Ke2 Ng4 (threatening ...Ke4) 5
+o+*+*Bo
38.Kd3 a3! 39.Bd8 (with the knight on 4 oP*Lm+*P
+*+*+*P*
g4 instead of e4, 39.f5 allows 3
39...Ne5+) 39...Nf2+ 40.Ke2 Ne4
41.Kd3 Nc3. Black stole a tempo and
2 p+*+*+*+
wins easily. 1 +*+k+*+*
a b c d e f g h
35...Ne5+

The start of a pretty dance that exploits

the weakness of the knight on d4. Black to move
“Loose pieces drop off!” they say, and
truer words have never been spoken in White is lost, yet the position remains
the world of chess. cloaked in murk when it comes to the
exact details:
36.Ke3 Ng4+ 37.Kd3 Ngf2+ 38.Ke3 1) The g3- and a2-pawns are
Nd1+! 39.Ke2 simultaneously targeted, one through a
direct attack and the other via a fork on
39.Kd3 Nef2+ drops the knight on d4,
c3.
and after 40.Kd2 Kxd4 41.g4 Nxg4
42.Kxd1 we get a slightly worse version 2) Black has a monster king position.
of what could have happened in the 3) White’s bad bishop is a glorified
game. rickshaw puller, while Black’s powerful
knight can pick and choose between
39...Kxd4! 40.Kxd1 outposts to sit on and weak pawns to
capture.
Exercise: Karpov must deal with
unfinished business. Which pawn
should he take?

Answer: The knight benefits from

stability, so taking the g3-pawn is an
easy win, while taking a2 should allow
White to hold a draw by creating
counterplay on the kingside. 8
*+*+*+*+
40...Nc3+?
7
+*+*+*+*
6
*+*+*+*+
Karpov was undoubtedly low on the 5
+o+*+*Bo
oP*L*+*P
clock. This was a blunder, which should 4
have nullified all of his previous superb
strategic play.
3 +*+*+*P*
40...Nxg3! is an easy win for Black. 2 m+k+*+*+
His king and knight will easily capture
one of White’s weaknesses, since all of
1 +*+*+*+*
a b c d e f g h
them are fixed on the color of the
defending bishop (as the a2-pawn will
soon advance to a3), meaning that all the
light squares belong to Black. To White to move
illustrate: 41.Kc2 Nf5 42.Bd8 Kc4
43.a3 Nd4+ 44.Kb2 Kd3 45.Bg5 Nb3 Exercise: Should White play
46.Bh6 Nd2! 47.Bg5 Nc4+ 48.Ka2 Kc2 42.Bd2, protecting b4, or should he
49.Bh6 Nb6 50.Bg7 Nd5 51.Bb2 Nc3+ sacrifice another pawn with 42.g4?
52.Ka1 Kb3 53.Bc1 Nd5 54.Bb2 Ne3
when ...Nc4 is coming, and the game is Answer: White should jump at the
over. chance to create a passed pawn on the
This technique, with a knight and king kingside. The sooner, the better.
picking off pawns fixed on the color of
the bishop, was not just known by 42.Bd2?!
Karpov, it was the main theme of his
As all sentient movie-robots say, “This
legendary 9th game win over Kasparov
doesn’t compute!” The passive defense
in their first world championship match of b4 loses a tempo and is clearly bad
in Moscow 1984/85, which is possibly Feng Shui. Amazingly, White can still
his most famous game. hold a draw, even after the inaccuracy,
41.Kc2! Nxa2 but that’s only in the world of engines.
If players this strong can flub it in such
simplified positions, then we shouldn’t
feel all that bad when we do the same.
With 42.g4! White cleanly saves the
game, despite being down three pawns!
After 42...Nxb4+ 43.Kd2 Ke5! (with
43...hxg4?? Black even manages to lose, 8
*+*+*+*Q
since after 44.h5 no one can catch the h-
pawn) 44.gxh5 Ke6 neither side can
7
+*+*+*+*
make progress, and the game will end in
6
*+*+*+*+
a draw.
5
+o+*+*+*
4 *L*+*+*+
O*+*+*+*
42...Kc4 43.g4! 3
There is absolutely nothing else. 2 k+*+*+*+
43...hxg4 44.Kb2?
1 +*+*+*W*
a b c d e f g h
Our suffering can be either physical,
or in our head. This move leads to a
painful loss.
Black to move
The engine holds it, but only barely:
44.h5! g3 45.h6 g2 46.h7! (46.Be3 loses The handsome prince awakens
to 46...Nxb4+ 47.Kd2 Nd5! with the Sleeping Beauty with a tender kiss,
double threat of ...Nxe3 and ...Nf6) soaked in true love. Then Sleeping
46...g1=Q 47.h8=Q Nxb4+ 48.Bxb4 Beauty immediately contacts her lawyer
Qg6+! (of course, not 48...Kxb4?? to sue the prince for sexual harassment,
49.Qc3 mate) 49.Kc1 Kxb4 50.Qd4+ since the kiss was clearly non-
and White can hold this, but not by consensual. I just described the level of
giving perpetual check – the checks disillusionment when we promote to a
soon run out. Computers realized that new queen only to be mated a few
this endgame is theoretically drawn a moves later. 50...Qg2+ 51.Kb1 Kb3! is
few years ago, but no one really cares. a nightmare for White. His queen lacks
Humans will nearly always lose it. even a single viable check despite the
board being wide open.
44...Nxb4 45.h5
45...Nd3+ 46.Ka2 g3!?
Also losing is: 45.Bxb4 Kxb4 46.h5
a3+ 47.Ka2 g3 48.h6 g2 49.h7 g1=Q Also winning and possibly simpler is
50.h8=Q. 46...Ne5! 47.Bf4 Nf7 48.h6 Nxh6!,
eliminating that troublesome pawn.

47.h6 Ne5 48.Bf4 g2! 49.Bxe5 g1=Q

50.h7 opponent strategically, in the end, we
must follow through by calculating
lines. More to the point though, it also
8
*+*+*+*+ teaches us that a bad piece, in this case
7
+*+*+*+p White’s dark-squared bishop, can be the
6
*+*+*+*+ reason behind a defeat.

+o+*B*+*
5 Moreover, we saw this bishop being
bad in a number of different ways. At
4 o+l+*+*+ first, it was bad mostly because White’s
3 +*+*+*+* pawns were fixed on the same color,
2 k+*+*+*+ restricting the bishop’s activity. Then,

+*+*+*W*
1 19...Qa5+! revealed that the bishop was
also vulnerable to an attack. 20.Nc3?
a b c d e f g h was impossible, so White felt the need
to enter an endgame. In the endgame,
the a3-f8 diagonal was clearly irrelevant,
Black to move and the bishop proved to be useless,
despite being outside the pawn chain.
Panic might be our first instinct when
confronted with this position, especially The next game presents a case where
with seconds on the clock. But Karpov being stuck with a bad bishop shouldn’t
is not one of us and keeps his cool. have been enough to lose the game. Yet
in such good knight versus bad bishop
50...b4! endings, it’s easy to blow it, since it can
be tough to work out the details to reach
Also winning is 50...Qf2+! 51.Bb2
a clear draw. The engines can draw, but
when both the calm 51...b4! and the
we humans blow such endings with
more panicky 51...a3!? 52.h8=Q Qxb2+
depressing regularity.
work out just fine.
Game 14
51.h8=Q b3+! 52.Ka3 Qc5+ 0-1 Alexander Khalifman – Dibyendu
Barua
The prison’s wall is too high for the Las Vegas 1999
white king to climb. White resigned.
The end would have been: 53.Kb2 Qf2+ 1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.Nf3 Be7
54.Kb1 Qa2+ 55.Kc1 Qc2 mate. 5.Bf4 0-0 6.e3 c6
This game teaches us the lesson that
no matter how much we outplay the This is the ultra-solid old Orthodox
Defense to the Queen’s Gambit, the 18.a5!
darling of Lasker and Capablanca’s eras.
I view the line like that friend whose Gaining space, putting pressure on b6,
college degree is in the History of Folk and trying to force even more black
Dance and who insists the degree is both pawns on to light squares.
relevant and will eventually get her a
18...b5?!
high-paying job. The line is super-solid,
yet utterly drained of dynamism. Principle: Don’t allow your pawns to
Today’s GMs prefer more active lines. be fixed on the same color as your
remaining bishop.
7.h3 Nbd7 8.Qc2 b6 9.Bd3 Bb7
10.cxd5 cxd5 11.0-0 a6 12.a4 Nb8 18...Nd7?! is met with 19.Na4!,
13.Ne5 Nc6 14.Nxc6 Bxc6 15.Rfc1 forcing Black to push the b-pawn
Bb7 16.Qb3 Bd6 17.Bxd6 Qxd6 forward anyway, but 18...bxa5! is
playable since White’s queen gets
Black’s bishop looks bad on b7, since trapped if she captures the loose bishop
the d5- and e6-pawns are fixed on the on b7. After 19.Rxa5! Rfb8 20.Qa2 Nd7
wrong color. But we must ask: is this Black looks better than in the game
single factor enough for White to win, continuation, since his pieces gained
given that everything else is some degree of activity and he may
symmetrical? The answer is: not yet. generate play down the open b-file, even
if the a6-pawn falls.
8
t+*+*Tl+ 19.Na2!?
7
+v+*+oOo
6
oO*WoM*+ Khalifman seeks to retain pieces to

+*+o+*+*
5 complicate the position. White’s plans
are:
4 p+*P*+*+ 1) The knight may transfer to b4, after
3 +qNbP*+p which it eyes c6 and has options of
2 *P*+*Pp+ shifting to d3 and then either c5 or e5.
1 R*R*+*K* 2) With a knight on a2, White has the
option to play Qb4. Black then has a
a b c d e f g h
choice of either swapping queens into a
depressingly passive ending or backing
off and refusing the swap, when White’s
White to move queen stands well and controls several
key dark squares in the black camp.
 However, this encounter is the first
game in Khalifman’s historic run that 8
t+t+*+l+
won him the 1999 FIDE World 7
+v+*+oOo
o+*WoM*+
Championship. The knockout format of 6
the tournament makes games with White
feel like must-win situations, and in
5
Po+o+*+*
such situations, players tend to avoid 4 *+*P*+*+
simplicity. 3 +q+bP*+p
nP*+*Pp+
In another game, Khalifman might 2
have found it psychologically easier to
go for 19.Na4!?, heading for c5. Now
1 R*R*+*K*
19...bxa4?! 20.Qxb7 Rfb8 21.Qc6! gives a b c d e f g h
White a decisive advantage because of
the weakness of the pawn on a6, so
Black needs to go for 19...Nd7,
White to move
transposing to the note above on
18...Nd7?!. After 20.Nc5 Nxc5 21.Rxc5 20.Qb4! Qxb4
Rfc8 22.Rac1 White has a large
advantage, especially in practice, but As we previously noted, backing off
Khalifman must have felt this wasn’t with the queen cedes control of the dark
enough, or that Black’s position is squares: for example, 20...Qd8 21.Rxc8
simpler and thus easier to defend Rxc8 22.Nc1! when White is planning
compared to the game. Nc1-b3-c5 and holds a clear advantage.

19...Rfc8 21.Nxb4 Kf8 22.Kf1 h6 23.Ke1 Ke7

24.Kd2 Kd6

I don’t trust 24...Ne4+?! 25.Bxe4 dxe4

26.Rc5!. Black has allowed a classic
good knight versus bad bishop scenario.

25.f3 Rxc1

Barua’s plan is to swap off rooks and

then try to erect a fortress draw.

26.Rxc1 Rc8
 Let’s forget the comment above. We
8
*+t+*+*+ once again reach one of those almost-
7
+v+*+oO* but-not-quite equal positions, which

o+*LoM*O
6 may theoretically be drawn, yet White
can play on for eternity, or perhaps even
5
Po+o+*+* a bit longer. This type of play always
4 *N*P*+*+ gives the better side good practical
3 +*+bPp+p chances to win.
2 *P*K*+p+ 28.Be2!
1 +*R*+*+* The d3-square is cleared to make way
a b c d e f g h
for Nd3, after which the knight can
move to either e5 or c5.

White to move 28...Bb7 29.Nd3 Nd7

27.Rxc8?! White’s knight is kept out.

Khalifman contradicts my note on the 8

*+*+*+*+
19th move and goes for simplicity. Even
though he definitely keeps an advantage,
7
+v+m+oO*
this is not the best move.
6
o+*Lo+*O
In positions with such a dominant
5
Po+o+*+*
position, and especially since White 4 *+*P*+*+
controls all the entry squares on the c-
file safely, it is better to keep the rooks
3 +*+nPp+p
on the board. So, any move with the
2 *P*Kb+p+
rook along the first rank is better. Let’s 1 +*+*+*+*
say 27.Rg1!. It might look silly at first, a b c d e f g h
but the rook can find future employment
on a different open file on the kingside
after we push our pawns. If we regret
White to move
our decision, whenever we want we
could go back to c1. 30.f4
27...Bxc8 This way, White discourages a break
with ...e5. White isn’t concerned about
the newly created hole on e4, since he 33.Kb4
will have his bishop ready to take
Black’s knight if it ever hops into the 33.f5! was winning again.
square.
33...Kd6 34.Be8 Nb8!
30...Bc6

Black just waits to see how his

8
*M*+b+*+
opponent makes progress. 7
+*+*+*O*
31.Bh5!
6
o+vLoO*O
5
Po+o+*+*
When trying to break down a passive 4 *K*P*P*+
+*+nP*+p
defense, almost every pawn push by our 3
opponent is a welcome sight. Each pawn
push creates new weaknesses, and those
2 *P*+*+p+
are what we need to eventually break 1 +*+*+*+*
through. a b c d e f g h
31...f6 32.Kc3

32.f5! was an excellent chance for White to move

White, as 32...e5? 33.Nb4 Bb7 34.Bf7!
ties Black in a knot. White has somehow Barua is doing a good job with his
managed to put pressure on both d5 and project of creating a fortress. Khalifman
a6, while his king is free to travel keeps trying though.
towards the kingside through the
weakened light squares and pick up the 35.Bg6 Bd7 36.Kc3 Nc6
g7-pawn. It’s logical to believe that if
This way White is forced to occupy b4
White manages to exploit all of Black’s
with a white pawn.
weaknesses at the same time, the
defense will collapse. 37.b4
32...Ke7 This is a disappointing moment for
White, as c3-b4-c5 was an important
Black should have hurried to place the
route the king hoped to use, while the
knight on c6 and the bishop on c8,
knight might have found access to the
starting with 32...Nb8!.
b4-square useful, as from b4 it could
target both the a6- and d5-pawns. better chances to hold by keeping his
However, Khalifman’s move keeps structure flexible, with a waiting move
some winning chances alive, unlike like 42...Ne7. Objectively, both moves
37.Nb4 Nxa5 38.Nxa6 Nc4 when Black should lead to a draw, so it is mostly a
no longer stands worse. matter of practicality.

37...Ne7 38.Bh5 Nf5 39.Kd2 Bc8 43.Bh5 Ke7

40.Nc5 Ne7 41.Bf3 Nc6 42.Kc3
Covering the e8- and f7-squares to
prevent the bishop from infiltrating.
8
*+v+*+*+ 44.Nd3 Bd7 45.Nc5 Bc8 46.Bf3 g6
7
+*+*+*O* 47.g4 Kd6 48.h4 fxg4 49.Bxg4
6
o+mLoO*O
5
PoNo+*+*
4 *P*P*P*+
8
*+v+*+*+
3 +*K*Pb+p
7
+*+*+*+*
2 *+*+*+p+
6
o+mLo+oO
1 +*+*+*+*
5
PoNo+*+*
a b c d e f g h
4 *P*P*PbP
3 +*K*P*+*
2 *+*+*+*+
Black to move
1 +*+*+*+*
a b c d e f g h
42...f5!?

Ugh! Principle: Don’t fix your pawns

on the same color as your remaining Black to move
bishop. GM Barua is aware of this
principle yet deliberately chooses to 49...h5!?
violate it, as he feels that the goal of
Ugly stuff. Every black pawn is on the
creating a fortress supersedes the
same/wrong color as the remaining
strategic considerations for his
bishop on c8. Still, the engine says this
remaining bishop.
move is OK, as it stops h4-h5, which is
There is no good way to prove my one of the only two pawn breaks White
point, but I have a feeling that Black had had at his disposal. There is a way to
deal with e3-e4, so Black has indeed something. Only three moves could
built an impenetrable fortress. ever be viable: A) 52...Kc7, B)
The alternative is the set-up 49...Ne7 52...Nd5+, or C) 52...Nf5.
50.h5 gxh5 51.Bxh5 Nf5 52.Kd2 Ke7, Everything else loses material.
which should also hold, but I think it is a What should Black play?
little more unpleasant compared to the
game, due to White’s ability to infiltrate Answer: Two of the three possible
the kingside with his king. If we say ‘A’, moves are solid, but Black chose the
we need to say ‘B’, and on move 42, third one.
Black decided to go for building a
fortress. Only one pawn sitting on a dark 52...Nd5+?
square on h6 doesn’t change that much
in Black’s camp – it’s not like If you fail to blend in, you are no
controlling one dark square on g5 was longer safe in society. This is a case of
the key. the right plan, wrong implementation.
However, the refutation isn’t all that
50.Bf3 Ne7 51.e4 dxe4 52.Bxe4 simple, and even a player of
Khalifman’s caliber did not find it – at
least, not at first.
8
*+v+*+*+ 52...Kc7?! holds, but the situation
7
+*+*M*+* remains a bit nerve-wracking, since
6
o+*Lo+o+ White can move around and continue
posing problems. For example, with
5
PoN*+*+o 53.Nd3 aiming to transfer the knight to
4 *P*PbP*P e5. In that regard, Black’s wish to take
3 +*K*+*+* concrete action is more than
2 *+*+*+*+ understandable, but d5 wasn’t the right
square for the knight.
1 +*+*+*+* The correct approach was: 52...Nf5!
a b c d e f g h (White must take the knight, otherwise
h4 falls) 53.Bxf5 gxf5 when Black has a
wretched bishop yet will save the game
Black to move via a fortress draw. He can continuously
move his king back and forth between
Exercise: White is threatening d6 and d5, and the f5-pawn controls the
nothing really, but Black is in some e4-square, so there is no way for the
sort of zugzwang. He needs to move white king to advance. For example:
54.Kd3 Kd5 55.Nb3 Bd7 56.Nd2 Bc6 56...exd5+ 57.Kd4, which is zugzwang
57.Nf1 Kd6 58.Ne3 Be4+ 59.Kd2 Bc6. number 2. Black must either lose his a6-
It is clear by now that Black has solved pawn or allow the white king to enter on
all his problems and White is unable to e5.
make progress.

53.Bxd5 exd5
8
*+v+*+*+
7
+*+*+*+*
Sadly, with the pawn on g6 instead of 6
o+*L*+o+
f5, 53...Kxd5 also fails, as after 54.Kd3
Black is in zugzwang. Black’s king must
5
PoNo+*+o
back off, giving White’s king access to
4 *P*P*P*P
e4: 54...Kc6 55.Ke4 Kd6. 3 +*K*+*+*
2 *+*+*+*+
8
*+v+*+*+ 1 +*+*+*+*
7
+*+*+*+* a b c d e f g h
6
o+*Lo+o+
5
PoN*+*+o
4 *P*PkP*P White to move
3 +*+*+*+* 54.Kd3?
2 *+*+*+*+
1 +*+*+*+* We treat our position with love, and it
returns our kindness with malice. Chess
a b c d e f g h
is a harsh game, where even the tiniest
infraction can be punished. White
needed d3 for his knight and moving the
White to move king there is enough to allow Black to
hold. In reality, this is mostly a case of
But wait, how does White make White not understanding that he has to
progress here? If the knight moves, the be concrete and win right now, instead
bishop will give a check on b7, pushing of moving around and looking for the
the white king back once again. The breakthrough further down the road.
answer is: 56.d5!! (when you give up Moves which look the same, don’t
too much space and get a position that is necessarily act the same. 54.Kd2!! wins,
nothing but weaknesses, such
but only because against 54...Ke7 White
breakthroughs are not all that surprising)
has the amazing 55.f5!!, creating an
entry path into the black position via the wretched bishop, since White cannot
f4-square. With the king on d3, the f5- improve his king position. As long as
pawn falls with check, and the bishop White can only use the knight to attack
has time to return to c8. In this case, Black’s weaknesses, Black has no
Black has to take with the pawn, but that problems keeping everything together.
always loses: 55...gxf5 56.Ke3 Kf7
57.Kf4 Kf6. This is a position that 59...Kf6 60.Nd3 Bg4+ 61.Kg3 Bf5
occurred in the game, but on move 74, 62.Nc5 Bc8 63.Nb3 Bg4 64.Kf2 Bf5
and it is winning for White. 65.Nd2 Bc8 66.Nf1 Be6 67.Ne3 Kf7?!

54...Ke7 55.Ke3 Kf6 56.Kf3 Bg4+ A clearer path to the draw is 67...Bf7!
57.Kf2 Bc8 58.Ke3 Kf5 59.Kf3 68.Kg3 Bg8! and shuffling the bishop
between f7 and g8. Black stays ready to
meet 69.f5?! with 69...g5!, when he
8
*+v+*+*+ might even start getting ambitious.
7
+*+*+*+* 68.Kg3 Kg7?
6
o+*+*+o+
5
PoNo+l+o Black should not have decentralized

*P*P*P*P
4 his king.
68...Kf6? loses to 69.f5!! when
3 +*+*+k+* 69...Bxf5 70.Nxd5+ Ke6 71.Nc7+ Kd6
2 *+*+*+*+ 72.Nxa6 Kd5 73.Nc5 is easily winning
1 +*+*+*+* for White despite the counterplay,
because Black has a really hard time
a b c d e f g h
catching the a-pawn. However,
68...Ke7! should still draw after 69.f5
Bxf5! 70.Nxd5+ Kd6, since there is no
Black to move Nc7. It still looks scary after 71.Ne3
Bd3 72.Kf4 Bb1 73.Kg5 Bd3, but
Black’s position is the slick corporate whenever the knight moves to help
attorney representing the giant asbestos capture the pawn on g6, the black king
company, who somehow convinces the gains access to d5, and the white pawns
jury that the thousands of people in the start falling.
small town who contracted lung cancer
are unrelated to their product, which is
perfectly safe. The game is now
objectively drawn, despite Black’s
 70.Kf4 Kf6
8
*+*+*+*+
+*+*+*L*
7 The next step is to create a zugzwang
situation. An ideal one would be
6
o+*+v+o+ White’s knight on c5, Black’s bishop on
5
Po+o+*+o c8 (since the a6-pawn requires
4 *P*P*P*P protection), and Black having to move.

+*+*N*K*
3 Watch how Khalifman achieves this:

2 *+*+*+*+ 71.Nc2
1 +*+*+*+* The knight can either head for c5 or
a b c d e f g h f4.

71...Bd7 72.Ne1 Bc8 73.Nd3 Be6

74.Kf3! Ke7 75.Kg3 Kf6 76.Nc5 Bc8
White to move
77.Kf4
Exercise: Do you see White’s

*+v+*+*+
winning plan? 8

Answer: Sacrificial
7
+*+*+*+*
breakthrough/clearance.
6
o+*+*L*+
5
PoNo+o+o
*P*P*K*P
69.f5!! 4
A clearance sacrifice, granting the 3 +*+*+*+*
white king access to the f4-square. 2 *+*+*+*+
69...gxf5
1 +*+*+*+*
a b c d e f g h
This is the structure Black should have
avoided, but in this case, taking with the
bishop is even easier to refute: 69...Bxf5
70.Nxd5 Kf7 71.Kf4!? Bc8 72.Kg5!? Black to move
followed by Nf4, pushing the d-pawn,
Shortness of breath is a red flag for
and taking on g6. Black is completely
poor health. The leaking dark squares
lost. More concrete ways were also
are an unending source of despair for
available.
Barua. Black’s king must give way.
And, just like that, Black loses. defense when we are hobbled with a
dysfunctional piece in our camp.
77...Kg6 78.Ke5 f4 Another important takeaway concerns
our perception of fortresses. As you saw,
78...Kf7 79.Nd3 is also hopeless for
Black did technically have a fortress at
Black, as the knight comes to f4 and
one point (after 54.Kd3?), and he had
starts picking up everything.
the ability to erect a safer one (with
79.Kxf4 Kf6 80.Kg3 Kg7 81.Kf3 Kg6 52...Nf5!). However, Carlsen’s
82.Nd3 Bg4+ 83.Ke3 exaggerated claim of “not believing in
fortresses” showed its merciless face:
Threatening Nf4+ and Nxd5. fortresses exist, but humans often botch
them. Thus, relying upon one to make a
83...Be6 84.Nf4+ Kf5 85.Nxh5 draw is rarely a good idea.
An outside passed h-pawn is added to Regarding how to dismantle a fortress,
White’s trumps. an important piece of advice would be:
believe in your chances of success, even
85...Kg4 86.Nf4 Bf7 87.Nd3 Be6 when the fortress is objectively
impenetrable. Move around, and your
87...Kxh4 88.Nc5 wins the a6-pawn opponent is likely to screw up.
and the game. Moreover, always think about possible
zugzwangs. A fortress, in order to hold
88.Nc5 Bc8 89.h5! 1-0 up, should be both impenetrable and
leave the defender room for passing the
White’s king forces decisive entry to
move. Many players tend to either forget
e5, and Black saw no reason to continue.
or not pay enough attention to that
An inconvenience, while annoying, is
second aspect.
not the same as a mortal threat.
Likewise, a single bad piece is often not
As we have already seen, a single
enough to lose a game. Black’s dark-
inferior piece is not always enough to
squared bishop on move 17 is indeed an
lose a chess game. That was again the
inferior piece, yet Black’s solidity and
case in the following game, but the
ability to create a fortress draw should
position was unpleasant, and even
have held the game. This game does also
Nakamura eventually faltered and lost.
show that the side with the good minor
piece does get practical chances to win, Game 15
even when the position is theoretically Levon Aronian – Hikaru
drawn, since it isn’t easy to organize a Nakamura
 Wijk aan Zee 2014 remaining dark-squared bishop.
3) Is Black’s remaining bishop really
that bad? What if it reaches c5? Is the
8
t+vT*+*L g3-bishop really that much superior?
7
Oo+*+*Vo Was the offer to trade bishops a
6
*+*+*O*M mistake on Nakamura’s part? It’s
difficult to say, and it wasn’t terrible,
5
+*+*O*+* but I would have avoided the swap and
4 *+bMp+oN retained Black’s light-squared bishop
3 +*N*+*B* with 18...Bd7!.
2 pP*+*PpP 19.Bxe6
1 +*+r+rK*
a b c d e f g h Of course, Aronian is happy to leave
his opponent with a sickly remaining
bishop.

Black to move 19...Nxe6 20.f3!

Both sides have their set of trumps. This way, White’s bishop emerges
White has an excellent bishop on c4, and from its prison and can challenge
his c3-knight is ready to jump to d5. Black’s knight when it reaches d4. If it
Black has a great knight on d4 and a bit wasn’t for this, Black would have every
more space. Both sides are stuck with a reason to claim that the bishop on g7 is
weird set of knights and dark-squared no worse than the one on g3.
bishops on the kingside, all of which
need improvement. 20...gxf3 21.gxf3!?

18...Be6!? Fighting for control over d4 with

21.Nxf3 was equally sensible.
This is a fateful decision resembling
an improvised stretcher for the injured
patient, with long-term implications:
1) White’s light-squared bishop is
powerfully posted on its diagonal, which
tempts Black to offer a bishop swap.
2) After a swap of light-squared
bishops, Black will be stuck with a bad
 22.Rxd1 Rd8 23.Rxd8+ Nxd8 24.Bf2
8
t+*T*+*L a6 25.Kf1 Kg8 26.Ke2 Kf7 27.Be3!?
7
Oo+*+*Vo Let’s assess. Black must negotiate
6
*+*+mO*M multiple dangers:
5
+*+*O*+* 1) White has options to play Bxh6,
4 *+*+p+*N which then gives the h4-knight access to
3 +*N*+pB* f5, while leaving Black with a
potentially bad dark-squared bishop.
2 pP*+*+*P 2) We can visualize White’s king
1 +*+r+rK* walking up the central light squares via
a b c d e f g h d3 and c4, entering Black’s territory,
especially if more pieces get swapped.
The engine says the position is equal,
which may be true with two engines
Black to move
playing. However, with humans playing,
21...Rxd1?! there can only be two results: White
wins, or the game ends in a draw. So,
I don’t much like this decision either, White clearly holds the better practical
since it violates the Principle: Don’t be chances.
the one to release the tension between
rooks facing each other on an open file, 8
*+*M*+*+
+o+*+lVo
especially when your opponent can 7
recapture with a rook.
When we are at a disadvantage as far
6
o+*+*O*M
as long-term factors are concerned, it is 5
+*+*O*+*
better to choose the more complicated 4 *+*+p+*N
route. In this case, 21...Nd4!, thinking
about playing ...f6-f5 next is a better
3 +*N*Bp+*
option compared to the game. Pay
2 pP*+k+*P
attention to how, on a board full of 1 +*+*+*+*
pieces, ...f6-f5 makes perfect sense, as it a b c d e f g h
increases Black’s dynamic potential.
With fewer pieces, it will both be harder
to accomplish and less effective if it
ever happens. Black to move
27...Ke6 28.Na4 29.Bxh6!

Aronian is in no rush to chop the h6- Aronian expertly calculates that Black
knight, since if it moves, White’s knight won’t be in time to set up his fortress, at
jumps into the outpost on f5. Taking least not in optimal fashion, due to the
immediately somewhat simplifies tempo lost on the previous move.
Black’s task: 28.Bxh6 Bxh6 29.Nf5 Bf8
30.Nd5 Nf7 followed by ...Nd6 or 29...Bxh6 30.Nc5+ Kf7!
...Nh6, trading off one of White’s
For Nakamura’s level, the trap
impressive looking knights. White’s
30...Kd6?? 31.Nxb7+! Nxb7 32.Nf5+ is
advantage becomes merely symbolic, as
child’s play. White gets an extra pawn
Black has a relatively easy-to-hold
and a winning position.
fortress.
31.Kd3
28...Bf8?!
A new issue arises for Black: White
This inaccuracy costs Black a precious
threatens to move his king up the board.
tempo.
It isn’t easy for White to make
progress if Black plays the correct
8
*+*M*+*+
28...Kd7!.
7
+o+*+l+o
6
o+*+*O*V
8
*+*M*V*+ 5
+*N*O*+*
7
+o+*+*+o 4 *+*+p+*N
6
o+*+lO*M 3 +*+k+p+*
5
+*+*O*+* 2 pP*+*+*P
4 n+*+p+*N 1 +*+*+*+*
3 +*+*Bp+* a b c d e f g h
2 pP*+k+*P
1 +*+*+*+*
a b c d e f g h Black to move

31...Bf4?!

White to move This should still hold, but it’s not the
most accurate, and small inaccuracies
tend to add up.
31...Bf8? 32.Kc4! is lost for Black, 8
*+*+*+*+
highlighting one of the ways in which 7
+o+*+l+o
the lost tempo influences the evaluation,
but 31...Bc1! 32.b3 a5! was the best
6
o+*+mO*+
chance, even though it still looks
5
+*N*OnV*
unpleasant. 33.Kc4 Be3! 34.Nf5 Bg1! is 4 *+*+p+*+
the line in which the pawn remaining on 3 +*+k+p+p
pP*+*+*+
h2 makes a big difference. 2
32.h3 Bg5? 1 +*+*+*+*
a b c d e f g h
This looks a bit like panic, as it invites
the knight to go where it wanted to go
anyway.
Black’s best chance still was 32...a5!, White to move
but it’s a worse version of the line
Exercise: Should White capture the
shown above after 31...Bc1!, and even
b7-pawn? Or should he swap
that line looked sketchy.
knights, aiming for a good knight
33.Nf5 Ne6!? versus bad bishop ending? One line
leads to a win, while the other
Desperation. After 33...Kg6 34.a4 should result in a draw.
Black is paralyzed, and White will walk
the king up the board via c4 and d5. Answer: Despite appearances, White
should capture the b7-pawn. The good
knight versus bad bishop ending
shouldn’t be enough to win.

34.Nxe6?

Chess is such a miserably difficult

game – I would have almost certainly
played the same move.
Correct was to turn the game into a
scary mutual promotion race, which
requires White to exhibit good
calculation: 34.Nxb7!! Nf4+ 35.Kc4
Nxh3 36.b4 Nf4 37.a4 h5 38.b5 axb5+ 36.a4 Bd2 37.b3
39.axb5.

8
*+*+*+*+
8
*+*+*+*+
7
+n+*+l+*
7
+*+*+*+o
6
*+*+*O*+
6
oO*+lO*+
5
+p+*OnVo
5
+*+*On+*
4 *+k+pM*+
4 p+k+p+*+
3 +*+*+p+*
3 +p+*+p+p
2 *+*+*+*+
2 *+*V*+*+
1 +*+*+*+*
1 +*+*+*+*
a b c d e f g h a b c d e f g h

Black to move Black to move

The b-pawn is surprisingly hard to It was the Polish GM and quipster

stop for all the black pieces: 39...h4 Savielly Tartakower who wrote that the
(going for the race is the most logical, worst bishop is still better than the best
but it is bound to fail; 39...Ne6 also knight. While that may not be true in
loses easily after 40.b6 Ke8 and now the this position, Black’s bad bishop is just
simplest would be 41.Nc5 Nd8 42.b7 good enough to hold the game, as
when even Ng7+ followed by Nxh5 is a White’s king lacks entry points and
threat, so White can win this any way he Black should be able to achieve a
likes best) 40.b6 h3 41.Nbd6+ Kg6 fortress draw with correct play.
42.b7. White will promote first, and two
37...h5!?
knights plus a queen will surely deliver
checkmate against the naked black king: This sets a clever trap, as the h5-pawn
42...h2 43.b8=Q (even 43.Ng3 is is immune, but it would be misguided to
available) 43...h1=Q 44.Qg8+ Kh5 think that Nakamura was playing for a
45.Qh7+ Bh6 46.Qxh6 mate. win. In reality, Black is in zugzwang –
anything he plays would make his
34...Kxe6 35.Kc4 b6!
position worse.
White’s king must be kept out. Trying to keep the fortress by passing
with the bishop would be the most Bc3 41.Kxa6 Bd2 42.Kb5 Bc3 when to
natural, but 37...Be1? allows 38.Ne3! make progress, White needs to go after
when White brings the knight to d5, and the a5-pawn, but 43.Ne3 is met with
the black fortress collapses after 38...Bf2 43...Kf7! 44.Nc4 Kg6 45.Nxa5 Kg5,
39.Nd5. There is no way to protect and Black holds his own in the ensuing
everything, as apart from the f6- and b6- race.
pawns, the a6-pawn is also threatened Even though keeping the pawn on h7
by Nc7+. was safer, Nakamura’s move also holds,
So, what happens if we pass with the but the constant threat of Ng7+ makes it
bishop on the c1-h6 diagonal, keeping slightly trickier.
control of e3? Let’s say, 37...Bc1!. This
does not allow the knight to e3, but it 38.b4!
relinquishes control over the e1-a5
Going for the pawn would be ill-
diagonal, allowing 38.a5!? bxa5 39.Kc5
advised: 38.Ng7+ Kf7 39.Nxh5?? Be1!
and the white king infiltrates the black
(oops!) 40.f4 Kg6. Have you ever had a
camp.
bug crawl out of your salad? White’s
knight is trapped.
8
*+*+*+*+
7
+*+*+*+o 38...a5!
6
o+*+lO*+ The only move, keeping the bishop on
5
O*K*On+* the square that keeps in touch with both
4 *+*+p+*+ e3 and e1.
3 +p+*+p+p Allowing Ne3 is out of question, and

*+*+*+*+
2 38...Bc1?? loses to 39.b5! a5 40.Ng7+
Kf7 41.Nxh5, when the bishop can’t get
1 +*V*+*+* to e1 to trap the knight, and Black can
a b c d e f g h resign.

Black to move

Understandably, Nakamura must have

been scared of this, but the engine
informs us that the fortress doesn’t
break. For example: 39...Bd2 40.Kb6
 8
*+*+*+*+ 8
*+*+*+*+
7
+*+*+*+* 7
+*+*+*+*
6
*O*+lO*+ 6
*+*+lO*+
5
O*+*On+o 5
Ok+*On+o
4 pPk+p+*+ 4 p+*+p+*+
3 +*+*+p+p 3 +*+*+p+p
2 *+*V*+*+ 2 *+*V*+*+
1 +*+*+*+* 1 +*+*+*+*
a b c d e f g h a b c d e f g h

White to move Black to move

39.bxa5! Exercise: Black’s king must move,

but in which direction? Our choices
Keeping winning chances alive. are d7 and f7. Only one of them
After 39.b5? there is no more holds the draw. Which one would
weakness on a6, and White’s king is left you play?
without a path forward on the kingside,
so Black can play 39...Be1!, preventing
Answer: The king is needed to cover
the capture of the h5-pawn. But, wait,
the weak h5-pawn. Therefore, going to
doesn’t this allow Nf5-e3-d5? Yes, but
f7 is correct.
after 40.Ne3 Bf2 41.Nd5 Bc5 there is
no pawn on a6 to capture with Nc7+, 40...Kd7?
and White has no way forward. Moving
the king to the kingside can be met with Making a quick decision to reach the
...f6-f5-f4, which completely blocks the first time control often feels the same as
kingside. turning in your high school test paper
blank and then expecting to pass with an
39...bxa5 40.Kb5 excellent grade. Nakamura is one of the
quickest thinkers under fire in chess, yet
here the calculation needed to find the
correct path required more time than he
likely had available on his clock. In such
situations, where we have, say, 30 it isn’t obvious at all: 47.Nxe5! fxe5
seconds to make the final move, all we 48.f4! exf4 49.e5 f3 50.e6 f2 51.e7 f1=Q
can do is guess and then pray. In this 52.e8=Q. The position is a theoretical
case, Nakamura’s guess turns out to be draw, but Black retains good practical
wrong. chances. Queen endgames with both
40...Kf7! holds the draw, using the kings vulnerable to checks confuse even
same strategy as the one shown in the the best among us.
note regarding 37...Bc1!. White can put
the pawn on h4 and stay still, but then 41.Ng7
the black bishop targets the pawn from
The knight can’t be trapped with
e1, and neither side could ever make
Black’s king on d7. Thus the h-pawn is
progress. Going for the full point with
forced to move forward.
41.Nd6+?! is more of a losing attempt
than anything else, as the bishop is a 41...h4 42.Nf5
much better minor piece when it comes
to races on opposite wings. For
example: 41...Kg6 42.Nc4 Be1 43.Nxa5 8
*+*+*+*+
Kg5 44.Nc4 Kh4 45.a5 Bxa5! 46.Kxa5
Kxh3.
7
+*+l+*+*
6
*+*+*O*+
8
*+*+*+*+
5
Ok+*On+*
7
+*+*+*+*
4 p+*+p+*O
6
*+*+*O*+
3 +*+*+p+p
5
K*+*O*+o
2 *+*V*+*+
4 *+n+p+*+
1 +*+*+*+*
a b c d e f g h
3 +*+*+p+l
2 *+*+*+*+
1 +*+*+*+* Black to move
a b c d e f g h
For Black, this is a non-sustainable
business model:
White to move 1) White’s knight dominates the
bishop. This situation resembles bad
White has only one path to a draw, and Feng Shui, as every black pawn is fixed
on the same color as the remaining
bishop, which is the wrong color. 52.a7 h2 53.a8=Q h1=Q 54.Qg8+ Kf2.
2) White’s king is much the more
active, having already infiltrated the 8
*+*+*+q+
opponent’s camp. 7
+*+*+*+*
3) Black’s a-pawn is a fixed target,
and all White needs to do to capture it is
6
*+*+*O*+
maneuver his knight to c4.
5
+k+*O*+*
4) Black’s king is too far from White’s
4 *+*+p+*+
kingside pawns to have any chance in a 3 +*+*+p+*
promotion race. 2 *+*+*L*+
42...Be1 43.Ne3 Ke6 44.Nc4 1-0
1 +*+*N*+w
a b c d e f g h
Nakamura resigned, as Black loses in
the ensuing race: for example, 44...Kf7
45.Nxa5 Kg6 46.Nc6 Kg5.
White to move
8
*+*+*+*+ With 55.Qa2+! (the knight can’t be
7
+*+*+*+* touched, or else Black loses his queen)
6
*+n+*O*+ 55...Kg3 56.Qd2 White remains up a
5
+k+*O*L* piece and consolidates. It’s quite
impressive that a player would resign
4 p+*+p+*O due to a 12-move variation, but for such
3 +*+*+p+p players, long lines are nothing special if
2 *+*+*+*+ they’re completely forced. Both of these

+*+*V*+*
1 players have the ability to calculate
something like this in mere seconds.
a b c d e f g h
In hindsight, Nakamura’s decision on
move 18 to trade off the light-squared
bishops could be considered an
White to move inaccuracy. Even though the game
wasn’t lost at that point by any means,
After 47.Nb4! the threat is to play and the position remained holdable all
Nd3, forever barring Black’s king from the way to the very end, creating a
entry. This means that there is only one scenario where Black clearly had more
try for Black: 47...Kf4!? 48.Nd3+ Kg3 weaknesses (on the light squares) made
49.Nxe1 Kxh3 50.a5 Kg3 51.a6 h3
the situation one-sided and practically
unpleasant. 8
*Ml+*+*V
Creating another weakness in the face 7
+*+*+*+o
of the h-pawn with 37...h5!? was also a
questionable decision. The h-pawn
6
*O*+o+o+
entered the knight’s field of vision, and
5
O*+o+oP*
Black added an extra headache to his 4 p+*P*B*+
already troubled position. 3 +p+*Pn+p
By the way, White technically also
had plenty of weaknesses: it wouldn’t be
2 *+*+*Pk+
easy to protect the f3- and h3-pawns,
1 +*+*+*+*
and the dark squares throughout the a b c d e f g h
board were under firm control by Black.
However, as we’ve said a few times
already, a weakness is relevant only White to move
when it can be attacked or otherwise
exploited. In the game, Nakamura never Let’s assess:
managed to make White’s weaknesses 1) White owns more space on the
felt, but it is no accident that both kingside.
chances we mentioned above (37...Bc1! 2) White for now controls a hole on
and 40...Kf7!) aimed to do just that. e5.
3) There are very few meaningful
Let’s stay on the topic of fortresses
pawn breaks in the position. White’s
and examine the following game:
only real one is on h5.
Game 16 4) A serious threat of a sealed, fortress
Boris Gelfand – Vladimir draw looms over White.
Malakhov
Exercise: Let’s examine the
Sochi 2005
accusatory blood-spatter at the
crime scene. Do you see White’s
winning plan?

Answer: Step 1: Trade off the f4-

bishop for the b8-knight, which would
otherwise be activated via c6. The
exchange on b8 also gains a tempo for
White, since Black’s king is moved one
step further away from the critical g6- 35...Kd7? is too slow. White wins a
square. pawn after 36.hxg6 hxg6 37.Nh4.

33.Bxb8!! 36.Kg3

The only winning move, planned by Step 3: White’s king walks over and
Gelfand a few moves before entering chops the h5-pawn.
this endgame. The plan for now exists
only within a dark, dusty corner of 36...Bg7 37.Kh4 Kd6 38.Kxh5 Bf8
Gelfand’s mind.
38...Ke7 39.Ne5 Kf8 40.g6! is similar
33...Kxb8 34.h4! to the game continuation.

Step 2: Push the h-pawn to h5, since 39.Ne5

this is White’s only viable break.
Also strong is the immediate 39.g6!.
34...Kc7 35.h5
8
*+*+*V*+
8
*+*+*+*V
7
+*+*+*+o
7
+*L*+*+o
6
*O*Lo+*+
6
*O*+o+o+
5
O*+oNoPk
5
O*+o+oPp
4 p+*P*+*+
4 p+*P*+*+
3 +p+*P*+*
3 +p+*Pn+*
2 *+*+*P*+
2 *+*+*Pk+
1 +*+*+*+*
a b c d e f g h
1 +*+*+*+*
a b c d e f g h

White to move

Black to move Black’s problems calcify into the

category of incurable, due to the
35...gxh5 following reasons:
1) White’s nimble knight is far
Black has no choice but to take on h5,
stronger than Black’s stationary bishop
since otherwise g6 can’t be protected.
because of the static nature of the
structure. Step 4: Break through on g6.
2) Frustratingly for Black, any swap of
bishop for knight leads to a lost king and 43...hxg6+
pawn ending.
Of course, this lets White’s king in.
3) White has a massive king position.
However, pushing to h6 also fails,
4) Black’s h-pawn is weak. This since the h-pawn falls and Black’s
means that when White breaks through desperate attempt at a fortress has no
with g5-g6, Black cannot afford to chance of staying upright. After 43...h6
bypass the break with ...h7-h6, since 44.Nf7 Kf6 45.Nxh6 Gelfand notes that
then White will win the h-pawn with the correct square for the knight is h5,
Nf7 and Nxh6. Thus, when Black swaps so: 45...Ke7 46.Nf7 Kd7 47.Kg5 Ke7
on g6, White’s king will recapture, 48.Ne5 Kd6 49.Nf3 Ke7 50.Ng1! Kd6
decisively infiltrating the black camp. 51.Ne2 Ke7 52.Ng3 Kf8 53.Nh5.
5) The white king and knight will White wins.
eventually muscle out Black’s king, and
in the end, the base of the pawn chain on 44.Kxg6 Bh8
e6 is doomed to fall.
After 44...Bf8 45.Nf3 Kd7 46.Kf7
39...Bg7 40.Nd3 Ke7 41.Ne5 Kd6 Bb4 47.Ng5 Bd2 48.Nxe6 Bxe3 49.Kf6
42.f4! Ke7 43.g6! White wins the f5-pawn and the game.

45.Nf7 Bf6 46.Ng5 Bh8 47.Nh7

8
*+*+*+*+ Gelfand starts playing against the
7
+*+*L*Vo restricted black bishop.
6
*O*+o+p+
5
O*+oNo+k 47...Ke8 48.Nf6+! Kf8
4 p+*P*P*+
3 +p+*P*+*
2 *+*+*+*+
1 +*+*+*+*
a b c d e f g h

Black to move
 Gelfand had to foresee that Black
8
*+*+*L*V couldn’t seal the position with a fortress
7
+*+*+*+* draw. However, for a player of his

*O*+oNk+
6 stature, it wasn’t hard to understand that
for Black playing with such a terrible
5
O*+o+o+* bishop shouldn’t be viable. The only
4 p+*P*P*+ thing he needed to take care of was
3 +p+*P*+* opening a path for his pieces to enter the
black camp, but the quick h3-h4-h5
2 *+*+*+*+ ensured such a path would always exist.
1 +*+*+*+*
a b c d e f g h Game 17
Dmitry Jakovenko – Laurent
Fressinet
White to move Plovdiv 2012

49.Nh5!

Keep in mind that Turtle took his

8
t+*+*+l+
sweet time, but won the race anyway.
7
Oo+*T*Oo
The black bishop’s status has been 6
*+*+oM*+
reduced to that of a beast of burden. It 5
+*+*+*+*
has run out of moves. 4 *+b+wP*+
49...Kg8 50.Kh6 1-0 3 +q+*P*+*
We are reminded of the point in the
2 p+*+*+pP
movie when Miss Scarlett’s ancestral
1 +*Rr+*K*
home burned to the ground. Zugzwang a b c d e f g h
sets in. White’s king first gains entry to
h7 and eventually to e7, which in turn
means that the base e6-pawn falls. To White to move
illustrate: 50...Kf7 51.Kh7 Bf6 52.Nxf6
Kxf6 53.Kg8 Ke7 54.Kg7 Ke8 55.Kf6 This game is another one that doesn’t
Kd7 56.Kf7 Kd6 57.Ke8 Kc6 58.Ke7 fit neatly into a single chapter. White did
wins. indeed own the superior minor piece,
33.Bxb8!! was a farsighted move. but he also won by piling up on two
weak black isolated pawns, the sum of notoriously difficult to win for the side
which was too much for the defense to having the extra pawn.
handle. In that regard, it could have also
easily been part of the previous chapter, 26.Bc4!
or even the first one, regarding the
The bishop returns to its best diagonal.
principle of two (or more) weaknesses.
White can even pretend everything was
Black experiences serious difficulties:
planned and that no mistakes were
1) The pinned e6-pawn, an oozing made, as this also stops the black queen
wound, is a clear target for White’s from going back to e4.
pieces.
2) White’s bishop is clearly the 26...Qxb3
superior minor piece over Black’s
knight. Sadly for Black, this is forced. The
alternative 26...Qd6 27.Bxd5 exd5
3) White leads in development, with
28.Qxd5+ Qxd5 29.Rxd5 is dead lost
rooks already posted on open files.
for Black, since White’s rooks control
24.Be2?! both open files, and the 4-2 kingside
majority will eventually advance, with
Jakovenko’s attempt to switch the decisive effect. In stark contrast to the
bishop’s diagonal is unjustified. The note above regarding 25...Qg6!, the
bishop was stronger on the a2-g8 queens have been traded, so there is no
diagonal, as Black’s main weakness is counterplay against the white king, who
the e6-pawn. will go to f3 and help push the kingside
Most accurate is: 24.Rd4! Qf5 25.Rd6 armada forward.
Rae8 26.Bb5 Rf8 27.Bd3! Qg4 28.h3!
when Black’s queen is driven away from 27.Bxb3 Nc7
her protection of the e6-pawn, which
falls on the next move.

24...Nd5 25.Kf2 Qb4?!

Black’s best defensive hope is to

induce White to exchange bishop for
knight, even if it gives up a pawn:
25...Qg6! 26.Bf3 Rae8!. White can win
a pawn by chopping the knight.
However, major piece endgames are
 8
t+*+*+l+ 8
t+*+*+*+
7
OoM*T*Oo 7
OoM*T*Lo
6
*+*+o+*+ 6
*+*+o+o+
5
+*+*+*+* 5
+*+*R*+p
4 *+*+*P*+ 4 *+*+*P*+
3 +b+*P*+* 3 +b+*P*+*
2 p+*+*KpP 2 p+*+*Kp+
1 +*Rr+*+* 1 +*+r+*+*
a b c d e f g h a b c d e f g h

White to move Black to move

28.Rc5! 31...Rf8

The rook directs itself towards e5, No solace is found in 31...gxh5, as

aiming to pile up the pressure on the e6- after 32.Rxh5 the h7-pawn looks to be as
pawn. weak or even weaker than a pawn on g6
would be. White can actually win it by
28...Kf8 29.Re5 g6?! force quite easily: for example, 32...b5
33.Kf3 a5 34.Bc2 Rh8 35.Rdh1 h6
This only makes matters worse, since
36.g4 and the coming g4-g5 wins.
it weakens the kingside dark squares.
Simultaneously, the pawn on g6 acts as 32.g4
a ‘hook’ White can latch on to, inviting
a pawn break on h5 that would create Why not gain even more space?
even more weaknesses.
32...Rff7 33.Rg5 Rd7?
30.h4! Kg7 31.h5!
To blunder in an already rotten
position is a hell within a hell. This is
Black’s “Why is this happening to me?”
moment.
Black had to try 33...Rf6 and then
pray, but he should have eventually lost
anyway. recaptures Black’s rook on d1 and
remains up a clean pawn, with a superior
8
*+*+*+*+ minor piece. Black only makes matters
7
OoMt+tLo worse and hangs a second pawn if he
tries 36...Kh7? which allows 37.Rxe6+
6
*+*+o+o+ Kg7 38.Rg6+ Kh7 39.Rg5+ Kh8
5
+*+*+*Rp 40.Bxd1.
4 *+*+*Pp+ 36.Rxg6 Rxd1 37.Bxd1 Rh7 38.Bb3
3 +b+*P*+*
2 p+*+*K*+ White’s queenside pawn is secure.
1 +*+r+*+* 38...Ke7
a b c d e f g h

8
*+*+*+*+
White to move
7
OoM*L*+t
6
*+*+o+r+
Exercise: Fressinet’s last move was
a mistake in an already lost
5
+*+*+*+*
position. Work out a way for White
4 *+*+*Pp+
to win a pawn by force: 3 +b+*P*+*
2 p+*+*K*+
Answer: Step 1: Exchange pawns on 1 +*+*+*+*
g6. a b c d e f g h
34.hxg6 hxg6

Step 2: Shift the bishop to c2, after White to move

which Black’s sickly g6-pawn falls.
39.f5!
35.Bc2! Kf8
The e6-pawn is going to be defended
35...Rxd1 can be met with 36.Rxg6+!. as many times as White can attack it. So,
Step 3: Zwischenzug. The store’s sign White’s best option is to simply swap it
reads: “Shoplifting is a crime. We off and get two connected passed pawns.
prosecute!” Black’s king must move to
f8, after which White leisurely 39...exf5 40.gxf5 Ne8
 The attempt to blockade White’s 8.cxd4 Bb4+ 9.Bd2 Bxd2+ 10.Qxd2 0-
pawns on the dark squares will fail. 0 11.Bc4 Nd7 12.0-0 b6 13.a4!?

41.e4 Nd6 42.f6+! 1-0

8
t+vW*Tl+
O*+m+oOo
White’s pawns will roar forward, and 7
Fressinet decided to call it a day. Three
factors cost Black the game:
6
*O*+o+*+
1) He had an inferior knight against 5
+*+*+*+*
White’s powerful bishop. 4 p+bPp+*+
2) He had a weak, isolated pawn on 3 +*+*+n+*
*+*Q*PpP
e6. 2
3) Eventually, after 29...g6?!, he got
an even weaker isolated pawn on g6.
1 R*+*+rK*
a b c d e f g h
I’ve learned from social media that
pretty much everyone claims to be
“special”. But how can this be so since,
Black to move
statistically, 49% of the population is
below average? On the other hand, I can One of many moves for White. The
claim with some degree of confidence idea is to play a4-a5 and exchange on
that Magnus Carlsen is that “special” b6, handing Black an isolated pawn
once-in-a-lifetime player, who soars which can later morph into a weakness.
above the rest of his era. A week before On the other hand, Black’s main source
I annotated this game, Carlsen played of counterplay in these positions is
Kramnik in a Titled Tuesday and creating a passed pawn on the
clobbered him by trolling with 1...h6. queenside, so White should be careful
Needless to say, Kramnik didn’t take it not to do Black’s work for him.
well and swore that he would never play
in the tournament again! 13...Bb7 14.Bd3 e5
Game 18
In a blitz game between the two earlier
Magnus Carlsen – Vladimir
in the year, Kramnik tried 14...Rb8.
Kramnik
Paris (rapid) 2016 15.Qe3!

1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.cxd5 Carlsen recognizes that a connected

Nxd5 5.e4 Nxc3 6.bxc3 c5 7.Nf3 cxd4 passed pawn on d5 won’t get him any
advantage after 15.d5 Nc5 16.Qe3 leaving Black with an isolated pawn on
Nxd3 17.Qxd3 f5!. Black successfully the queenside. However, the pawn is
undermines the white pawn chain and also passed. Is it a strength or a
gets good active counterplay. weakness? Can it advance? Concrete
analysis is needed to answer these
15...exd4 16.Nxd4 Nc5 17.Bc2 Re8 questions.

19...Rad8 20.Rfd1 Ba6 21.axb6 axb6

8
t+*Wt+l+
7
Ov+*+oOo
6
*O*+*+*+
8
*+*Tt+l+
5
+*M*+*+*
7
+*+*+oOo
4 p+*Np+*+
6
vO*+*W*+
3 +*+*Q*+*
5
+*M*+*+*
2 *+b+*PpP
4 *+*Np+*+
1 R*+*+rK*
3 +*+*Qp+*
a b c d e f g h
2 *+b+*+pP
1 R*+r+*K*
a b c d e f g h
White to move

18.f3! White to move

Many would find this to be too 22.Ba4!
passive, but that’s not the correct way to
look at it. Activity in chess is a two-way A creative way to get rid of the strong
street, and restricting the opponent is as knight on c5.
good or sometimes even better than
increasing the activity of our own 22...Nxa4
pieces. Solidifying the pawn on e4
Black had no choice since White takes
restricts Black’s knight, rook and bishop
over the initiative after the passive
all at the same time.
22...Rf8?! 23.e5!.
18...Qf6 19.a5!
23.Rxa4 b5
Magnus fulfills his minority attack,
 The pawn is pushed, but looking at the This natural move may not be best,
bishop on a6 leaves a bitter aftertaste. since it allows White’s e-pawn to move
No better for Black is 23...Bb7 forward, with tempo. But the question
24.Rb4. then is: what plan does Black have,
except to wait?
24.Rb4 Rd7 25.Rc1
26.e5!
The engine assesses this as dead even.
As you may have guessed, your human White gains space, while worrying his
writer doesn’t believe the assessment, opponent about sudden pushes to e6.
for the following reasons:
26...Qb6 27.Rc6! Qa7 28.Nf5!
1) Black’s isolated b-pawn is
vulnerable. The knight was unstable on d4, so
2) White controls the center with his Magnus offers to swap queens.
pawns and pieces.
3) White controls the open c-file. 28...Qxe3+
4) Most importantly, White’s knight is
Tossing in checks doesn’t endanger
clearly superior to Black’s hemmed-in
White’s king in the slightest after
bishop.
28...Rd1+ 29.Kf2 R1d2+ 30.Kg3.
8
*+*+t+l+ 29.Nxe3
7
+*+t+oOo
6
v+*+*W*+ 8
*+*T*+l+
5
+o+*+*+* 7
+*+t+oOo
4 *R*Np+*+ 6
v+r+*+*+
3 +*+*Qp+* 5
+o+*P*+*
2 *+*+*+pP 4 *R*+*+*+
1 +*R*+*K* 3 +*+*Np+*
a b c d e f g h 2 *+*+*+pP
1 +*+*+*K*
a b c d e f g h
Black to move

25...Red8
 Black to move

Black to move
8
t+*+*+l+
7
+*+t+o+o
The engine’s assessment remains dead 6
v+r+*+o+
even and in an engine versus engine
game, this may be true. But a human
5
+o+*P*+*
must still overcome obstacles to hold a
4 *R*+*+*P
draw for Black, who has two 3 +*+*Np+*
weaknesses in the shape of the bad 2 *+*+*Kp+
+*+*+*+*
bishop on a6 and the weak pawn on b5, 1
while White’s position doesn’t have any
real defects.
a b c d e f g h

29...Ra8 30.Kf2 g6!?

This weakens the dark squares around Black to move

the black king and makes the bishop
31...h5?!
even worse.
Kramnik probably wanted to stop the This natural reaction might be what
knight from reaching d6, since after pushes Black’s position into becoming
30...h6 31.Nf5 Bc8 32.Nd6 the b-pawn truly unpleasant.
will fall, and Black will have to suffer in Kramnik had better chances to hold
order to gain the half-point. the game by immediately giving up the
b-pawn: 31...Rd2+ 32.Kg3 Bc8! 33.Rc7
31.h4!
Be6!. Amputation of a mangled limb is
Now Black’s worries need to include never an easy decision for either doctor
h4-h5-h6. or patient. Yet in this case, activating the
bishop is more important than holding
on to the sickly b-pawn. Black has better
chances to hold here than in the game
continuation, as the bishop firmly
protects the kingside and the black rooks
are free to create counterplay against the
untethered knight on e3 and the g2-
pawn.

32.e6!
 Magnus recognizes that g6 is a 38...Rxe4 39.Nf6+ Kg7 40.Nxd7 Re7!
weakness in Black’s position, and it is
harder to hold on to g6 than it was to What? Did Kramnik just hang a piece?
hold on to f7. Actually not, and his last move is a
brilliant idea which should hold the
32...fxe6 33.Rxe6 Kh7 34.Rf4 draw. White has only one computer-like
move which offers him practical
Starting with 34.Rb6! is stronger, since chances to win. After a long think,
it awkwardly ties Black’s clunky bishop Carlsen found it.
to the b-pawn.
41.Nb8!!
34...Bc8 35.Rb6 Ra2+ 36.Kg3 Re2
37.Re4 Bb7 Carlsen finds his only practical chance
to win. Wild, imaginative ideas are often
defeated by the details of reality. Not
8
*+*+*+*+ here though. This disorienting idea
7
+v+t+*+l actually works. I was watching this
6
*R*+*+o+ game live, and this was the moment that
my mouth hung open in shock, like a
5
+o+*+*+o disoriented trout on a hook.
4 *+*+r+*P 41.Nc5?! is tempting, yet not best
3 +*+*NpK* since it allows Black to hold a draw with
2 *+*+t+p+ 41...Bc8 42.Rxb5 Re5!. Suddenly White

+*+*+*+*
1 is caught in an unpleasant pin. The only
way to break it is to trade rooks, which
a b c d e f g h allows Black to hold a draw: 43.Ra5
Kh6! 44.Nb3 Rxa5 45.Nxa5 g5! and the
reduced number of pawns severely
White to move reduces White’s winning chances.
Even worse is 41.Rxb7?!, since after
38.Nd5! 41...b4! White will need to give back the
knight on d7 and enter a 3-2 rook
I like this simplifying idea, which robs
endgame which, unlike having a knight
Black of potential counterplay against
and a bishop respectively, is dead drawn
White’s g2-pawn. After the trade of one
at this level. Of course, 42.Nc5?? Rxb7
pair of rooks, the only thing that remains
43.Nxb7 b3 is a tragicomedy.
for Black is to sit and wait.
41...Rc7 42.Rxb5
 The price of winning a pawn is that doesn’t cut it in a binary choice
White’s knight is weirdly trapped on b8. situation.
How will Magnus extricate the knight Correct was: 49...Rb4! 50.Nd7 Bc8
from its current imprisonment? Watch. 51.Rb6 Rd4 (nyet, no rook swap!)
52.Ne5 Bf5. Of course, White has
42...Kf6 43.Rb2 Kf5 44.Rb6 Re7 decent practical chances to win,
45.Rb5+ Kf6 46.Kf4 Rc7 47.Rb6+ Kg7 especially with Carlsen playing White.
Yet from an objective point of view,

*N*+*+*+
8 Black should hold the draw with perfect
play.
7
+vT*+*L*
6
*R*+*+o+ 50.Rd7+! Rxd7 51.Nxd7
5
+*+*+*+o
4 *+*+*K*P 8
*+*+*+*+
3 +*+*+p+* 7
+v+n+*L*
2 *+*+*+p+ 6
*+*+*+o+
1 +*+*+*+* 5
+*+*+*+o
a b c d e f g h 4 *+*+*+*P
3 +*+*+pK*
2 *+*+*+p+
White to move 1 +*+*+*+*
48.Rd6! a b c d e f g h

The d7-square is the knight’s path

towards freedom.
Black to move
48...Rc4+ 49.Kg3 Rc7?
How can this possibly be a win for
Sometimes in a chess game, we are White, with so few pawns remaining?
dead even though we don’t know it. This 1) Black’s pawns will be eternally
logical yet incorrect move, protecting fixed on their present squares.
the seventh rank, is losing. By allowing 2) Black is weak on the kingside dark
a rook swap, Kramnik seals his fate. squares.
Black must keep rooks on the board to 3) Black will lose if White’s king can
have a chance to hold a draw. “Close” reach g5, as then f3-f4-f5 will isolate the
h5-pawn, and the knight can go grab it.
Carlsen, with his almost infallible 8
*+*+*+*+
endgame intuition, foresaw that Black 7
+*+*+*L*
*+*+*+o+
had no way to stop the plan of White’s 6
king seizing control over g5.
5
+*+*+*Ko
51...Ba6 52.Nc5 Bf1 4 *+vN*+*P
Kramnik attempts to tie White down
3 +*+*+pP*
to defense of g2. The problem is that the 2 *+*+*+*+
pressure is only temporary. 1 +*+*+*+*
53.Ne4 Kf7 54.Nd2 a b c d e f g h

Out!

54...Bd3 55.Kf4 Ke6 White to move

55...Kf6 56.Ne4+! leads to the 60.f4!

winning plan used in the game.
Kramnik is confronted by a grim
56.Kg5 finality, and he is helpless against the
upcoming plan.
Step 1 of White’s winning plan is
accomplished. 60...Bd3 61.f5! gxf5 62.Nxf5+!

56...Kf7 57.Nb3 Bf1 58.g3 Kg7 Certainly not 62.Kxh5?? Kf6 when
59.Nd4 Bc4 Black should hold the draw.

62...Kh7

The king and pawn ending is dead lost

for Black as well. After 62...Bxf5
63.Kxf5 Kh6 64.Kf6 Kh7 65.Kg5
White wins.

63.Ne3 1-0

The h5-pawn falls when White’s

knight reaches either f4 or f6, so Black
resigned. writes Giri. Each week, mean-spirited
When people announce that friends post some version of a London-
Capablanca or Fischer were the greatest System-Sucks meme on my social
endgame players of all time, I shake my media. I’m thinking of spearheading a
head in pity, snort disdainfully, and class-action suit.
inform them that they believe in false
prophets. It’s not even close. Carlsen, at 3...c5 4.e3 Nc6 5.Nbd2 Bg4 6.c3 e6
least in that regard, is in the number one

t+*WlV*T
spot of all time. 8
This was a perfect example of the
Principle: Don’t allow the opponent to
7
Oo+*+oOo
fix your structure on the same color as
6
*+m+oM*+
your remaining bishop, since those 5
+*Oo+*+*
pawns will become targets and you will
suffer on the opposite color of your
4 *+*P*Bv+
bishop. Even with as few as three pawns
3 +*P*Pn+*
versus two on the same side, Black still 2 pP*N*PpP
wasn’t able to hold, once rooks were 1 R*+qKb+r
removed from the board. a b c d e f g h
The next game is an example of a
piece that is not trapped, but is
perpetually offside, which negatively
affects Tari’s position. Black’s knight White to move
was stuck on h5 for a very long time,
7.Qb3
unable to participate in the game.
Game 19 This move is why so many players
Anish Giri – Aryan Tari choose to keep Black’s light-squared
Stavanger 2022 bishop inside the pawn chain. Black
must contort to defend b7.
1.Nf3 d5 2.d4 Nf6 3.Bf4
7...Qc8 8.h3 Bh5 9.Be2 Be7
You can’t see me, but my hands are
reverently folded in a Namaste prayer Giri had previously reached this
whenever White brings this bishop to f4. position against Wojtaszek and Gelfand.
“The London System has not become In both games he castled kingside. This
entirely mainstream but is no longer that time he deviates.
rare of a guest at the highest level,”
10.dxc5!? 11...Bg6 12.Nh4 Qd7

This way, Black is deprived of the Black is rightfully nervous about

...c5-c4 mechanism. castling kingside, since then White has a
ready-made pawn storm.
10...Bxc5
13.Nxg6 hxg6 14.g5!

8
t+w+l+*T An idea which has been incubating
7
Oo+*+oOo now takes birth. Giri gives his opponent

*+m+oM*+
6 a difficult choice: head back to g8,
losing time, or move the knight to h5,
5
+*Vo+*+v where it is clearly offside.
4 *+*+*B*+
+qP*Pn+p
3 14...Nh5

2 pP*NbPp+ This fateful decision is the no-frills

1 R*+*K*+r version. The knight is clearly out of play
on h5, yet it’s still likely the correct
a b c d e f g h
decision since the retreat to g8 loses
time.
White looks clearly better after
White to move 14...Ng8!? 15.0-0-0. His ideas include
pawn breaks on c4 and e4, or initiating
11.g4! piece play with Ne4. Both positions are
unpleasant for Black – defending one or
“Believe what I tell you to believe!”
the other is partly a matter of taste.
orders Big Brother opening theory.
Fortunately, there are those who don’t 15.Bh2
listen and produce new ideas. Giri’s idea
is to chase down the bishop pair, as well
as gain useful kingside space. The
danger of such a move is the potential
for future overextension. White’s king
suddenly doesn’t look so safe on the
kingside, so he will likely castle long,
which is also full of danger, since
Black’s c-file is open on that wing.
 most areas of the board, and that is the
8
t+*+l+*T most influential of the factors stated
7
Oo+w+oO* above.
6
*+m+o+o+ 15...a6
5
+*Vo+*Pm
*+*+*+*+
4 15...Bd6?! is actually an exchange
sacrifice: 16.Bxd6 Qxd6 17.Qxb7 (the
3 +qP*P*+p knight on c6 is unstable and Bb5 is a
2 pP*NbP*B threat) 17...Rb8 18.Qa6 Rxb2 19.Nb3 0-
1 R*+*K*+r 0 20.0-0-0!. White wins the exchange,
but Black must have some compensation
a b c d e f g h
due to the extremely exposed white
king. Then we look at the knight on h5,
and a deflating feeling sinks in: Black’s
Black to move attack has almost no chances to succeed,
as he’s effectively attacking without a
We sense that something is wrong in full rook.
Black’s position. In fact, it’s several
somethings. Giri’s creative opening play 16.0-0-0 b5 17.Kb1 Bd6?!
earned him an advantage. Let’s gather
data: How awful when we follow a
1) Black’s knight on h5, a portrait of principle, and the move is still wrong!
agitation, while perfectly safe there, is Black’s last move follows the Principle:
simultaneously out of play and unable to If the opponent owns the bishop pair,
enter the game. If the game opens in the then swap one of the bishops off. The
center later on, Black will miss the trouble is that the move is mistimed.
offside piece. A more accurate move order is
2) White owns the bishop pair, and the 17...Na5 18.Qc2 Bd6.
structure is fluid. 18.Bxd6 Qxd6
3) Black controls a greater share of the
center.
4) Black will possibly have the safer
king after castling kingside.
Conclusion: The position is a mess,
yet White is better, as the knight being
on h5 means having an extra piece in
 grabbing a pawn with 21.Ne4? Qe7!
8
t+*+l+*T 22.Qxd5, which allows Black to escape
7
+*+*+oO* via a clever repetition: 22...Nb4 23.Qd4
Nc6 24.Qd5 Nb4, etc.
6
o+mWo+o+
5
+o+o+*Pm 21...Rad8 22.h4 Rfe8 23.Nd4 Ne5
4 *+*+*+*+
3 +qP*P*+p 8
*+*Tt+l+
2 pP*NbP*+ 7
+*+*+oO*
1 +k+r+*+r 6
o+*W*+o+
a b c d e f g h 5
+o+oM*Pm
4 *+*N*+*P
3 +q+*P*+*
pP*+bP*+
White to move 2
19.c4! 1 +k+r+*+r
a b c d e f g h
Principle: Open the position and create
confrontation when leading in
development. “I was very happy with
this idea, which occurred to me while White to move
my opponent was contemplating
17...Bd6. All tactics work for me, and 24.Qa3?
opening the position only highlights the
Giri reasons that he will be up a piece
fact that the knight on h5 is out of play,”
(the offside h5-knight) if the queens
writes Giri.
come off the board. Yet, this is too large
19...0-0 a concession.
A better approach was 24.a3! Nc4
Tari accepts a weakness on d5. After 25.Ka2, safeguarding the king. An
19...Na5 20.Qc3 Nxc4 21.Nxc4 bxc4 illustrative line could be: 25...Rb8
22.Qxc4 0-0 23.h4 Black is in deep 26.Nc2! (covering b4 against Black’s
trouble due to the offside knight. intended push, while opening an attack
on the d5-pawn) 26...a5 27.Bf3 Re5
20.cxd5 exd5 21.Nf3! 28.Rd3. The d5-pawn will soon fall,
adding a missing pawn to the list of
Giri correctly resists the temptation of
Black’s worries. White’s advantage is Black should be doing everything in
decisive. his power to emancipate the h5-knight.
Therefore, best is 25...f5!, intending to
24...Qb6? break free with ...f5-f4 at some point if
White refuses to win a pawn with Nxf5.
Tari misses the promising opportunity
If White grabs the pawn, then the h5-
of entering the endgame with 24...Qxa3!
knight is eliminated, and Black couldn’t
25.bxa3 and then playing 25...f5!,
be happier.
introducing the idea of ...f5-f4, which
would extricate the knight from h5. 26.Qb2 Qd6 27.Rc1
Black’s position remains unpleasant,
and White could grab a pawn with Every swap benefits White. Exactly
26.Nxf5 if he wants, but that eliminates like in positions with a lot of extra
the h5-knight, and Black has enough material, simplification will highlight
positional compensation to hope for a the absence of Black’s knight from the
draw. main part of the battlefield.

25.b3!? 27...Qd7 28.Rxc8 Rxc8 29.Qd2

Black’s e5-knight is denied use of c4. Giri is worried about a second rook
trade, fearing counterplay after Black’s
8
*+*Tt+l+ queen infiltration to h3.
7
+*+*+oO* 29...Ng4
6
oW*+*+o+
5
+o+oM*Pm
4 *+*N*+*P
8
*+t+*+l+
3 Qp+*P*+*
7
+*+w+oO*
2 p+*+bP*+
6
o+*+*+o+
1 +k+r+*+r
5
+o+o+*Pm
a b c d e f g h
4 *+*N*+mP
3 +p+*P*+*
2 p+*QbP*+
Black to move
1 +k+*+*+r
a b c d e f g h
25...Rc8?!
 White to move fears aren’t real, except within the mind,
where the terror is very real. “I didn’t
Black’s knights look a bit annoying, know how to further improve my
but their activity is temporary – they’re position, but actually I had to let go of
fighting an uneven battle. the h2-square control and reposition my
rook,” writes Giri. We can see from the
30.Bf1
engine analysis that this is White’s best
30.Bf3 is more natural. plan, yet for a human to allow Black’s
queen to infiltrate on h2 is
30...Qe7?! psychologically difficult.

Evading the coming pin with Bh3 35.Bd1

while threatening the cheapo ...Qe4+
followed by ...Qxh1, but 30...f5! was Stronger is 35.Rb1! Qh2 36.Rb2!
still best. Qxh4 37.Ne6!, followed by taking the
pawn on d5 and winning due to the
31.Bg2 Qe5 32.a3 Kh7 33.Ka2 f5! atrocious temporary placement of the
34.Bf3 Qd6 black queen and knights.

35...Re8 36.Bf3
8
*+t+*+*+
+*+*+*Ol
7 Giri drifts, unable to find something
concrete.
6
o+*W*+o+
5
+o+o+oPm 36...Rc8 37.Qe2 Rf8 38.b4!?
4 *+*N*+mP White toys with a future Nb3 and
3 Pp+*Pb+* Nc5.
2 k+*Q*P*+ The strongest plan is still to activate
1 +*+*+*+r the rook, even though this abandons the
h-pawn and allows Black’s queen to
a b c d e f g h
enter h2. For example: 38.Rc1! Qh2
39.Rc2! Qxh4 40.Ne6 Re8 41.Bxd5 and
White wins.
White to move
38...Qe5 39.Qd2
When we think about it, we come to
the conclusion that the majority of our 39.Qc2!, seizing the open file, is
stronger.
39...Rc8 40.Nb3 Rc4 41.Rd1 f4!? draw offer since his opponent was low
on the clock.

8
*+*+*+*+ 44.Kb1 Qf7 45.Qe2
7
+*+*+*Ol
6
o+*+*+o+ 8
*+*+*+*+
5
+o+oW*Pm 7
+*+*+wOl
4 *Pt+*OmP 6
o+*+*+o+
3 Pn+*Pb+* 5
+o+o+*Pm
2 k+*Q*P*+ 4 *P*N*TmP
1 +*+r+*+* 3 P*+*+b+*
a b c d e f g h 2 *+*+qP*+
1 +k+r+*+*
a b c d e f g h
White to move

This is a bluff that ends up confusing

Giri. The idea is that the g4-knight isn’t Black to move
really hanging due to the discovered
attack ...f4xe3. 45...Nh2!

42.exf4? The pressure down the f-file forces

White to make a concession.
Black’s h5-knight begins to see
daylight. 46.Bxh5
42.Na5! wins material since Black’s
This must have been a satisfying
rook and g4-knight hang
moment for Tari, since his h5-knight has
simultaneously: 42...fxe3 43.Nxc4! is
been vindicated.
the key point. Black’s queen is
threatened and after 43...dxc4 44.fxe3 he 46...gxh5 47.Qd3+?
doesn’t have enough for the exchange.
47.Qc2+ was the correct check.
42...Rxf4 43.Nd4 Qc7
47...Qg6!
Giri realized he had thrown away his
advantage, yet he still refused Tari’s The advantage has miraculously
swung to Black. fork available on e5. White cannot hold
the g-pawn.
48.Ne6 Rxf2?
58.Kc1!
48...Rf5! leaves Black in control. If the
queen were on c2, then White would The g-pawn can no longer be halted.
have Rxd5. That’s why checking from
d3 was mistaken. 58...Ne4 59.Re7! Nc3 60.Rxe2 Nxe2+
61.Kd2 1-0
49.Qxg6+ Kxg6 50.Rxd5 Kf7 51.Nc7
Suddenly, Giri is so rich that he has
Now White is OK again. The game accumulated more money and assets
threatens to turn into an opposing-wings than he could ever spend in a lifetime,
promotion race. mainly because he is promoting to a new
queen.
51...Nf3 52.Rd7+ Kg6 53.Ne8! In this game, Giri masterfully
displayed his strategic advantage by
The g7-pawn falls.
imprisoning his opponent’s knight on
53...Re2?! the rim at h5. Time pressure began to
alter the natural flow, and the game
53...Rg2 was better, watching over devolved into chaos in its later stages.
White’s g-pawn. However, it is not only time pressure
that led to problems in converting the
54.Rxg7+ Kf5 55.Rf7+ Kg4 56.Nf6+ advantage. Like most humans, when
Kxh4 57.g6 owning something as valuable as
managing to imprison an enemy piece,
This is pretty scary: White threatens to
Giri was too reluctant to let it go. The
promote. Black still has a miracle
static advantage often has to be
defense.
transformed into an attack if it is to
57...Nd2+? succeed, and that means relinquishing
some control. Constant denial of doing
Tari didn’t have enough time to find so keeps the game going longer than it
the drawing idea, which is to force should, and that inevitably creates
White’s king to d3: 57...Re1+! 58.Kc2 practical chances for the defender. In
Re2+! 59.Kd3 (if the king refuses to that regard, the lines showing a more
move to d3, Black’s rook then gives concrete approach (24.a3!, 35.Rb1!, and
endless checks) 59...Rg2! when Black is 38.Rc1!) are especially instructive.
threatening ...Rxg6, while also having a Round 5 of the Tata Steel tournament
in 2024 turned out to be historic for
women’s chess. Shocks arrived in 8
tMvWl+mT
stereo, with two massive upsets. First, 7
OoO*+oOo
*+*V*+*+
Dutch Women’s Champion Eline 6
Roebers won her first game by taking
down 2692-rated GM Hans Niemann,
5
+*+o+*+*
the highest-rated player she had ever 4 *+pP*+*+
faced. Then, the reigning women’s
world champion, Ju Wenjun, defeated
3 +*+*+n+*
2759-rated GM Alireza Firouzja.
2 pP*+*PpP
Game 20
1 RnBqKb+r
Ju Wenjun – Alireza Firouzja a b c d e f g h
Wijk aan Zee 2024

1.d4 e6 2.e4 d5 3.exd5

Black to move
If you are a fan of the chess version of
Do we sense some ambition? White is
the sex, drugs, and rock and roll
first to break the symmetry!
lifestyle, then the puritanical Exchange
French won’t appeal to you. As a 5...Nf6 6.c5!?
lifelong French Defense player, I always
fight the urge to strangle my lower-rated White gains a tempo on the bishop
opponents when they play the hated while grabbing some space on the
Exchange Variation. In this case, it queenside. On the downside, the central
might actually be an excellent choice for tension is released, and Black gains the
the lower-rated player who faces a fiery option of chipping away at White’s
tactician. She essentially announces, “I structure with ...b7-b6.
am happy to draw. You will need to take
risks to beat me.” Moreover, the 6...Be7 7.Bd3 0-0 8.0-0 Bg4
simpler, more positional nature of these
8...b6! is more accurate.
lines can allow Ju Wenjun to highlight
the strengths in her game. 9.h3 Bh5 10.Be3
3...exd5 4.Nf3 Bd6 5.c4!? White keeps her options open
regarding the development of the b1-
knight, which can often choose d2
instead of c3, protecting its counterpart
on f3 and thus relieving any pressure felt 12.Bb5!?
by the pin.
Ju is willing to hand over one of her
8
tM*W*Tl+ bishops and relinquish some light-square
control if it means inflicting a weakness
7
OoO*VoOo on the black structure. Furthermore, by
6
*+*+*M*+ eliminating the c6-knight, she could get
5
+*Po+*+v the chance to jump forward with Nf3-
4 *+*P*+*+ e5. Hence, it’s imperative for Firouzja to
react in an active, dynamic fashion.
3 +*+bBn+p
2 pP*+*Pp+ 12...Ne4! 13.Qa4!?
1 Rn+q+rK* Also possible was 13.g4!? Bg6
a b c d e f g h 14.Bxc6 bxc6 15.Ne5, which the engine
claims is dynamically balanced.

13...Nxd2 14.Nxd2
Black to move

t+*Wt+l+
10...Nc6!? 8
The fundamentalist believes that every 7
OoO*VoOo
word of their scripture is truth and that 6
*+m+*+*+
+bPo+*+v
all religions – except theirs – are false. 5
Firouzja is a player who believes in his
DNA, that activity and dynamic forces
4 q+*P*+*+
supersede structural weakness. He is 3 +*+*B*+p
unafraid of White playing Bb5 and 2 pP*N*Pp+
R*+*+rK*
Bxc6, inflicting damage to his structure, 1
since in return Black receives the bishop
pair and enhanced light-square power. a b c d e f g h
10...c6 is a safer alternative.

11.Nbd2 Re8!? Black to move

Firouzja is not the type to take time 14...Bg5!
out for a prophylactic measure like
11...a6. It’s crucial for Black to continue
creating threats. If White is allowed to White has the better structure and
consolidate, she would be guaranteed a potentially a better minor piece, as the
permanent positional edge. bishop on h5 looks useless, targeting
empty and irrelevant squares. Firouzja is
15.Bxc6 bxc6 under pressure to create quick
counterplay, and his move is logical,
White has handed Black a pawn
aiming to generate more confrontation
weakness on c6, yet Black has managed
on the kingside with ...f5-f4. However,
to create some dynamic pressure on the
this weakens some dark squares,
kingside and down the e-file. The engine
particularly e5.
rates the game as even.
20.Qa6!? f4!?
16.Rae1 Bxe3 17.Rxe3 Rxe3 18.fxe3
Qe8! A double-edged decision. The engine
claims it’s fine, but I don’t like it. It
The queen covers c6, while
creates confrontation on a part of the
simultaneously threatening e3.
board where White is about to have an
19.Re1 extra piece participating. It only works
because of a hidden resource Black had
on the 22nd move, but even Firouzja
8
t+*+w+l+ didn’t find it. Something calmer like
7
O*O*+oOo 20...h6 21.Qf1 Rb8 22.b3 Qd7 feels
more appropriate.
6
*+o+*+*+
5
+*Po+*+v 21.Qf1!
4 q+*P*+*+ Forcing Black to take on e3.
3 +*+*P*+p
2 pP*N*+p+ 21...fxe3 22.Qd3!
1 +*+*R*K*
a b c d e f g h

Black to move

19...f5
 This might have been an attempt to
8
t+*+w+l+ keep the position complicated and play
7
O*O*+*Oo for a win, but it is wildly overoptimistic

*+o+*+*+
6 and in violation of simple positional
principles. The knight is obviously a
5
+*Po+*+v better piece than the bishop due to the
4 *+*P*+*+ stability of the structure.
3 +*+qO*+p Black could have tried for dynamic

pP*N*+p+
2 equality with 23...Qe4!?, or just gone for
the simple 23...Bxf3, trading off the
1 +*+*R*K* problematic bishop and ensuring
a b c d e f g h equality. White might be slightly better
in the endgame that follows, but it is
nothing special. The d4-pawn is
somewhat vulnerable, as is the white
Black to move
king, so the slight damage to the black
Getting ready to take the pawn. structure should prove to be
inconsequential.
22...e2?

This natural move is what gets

8
t+*+*+l+
Firouzja into trouble.
7
O*O*+wOo
The only way to keep the balance was 6
*+o+*+*+
22...Qb8!, using the tempo gained by 5
+*Po+*+v
threatening the knight to take the pawn
on b2.
4 *+*P*+*+
3 +*+q+n+p
23.Nf3?! 2 pP*+o+p+
Missing a big opportunity. White
1 +*+*R*K*
probably didn’t want to weaken the a b c d e f g h
shelter around her king, but keeping a
potentially good knight against a bad
bishop was already possible: 23.g4! Bg6 White to move
24.Qxe2 is a large advantage for White.
24.Ng5!
23...Qf7?
The double threat on the h7-pawn and
Black’s queen forces Black to abandon Black to move
the defense of the pawn on e2.
Black is in serious trouble:
Firouzja might have been hoping for
the rushed 24.Ne5? Qf4!, with some 1) Both white rook and knight look
counterplay. superior to their counterparts. White’s
rook threatens to infiltrate on e7, while
24...Qg6 the knight may later hop into e5.
2) The doubled black c-pawns are
A sad necessity. How badly must targets.
Firouzja have wanted to turn back time
3) Black has a bad bishop, with pawns
and take on f3. This is a reminder to all
fixed on the same color on c6 and d5.
of us: play for the win, but don’t overdo
4) Partially due to number 3 on this
it! We need to know when it’s in our
list, Black has weaknesses around the
best interest to just go for the draw. An
dark squares.
equal position is, anyway, a better
battlefield for us to outplay our 26...h6
opponents. A lost one, on the other
hand, not so much. After 26...Rb8 27.b3 Rb4 28.Re7!
24...Bg6 25.Qxe2 is also clearly in Rxd4 29.Rxc7 Black’s pawns are far
White’s favor. more vulnerable than White’s, and there
is already a threat of checkmate with
25.Qxg6 Bxg6 26.Rxe2 Rc8+.

27.Nf3 Be4
8
t+*+*+l+
7
O*O*+*Oo Firouzja gives up his c6-pawn to keep
White’s rook out of the seventh rank.
6
*+o+*+v+
5
+*Po+*N* 28.Ne5 a5 29.Nxc6 a4 30.Re3!
4 *+*P*+*+ White uses her superiority on the dark
3 +*+*+*+p squares in all kinds of ways. With her
2 pP*+r+p+ last move, she introduces the option of a
1 +*+*+*K* sudden Re3-a3xa4.
a b c d e f g h 30...Ra6 31.Nb4?!

This unnecessarily complicates the

technical task. 31.Ne5 was more natural
and accurate, making it tougher for the well-posted bishop and the a4-pawn.
black rook to get activated.
38...Re6 39.Kd2?
31...Rg6 32.g3 h5 33.Ra3 Kf8 34.Kf2
Rf6+ 35.Ke2 Rg6 36.h4 This is almost certainly an attempt to
reach the time control before making
Putting everything on dark squares. any major decisions. The Soviet School
Not only is White up a clean pawn of Chess taught us that we should not
already, but the a4-pawn is also difficult rush in technical endgames. However,
to defend. Black’s only hope is that the there is an opposing Principle: Your
pressure against g3 offers him non-reaction in a critical situation might
counterplay. be enhancing the status quo of a
degenerating advantage. This
36...Ke8 37.Re3 Kd7 unnecessary move gives Black a crucial
tempo and endangers White’s potential
win.
8
*+*+*+*+ White is winning after the immediate
7
+*Ol+*O* 39.Na2!.
6
*+*+*+t+
+*Po+*+o
5 39...Rf6 40.Ke2
4 oN*Pv+*P The second rank must be protected.
3 +*+*R*P*
2 pP*+k+*+
8
*+*+*+*+
1 +*+*+*+*
7
+*Ol+*O*
a b c d e f g h
6
*+*+*T*+
5
+*Po+*+o
4 oN*Pv+*P
White to move 3 P*+*R*P*
38.a3!
2 *P*+k+*+
1 +*+*+*+*
The knight on b4 is White’s a b c d e f g h
problematic piece – that’s why it should
have gone to e5. However, with her last
move, Ju prepares Nb4-a2-c3, after
which the knight targets both Black’s Black to move
40...Rf8!
8
*+*+*T*+
+*Ol+*O*
Firouzja finds another weakness in the 7
white camp, in the form of the now
fixed pawn on b2. It becomes much
6
*+*+*+*+
harder for White to shift her knight to c3 5
+*Po+*+o
due to the possibility of ...Rb8. 4 *N*P*+*P
41.b3!
3 Pr+*+*P*
2 *+*+k+v+
+*+*+*+*
White’s only chance is to create a 1
passed a-pawn. However, this is a partial
admission of guilt, since we’ve already a b c d e f g h
said that trading off weaknesses, in this
case b2 for a4, always benefits the
defender, who has one less thing to be White to move
worried about.
43.Re3!
41...axb3 42.Rxb3 Bg2?!
Reminding Black that there are major
Firouzja is trying to move the bishop weaknesses that need coverage on the
to c4 via f1. However, the bishop is kingside.
needed on e4 to block the e-file and Also strong is to engage the Principle:
maintain the cohesion of the central and Passed pawns should be pushed, with
kingside black structure. 43.a4!, after which Black is in big
Correct was to curb White’s passed trouble.
pawn with 42...Ra8!.
43...Be4

Forced. The rook cannot be allowed to

enter the black camp.
43...Ra8 44.Re5! Rxa3 45.Rg5 is an
easily winning position for White.

44.Na2!

The knight returns to action via c3,

and the blockade on e4 will soon be
broken.
44...Rb8 despite being down only one pawn. The
white rook is just way too active after
51.Re5 g6 52.Rg5, and more pawns are
8
*T*+*+*+ about to fall.
7
+*Ol+*O*
6
*+*+*+*+ 8
*+*+*+*+
5
+*Po+*+o 7
+*Ol+*O*
4 *+*Pv+*P 6
*+*+*+*+
3 P*+*R*P* 5
+*Po+*+o
2 n+*+k+*+ 4 *+*P*+*P
1 +*+*+*+* 3 PtN*R*P*
a b c d e f g h 2 *+*+*K*+
1 +*+*+*+v
a b c d e f g h
White to move

45.Nc3!
White to move
White can live with Black’s rook on
48.Nd1?!
her second rank! Let’s remind ourselves
of the concluding remarks we made after The black rook is denied access to the
the Giri-Tari game we studied b2-square and is forced to retreat.
previously. In the end, what gives Ju the
48.a4! is better technique, pushing the
win is concrete calculation. We cannot
passed pawn first.
always keep control until the end of the
game. In order to win, some concession 48...Rb8 49.Re5
has to be made. In this case, it’s
allowing the black rook to be activated. The rook is activated, trying to force
yet another black pawn on to the
45...Rb2+ 46.Ke1 Rb3 47.Kf2 Bh1 same/wrong color as the bishop.
After 47...Rxa3?? 48.Nxe4 Black can 49...g6?!
resign. Instead, 47...Rb2+ is met with
48.Re2. Then a transition to a rook As often happens, best for Black
endgame with 48...Rb3 49.Nxe4 dxe4 would have been to sacrifice another
50.Rxe4 Rxa3 gives Black no hope, pawn and go for counterplay with
49...Kc6! 50.Rxh5 Rb3!. 53.Nxe4 dxe4 54.Kg2 Rxa3 55.Rxe4

50.Re1?! Black is unable to save the rook

ending since g6 is chronically weak.
Correct is 50.Nc3! c6 51.a4. He’s also a tempo short in his
counterplay – the king should go to d5
50...Be4 51.Nc3 Rf8+ 52.Kg1
in order to hold.

55...Rd3
8
*+*+*T*+
7
+*Ol+*+* 55...Kc6 loses easily to 56.Re6+ Kd5
6
*+*+*+o+ 57.Rxg6 Kxd4 58.Rg5.
5
+*Po+*+o 56.Kf2 c6
4 *+*Pv+*P
3 P*N*+*P* 8
*+*+*+*+
2 *+*+*+*+ 7
+*+l+*+*
1 +*+*R*K* 6
*+o+*+o+
a b c d e f g h 5
+*P*+*+o
4 *+*Pr+*P
Black to move
3 +*+t+*P*
2 *+*+*K*+
52...Rf3? 1 +*+*+*+*
The rook ending is completely lost for a b c d e f g h
Black.
White’s win is not a given after
52...Kc6!!. The main tactical idea is that White to move
53.Nxe4 is met with 53...Re8. Being
more patient with 53.Re3 Rf7, and then 57.Kg2!
trying to distract the black rook with
54.a4 Rf8 55.a5, also fails to: 55...Ra8! Zugzwang. Black must give way.
56.Nxe4 dxe4 57.Rxe4 Kd5! and the
57...Rd1
active black king generates enough
counterplay for a draw. Another important line is: 57...Rc3
58.Re5 Rd3 59.Rg5 Rxd4 60.Rxg6 Rd5
61.Rg5! Rxg5 62.hxg5 Ke6 63.Kh3! transform our advantage into something
Kf5 64.Kh4 and White wins. more tangible is often required to
actually secure the full point.
58.Kf3 Rf1+ 59.Ke3 Rf7 60.Rf4 Rg7 Let’s move on to a few studies that
61.Kd3 Ke6 62.Rf8 Ra7 63.Ke4 Ra1 showcase the theme of a weak piece. So
64.Rd8! 1-0 subtle and incomprehensible are some
endgame studies that we get the feeling
Black’s king and rook will be
we are being talked about in a language
hopelessly tied down in passivity once
we don’t understand. They tend to
White’s rook reaches d6, so Firouzja
intimidate club-level players, and one of
resigned.
my goals in life is to break this fear,
This game is a good example of
since they are incredibly beneficial to
exploiting the opponent’s multiple
the development of our chess
weaknesses. It started off as mainly a
understanding. Endgame studies teach
structural affair when White decided to
us to look past the surface. In fact, even
cripple her opponent’s queenside
though solving them would obviously be
structure, starting with 12.Bb5!?
ideal, we don’t even need to do that. We
followed by 15.Bxc6. Then, possibly
can also benefit from trying for a few
because Firouzja played too ambitiously
minutes and then simply looking up the
and missed his chance for 23...Bxf3, the
answer.
situation turned into an unpleasant
endgame where Black had a bad bishop. Game 21
Robert Fontana
White generally played well and
Schweizerische Schachzeitung, 1945
converted her advantage to a full point
without allowing too many chances. We
shouldn’t be harsh on the players
because we have access to a heartless
engine. When compared with those
silicon monsters, everyone’s technique
would be deemed inadequate. However,
we once again encountered the same
motif as in the previous game: a slight
hesitance to take concrete action (for
example, 39.Kd2) made White’s
advantage shrink. Chess is a brutally
concrete game. Yes, positional trump
cards are great, but the courage to
 isn’t all that obvious that White can
8
*+*+*+l+ even hold the game. Incredibly, it’s a
7
+*+*+*+* White to play and win study! I found the

*+*+*KpB
6 draw with the freeing maneuver of g6-
g7, g5-g6, and Bc1, which saves the
5
O*+*+*P* game for White, yet fails to win.
4 *O*+*+*+ White’s odds for a win don’t look so
3 +p+*+*+* good. Yet keep in mind that “not so
good” is not the same as “impossible”.
2 *+*+*+*+
+*+*+*+*
1 Exercise: Black’s position is the
otherwise functioning car, with
a b c d e f g h growing rust on its underside.
Children believe in magic, while
few grown-ups do. Yet real magic
White to move lies inert in this position. How on
White to play and win earth can White win?

Let’s assess: Answer: The key is to activate the

1) White is up a piece and a pawn. Our locked-in bishop by transferring it to the
problem is that we are not able to force a a1-h8 diagonal, but in three moves
queen with our forward g-pawn, since instead of four (as with g6-g7, g5-g6,
Black is in firm control over the and Bh6-c1-b2). White has an amazing
promotion square on g8. idea to achieve this Houdini-like
2) White’s bishop is horribly out of the untangling trick.
game on h6, since it is hemmed in by
our own g5-pawn. 1.Bg7!!
3) Black threatens to promote a pawn
What could possibly be the point of
by creating a queenside passer with
such a move? After all, if we ever move
...a5-a4.
our king, then Black simply chops our
4) Not only is White’s bishop bishop with ...Kxg7.
dysfunctional, but the white king is as
In a tournament game, 99% of us
well. So far forward is White’s king that
would be thankful that we found a draw
it is useless in halting Black’s coming
with 1.g7? a4 2.g6 a3! 3.Bc1 a2 4.Bb2
passed pawn.
a1=Q 5.Bxa1, which is stalemate.
At first, I thought this might be a
White to play and draw study, since it 1...a4
 Now what? Isn’t White dead lost? We 3.Ke7 and 3.Ke6. What would you
need to find the follow-up idea for the play?
seemingly useless posting of the bishop
to g7. Answer: Only e6 is the correct square
for White’s king, since it gives Black’s
2.Bh8!
king air. Were we to choose e7, Black
This is the point! White’s bishop is would pull off another stalemate cheapo.
safe on h8, since Black gets mated if it is
3.Ke6!
taken.
People who make a mess of things
2...a3
always believe in second chances.
If Black swallows up our fortune with 3.Ke7? tosses away the win: 3...a2
2...Kxh8, then we respond with 4.Bb2 (4.Ba1 is stalemate) 4...a1=Q
gratitude. White wins simply by seizing 5.Bxa1 is stalemate.
control over g7 and g8 with 3.Kf7,
3...a2
followed by mate in two: 3...axb3 4.g7+
Kh7 5.g8=Q mate. Again, taking the bishop with
3...Kxh8 allows 4.Kf7 a2 5.g7+ Kh7
8
*+*+*+lB 6.g8=Q mate.
7
+*+*+*+* 4.Ba1
6
*+*+*Kp+
5
+*+*+*P* We see why White’s king chose e6
4 *O*+*+*+ rather than e7, as the black king can go
to f8, after which White wins easily,
3 Op+*+*+* either by promoting one of the g-pawns
2 *+*+*+*+ or by coming back with the king and
1 +*+*+*+* winning the remaining black pawns on
the queenside.
a b c d e f g h
This quirky study displays a situation
where we believe one side owns a weak,
ineffective piece, yet it doesn’t if the
White to move correct plan is found. White’s weak
piece comes into play with the amazing
Exercise: It’s still not too late to maneuver of first 1.Bg7!!, then 2.Bh8!.
blow it. White’s choice is between What I’m trying to say is that
appearances can be deceptive, Upon closer examination, we see that
weaknesses are relative, and tactics are Black is about to lose a piece due to a
all-important. tactical anomaly.
Exercise: How can White force the
Game 22 capture of a piece?
Jan Rusinek
The Problemist, 1993 Answer: Our bishop can control the
e8-square, and the knight is one move
away from c6, where it threatens both e7
8
*N*+l+*+ and d4, hinting at a possible fork.
7
+*+*+*+* 1.Bb5+!
6
*+*+p+*+
5
+*+*+*+* Black’s king and d4-bishop are unable
4 o+*V*+*+ to dodge the coming knight fork.
3 +*+*+*+* 1...Kf8!
2 vP*+*+*+ If the bishop is to fall, let’s at least
1 +*K*+b+* force White to give up the e-pawn first.
a b c d e f g h 1...Kd8? 2.Nc6+ and 1...Ke7? 2.Nc6+
Kxe6 3.Nxd4+ Kd5 4.Nc2 win easily.

2.e7+!
White to move
White to play and win
Attraction. The king is forced to step
White faces colossal obstacles to the up to e7.
win, which carries with it a heavy load
2...Kxe7 3.Nc6+ Kd6 4.Nxd4 Kc5
of expectations:
1) White is up by only one pawn. Oh no! Double attack! Are we losing
2) The e6-pawn is shaky and may fall. our piece back?
3) There are only a few pawns After 4...a3 5.b4!? the black king is
remaining on the board, which favors denied access to the c5-square and
the defender. White wins.
4) It’s hard to believe that White is
winning with just one unstable extra
pawn.
 back our extra piece, or else Black will
8
*+*+*+*+ liquidate all the pawns from the board,
7
+*+*+*+* resulting in a drawn position.
6
*+*+*+*+ 6...Kxb5 7.Nc3+ Kc4
5
+bL*+*+*
4 o+*N*+*+ 8
*+*+*+*+
3 +*+*+*+* 7
+*+*+*+*
2 vP*+*+*+ 6
*+*+*+*+
1 +*K*+*+* 5
+*+*+*+*
a b c d e f g h 4 *+l+*+*+
3 P*N*+*+*
White to move
2 v+*+*+*+
1 +*K*+*+*
Exercise: Find the only move a b c d e f g h
which keeps White’s winning hopes
alive.

White to move
Answer: Another fork saves White.
Exercise: We have a problem. If
5.Ne2! a3!
our knight captures the bishop, then
The best practical try. Black plays ...Kb3, winning our
Allowing the fork first with 5...Kxb5 final pawn and holding the draw.
6.Nc3+ Kc4 7.Nxa2 and then trying to However, if we examine the
trade off the last pawn fails, no matter position more deeply, we see that
the move order: 7...Kb3 loses to 8.Kb1 Black’s bishop is restricted and still
a3 9.Nc1+, and 7...a3 is convincingly in danger. Look closely, and you
met with 8.b4!. may find a flash of insight. How
does White exploit Black’s
5...Kb4 also loses easily, to 6.Bxa4!
dysfunctional bishop to force a win?
Kxa4 7.Nc3+ Kb3 8.Nxa2.

6.bxa3 Answer: Only the c2-square works.

Other tries are fumbling apologies
There is no choice. We must hand which just don’t cut it.
8.Kc2! Game 23
Gleb Zakhodyakin
We finally let Black in on our joke. 64, 1931
Black is in zugzwang.
Aren’t we all attracted to shiny
objects? Thinking is hard work, and it’s
much more relaxing to make automatic,
8
m+*+*+*L
shallow moves like 8.Nxa2??, but after
7
+*+*+*+*
8...Kb3 we are back to the harsh cold 6
*+*+*+*+
reality: our final pawn is eliminated. 5
+*+*+*P*
There is the easy, shallow way, and 4 *+*K*+*+
+*+b+*+*
then there is the difficult path of depth 3
and accuracy. 8.Kb2? falls under the
first category, as after 8...Bb3 White is
2 *+*+*+*+
in zugzwang, and Black holds. 1 +*+*+*+*
a b c d e f g h
8...Bb3+ 9.Kb2

Comprehension comes crashing down.

Black is in zugzwang and has to White to move
abandon the poor bishop. White to play and win
A fear can be specific, or just a vague
suspicion that something is wrong. Winning for White will not be an easy
Black’s weak piece, the light-squared task. The following are points that must
bishop, was not obvious to identify in be pondered for our side:
this study. We only saw that it was 1) There is only one remaining pawn
trapped at the end. Sometimes, a piece on the board. If it is blockaded or lost,
may look fine to the eye’s superficial Black achieves a draw.
scan, yet inner rot can still exist under 2) Black’s knight is both offside and
the surface. The exploitation might only rootless. Can we trap it while holding on
be possible to achieve through a deeper to our final pawn? The answer to this
understanding of the position. question is not so clear. 3
3) Black’s king can never approach
The following is a classic study on the our remaining pawn, as long as our
theme of domination, demonstrating the bishop covers g6.
struggle between a good minor piece 4) Due to point number 3, our
and a dysfunctional one. domination in terms of activity extends
beyond the minor pieces and includes
the kings, as White’s king is infinitely 8
*+*+*+*L
more powerful than Black’s, who 7
+*+*K*M*
*+*+*+*+
cringes in the corner. 6
1.Kc5! 5
+*+*+*P*
We threaten Kc6 and Kb7, which
4 *+*+*+*+
would win the knight, forcing Black’s
3 +*+b+*+*
next move. 2 *+*+*+*+
1...Nc7 2.Kd6 Ne8+
1 +*+*+*+*
a b c d e f g h
The knight gives White’s king and
bishop a wide berth. Once again, there
was no choice for Black.
White to move
2...Na8?? would be a delightful gift
for us, as after 3.Kc6 there is no remedy 4.Bg6!
to the coming Kb7, trapping the knight.
Oh, nyet, you don’t! Black’s knight is
3.Ke7 Ng7 imprisoned, but it’s protected, so it
cannot be taken – at least not yet. Our
We learned from Star Wars that if plan is to run Black’s king out of moves,
your ship is being chased by an Imperial thereby forcing the knight to move and
Cruiser, then your best chance for be captured.
survival is to head straight into the
asteroid belt. Black’s knight has evaded 4...Kg8 5.Bf7+ Kh7 6.Kf6
capture and reached the black king, yet
its safety status is still uncertain – its Preventing ...Nf5.
mobility remains severely restricted.
6...Kh8
3...Nc7?? loses swiftly to 4.Kf7! Nd5
5.g6 when g6-g7 mate is unstoppable.
 7...Kh7 8.Ke4!
8
*+*+*+*L
+*+*+bM*
7 Losing a tempo on our way to g4.
The immediate 8.Kf4 doesn’t spoil
6
*+*+*K*+ anything, but after 8...Kh8 9.Kg4 Kh7
5
+*+*+*P* we may realize that we need to go back
4 *+*+*+*+ to the squares e4, e5, and f4, and
3 +*+*+*+* triangulate with the king to lose a tempo.
The problem is that 10.g6+? allows
2 *+*+*+*+ 10...Kh6 with a draw. The black king
1 +*+*+*+* needs to be on h8 for g5-g6 to work.
a b c d e f g h
8...Kh8 9.Kf4! Kh7 10.Kg4! Kh8
11.g6!

White to move The knight has to move, and it’s all

over.
Exercise: The position contains all I like the clarity of this study. The
the requisite qualities to produce a knight is slowly pushed into its corner,
zugzwang. How do we force the where it eventually falls, due to
win of the knight? zugzwang. There are few ways to
illustrate better how bishops thrive on an
Answer: The king should go to g4, open board while knights suffer, how
followed by g5-g6. Trying to eliminate knights on the rim are dim, and how
the h7-square without pushing the g- kings and knights are the most
pawn requires using the king, but using susceptible pieces to being put in
the king allows stalemate tricks. Pushing zugzwang.
the pawn to g6 breaks the connection
between the bishop and the h5-square, Conclusion
which is why the king should go to g4 In this chapter we tried to understand the
first. difference between a functional and a
non-functional piece. Anyone who has
7.Ke5!
suffered a heart attack understands the
The tempting 7.Kg6? runs into horrors of constricted arteries. Any time
7...Ne6! and capturing the knight would one of our pieces is not working
lead to stalemate. properly, a feeling of restricted
dysfunction arises. We have three goals:
 1) We must become adept at
recognizing weak/dysfunctional pieces
and avoiding getting them.
2) When we do get them, we should
Chapter Four
strive to either trade them off or improve Exploiting Holes
them.
3) When our opponent has a bad piece, Exploiting a hole is just a microcosm
we must be skilled at finding clear plans of Chapter Five and the two can be
to exploit it. These could include somewhat intermingled. What is a hole
trapping it, like in the final study, and what is a weak square in chess?
highlighting its defects, as in Khalifman- A weak square is one which the
Barua, or ignoring it and exploiting its defending side is unable to control with
weakness by initiating active play, as in pawns. A hole is an even more
Giri-Tari. vulnerable square, which the opponent
Most of all, the goal of this chapter is threatens to occupy with a piece which
to showcase how dysfunctional pieces cannot be ejected by the defending
can and should also be considered side’s pawns. This means that if our
weaknesses, expanding the concept of a piece jumps into it, only the opponent’s
weakness from something concerning pieces can fight for it, and never the
only pawns and squares into something opponent’s pawns. Such squares tend to
that can also refer to more dynamic be ripe for piece occupation from our
factors. side.
In the following game, Carlsen
demonstrates a typical way of creating
holes in the enemy camp. He then
occupies it with his knight, securing a
large advantage in the center and,
consequently, on the kingside.
Game 24
Magnus Carlsen – Anish Giri
Internet (rapid) 2020

1.d4 Nf6 2.c4 e6 3.Nf3 d5 4.Nc3 c5

5.cxd5 Nxd5 6.e4 Nxc3 7.bxc3 cxd4
8.cxd4 Bb4+ 9.Bd2 Bxd2+ 10.Qxd2 0-
0 11.Bc4 Nd7 12.0-0 b6
 8
t+vW*Tl+ 8
*+tWt+l+
7
O*+m+oOo 7
Ov+m+oOo
6
*O*+o+*+ 6
*O*+o+*+
5
+*+*+*+* 5
+*+*+*+*
4 *+bPp+*+ 4 *+*Pp+*+
3 +*+*+n+* 3 +b+*+n+*
2 p+*Q*PpP 2 p+*Q*PpP
1 R*+*+rK* 1 +*+rR*K*
a b c d e f g h a b c d e f g h

White to move White to move

13.Rad1 The idea is to clear f8 for the knight,

which in turn protects the under-
Earlier in the book we looked at a defended black king. Moreover, the rook
game where Carlsen played 13.a4 is well-placed on e8, anticipating an
against Kramnik. increase in activity whenever White
plays d4-d5.
13...Bb7 14.Rfe1 Rc8 15.Bb3
16.Re3!?
On b3, the bishop is stable, fights for
control over the d5-square, and leaves From e3 the rook can quickly swing to
the d-file unblocked, which increases the the kingside and take aim at the black
activity of the white queen and d1-rook. king.
15.Bd3 is also sensible, as the bishop 16.h3 was White’s main move.
takes better aim at the black king. Both
moves are interesting. 16...Nf6 17.d5!?

15...Re8 At club level, we ordinary players tend

to follow the commands of theory like
soldiers who never in their lives dare to
disobey an order from a superior officer.
That’s not true for players like Carlsen,
who routinely create new theory with
their deep, pre-game preparation. What is Carlsen expecting for the
Disregarding the tactics involved for a sacrificed pawn?
moment, we can say that this thrust is 1) White created a kingside pawn
logically timed, since Black’s knight just majority, which means some attacking
left d7, from where it was ideally posted prospects.
to meet d4-d5 by jumping to the outpost 2) By preceding e4-e5 with d4-d5,
on c5. White cleared d4 for the f3-knight.
17.Qe1 and 17.Qd3 had been tried 3) By making Black capture with a
previously. pawn on d5 before moving e4-e5, the
b7-bishop is restricted, and the black
17...exd5 18.e5!
knight is denied the opportunity to
The milquetoast 18.exd5 looks occupy a hole of its own on d5, as would
comfortable for Black after 18...Qd6. have been created by the immediate e4-
e5.
8
*+tWt+l+ The sacrifice is completely sound, and
actually extremely typical. Carlsen has
7
Ov+*+oOo both seen and performed such sacrifices
6
*O*+*M*+ thousands of times, including the exact
5
+*+oP*+* same one. For example, in Carlsen-
4 *+*+*+*+ Firouzja, Wijk aan Zee 2021, the
position below was reached:
3 +b+*Rn+*
2 p+*Q*PpP
1 +*+r+*K* 8
t+*T*+l+
a b c d e f g h 7
+v+*+oO*
6
o+*+oM*O
5
R*+*+*+*
Black to move 4 *+*PpW*+
Sacrifice is one of the most “real”
3 +*+b+n+p
parts of chess, since it involves giving 2 *+*+*Pp+
up something, which is a type of
suffering. Of course, for the sacrifice,
1 Q*+*R*K*
a b c d e f g h
we expect to inflict some kind of
compensatory counter-suffering upon
the opponent!
 White to move

Carlsen played – you guessed it –

8
*+t+t+l+
23.d5!? exd5 24.e5! and went on to win.
7
OvW*+oOo
Back to our present game: 6
*O*+*+*+
18...Ne4
5
+*+oP*+*
4 *+*+m+*+
The most natural move, since the 3 +b+*Rn+*
knight goes forward and gains a tempo
against White’s queen. However, it
2 p+*+*PpP
itsn’t as great as it appears, as White
1 +*+rQ*K*
will soon move the knight from f3 and a b c d e f g h
oust the black knight from e4 with f2-f3.
18...Nd7 is equally sensible for Black
according to the engine. White to move

19.Qe1! 20.Nd4!

Threatening Rxe4. The queen on e1 is Tactics allow this positionally

better than it first appears, as the soon- desirable move despite the e5-pawn
to-be-played f2-f3 will clear the path for seemingly hanging.
the queen to transfer to the kingside.
20...a6
19...Qc7
Preventing any possible Nd4-b5-d6
maneuvers.
20...Rxe5?? is fool’s gold: 21.f3 Rh5
22.h4! Nf6 23.g4! Rh6 24.Nf5 Rg6
25.Ne7+ and Black is crushed.

21.h4!

Played with a shout-out to AlphaZero,

who taught us that it’s always a good
idea to push our h-pawns! The pushed h-
pawn could be useful in the attack about
to follow, the black knight is deprived of
the option to retreat to g5, and luft is
created for the white king if he ever
needs to go to h2. 8
*+*Tt+l+
After the obvious 21.f3 Black has 7
+vW*+oOo
21...Nc3!, aiming for ...Nb5 to
challenge the d4-knight, and if 22.Rc1
6
oO*+m+*+
then there is 22...Qc5! 23.Qxc3 Qa3!,
5
+*+oP*+p
forcing 24.Bc4 with equality. 4 *+*N*+*+
21...Rcd8?!
3 +b+*Rp+*
2 p+*+*+p+
Giri was thrown by Carlsen’s opening
preparation and at this stage found
1 +*+rQ*K*
a b c d e f g h
himself around 10 minutes behind on the
clock.
21...Nc3?? was the antidote we
mentioned against 21.f3, but it now White to move
loses to 22.Rc1 Qc5 23.Rd3!, and Black
24.Nf5!
can resign. With the pawn on f3, the d4-
knight is pinned, and ...Nc3-e2+ would This is scary stuff for Black:
be winning for Black.
1. Both Nd6 and Qg3 are threats.
A more circumspect and better
2. The dark squares in Black’s camp
approach would have been 21...b5!,
are highly vulnerable. It has become
creating the b6-square for the queen and
obvious that White’s activity and
introducing the idea of ...b5-b4 to
attacking chances are worth more than
possibly hold on to the knight about to
Black’s extra pawn.
appear on c3.
24...d4!
22.f3 Nc5 23.h5! Ne6?!
Black follows the Principle: Meet the
Giri possibly underestimated the force
opponent’s flank attack with counterplay
of the upcoming attack, but it was
in the center. By removing the pawn
already somewhat difficult to play for
from d4, Black activates his bishop, but
Black.
the b3-bishop also becomes more active.
The best idea was again to try creating
some counterplay with 23...b5! followed 25.Red3 Nc5?
by ...Qb6.
Nope, that’s incorrect. Black’s
counterplay arrives with Mastodonian
slowness. produced one tactical blow after
25...Bd5! was Black’s final hope, as another. How should White
the bishop on b3 should be eliminated. continue?
The b7-bishop is useless on the long
diagonal, while the e6-knight is Answer: The defense is completely
excellent and currently the only disrupted with the push, which plays on
defender of the black kingside. Hence, the theme of Black’s back-rank
the bishop should offer itself up for the weakness.
trade, not the knight.
30.e6!
26.Rxd4 Rxd4 27.Rxd4 Nxb3 28.Qg3!
We suspect that Magnus isn’t pursuing
Zwischenzug, threatening mate on g7 honest business dealings.
and further underlining the kingside
weaknesses created by the absence of 30...Qc1+
the e6-knight.
30...Qxg3?? misses the entire point:
28...g6 29.axb3 Rd8 31.Rxd8 mate. Instead, 30...Qc5 is
stunningly met with 31.Qe5!!, again
exploiting Black’s back rank and
8
*+*T*+l+ threatening Qg7 mate.
7
+vW*+o+o 31.Kh2 Rxd4
6
oO*+*+o+
5
+*+*Pn+p 31...Re8 32.Qe5 is a game-ender after
4 *+*R*+*+ 32...gxf5 33.exf7+ Kxf7 34.Rd7+.
3 +p+*+pQ* 32.e7!
2 *+*+*+p+ White’s threats continue to burn, white
1 +*+*+*K* hot.
a b c d e f g h
32...Qc8 33.Qe5!

White to move

Exercise: I watched this game live

and from this point, Carlsen
 Both win easily.
8
*+w+*+l+
+v+*Po+o
7 34.Kg3 1-0

6
oO*+*+o+ Of course, Carlsen wants the style
5
+*+*Qn+p points. White’s threats remain.
4 *+*T*+*+ This was a masterful display of how to
create and exploit holes. The advantage
3 +p+*+p+* of occupying a hole on d4 led to a
2 *+*+*+pK generalized weakness on Black’s dark
1 +*+*+*+* squares, which then transformed into a
direct attack upon the enemy king. It is
a b c d e f g h
almost mesmerizing to watch how
behind Carlsen’s play, even when it
involves dazzling sacrifices, the
Black to move fundamental positional principles are
able to shine through.
Talk about achieving cosmic oneness.
Even if you forget everything else
Black’s position, which drips blood
about this game, I urge you to keep in
from multiple wounds, is exposed to a
mind the typical mechanism of altering
plethora of dangers as white attackers
the structure that took place with 17.d5!
boil over the defensive wall. The main
exd5 18.e5!. Essentially, what that
threats are Qg7 mate, Nh6 mate and e7-
mechanism does is ensure that we have
e8=Q+ or e7-e8=R+, followed by
control of a hole on d4, while our
checkmate once more. Almost as an
opponent doesn’t get to occupy a hole
afterthought, Black’s rook hangs as
on d5. You might be thinking that this is
well.
way too specific, but it’s not. I can
33...Rh4+ 100% guarantee that you’ll have the
chance to use this motif in your own
I suppose that the oppressed don’t games, just as long as you stay on the
always have the resources needed to lookout for it!
stand up to injustice. What would Furthermore, we should note that the
Doctors Freud and Jung say about the hole on d4 was then occupied by a
good old spite check, our final act of knight, which turned out to be a real
defiance, before resigning? When we are beast. Knights are usually the best piece
losing, the truth turns into a liability. for occupying a hole. Their short-range
White can take the rook or ignore it. abilities shine best when they remain
relatively stationary, while other pieces of needing to win against a stronger
tend to care more about their ability to player to have a shot at first place. Let’s
move around freely. Thus, when we try to work out what is happening:
have a choice of which piece to occupy 1) White’s rook and bishop hang
a hole with, the knight should take first simultaneously, and it’s obvious that
priority. White loses the initiative if I chicken out
by passively backing my rook to d2.
Game 25 2) There are holes on h6, g7, and f6
Cyrus Lakdawala – Varuzhan around the black king. They look
Akobian enticingly weak, and I was seeing
Los Angeles 2001 visions of Qd4, Nh6+, and Qh8 mate.
Exercise: Desperation (for the win)
sometimes allows the dullard to
8
*+t+t+l+ produce ingenuity. When in an
7
O*+*Wo+o interview with Chess Life, I was
6
*O*Ro+o+ asked for the best chess move of my
life, I cited this game and my next
5
+*M*+*+* move. It’s White to play and win.
4 *+*+v+n+ Can you find it?
3 +*P*Q*+*
2 pPb+*PpP Answer: Sacrifice/undermining.
1 +*+*R*K* 25.b4!!
a b c d e f g h
Your writer is not normally associated
with the valor level of Richard Lionheart
or William Wallace (Braveheart). Yet in
White to move this case, desperation produced an
anomaly that leads to a forced win. The
GM Akobian, the tournament favorite,
knight on c5 is what partially holds the
and I reached this position in the final
black position together. Along with the
round of the Southern California State
Championship of 2001. California is queen, it is the only black piece that can
such a large state that the US Chess defend the fatally weakened dark
Federation split it into two smaller squares around the black king.
“states”. Varuzhan led by a half-point, 25...Qxd6
which put me in the awkward position
 All other tries lose as well. two pieces for a rook, while his king is
25...Bxc2 is answered with 26.Qe5! still under attack due to the weak dark
(threatening Nh6+ followed by Qh8 squares around his king.
mate) 26...h5 27.Nf6+ Kf8 28.Nxe8
(again threatening Qh8 mate) 28...Qxe8 27...Qd8
29.bxc5 when White is already up an
Alternatively, 27...Rxc5 28.Nxe4 will
exchange, and the weakness of Black’s
leave White up a piece, 27...Qxc5
dark squares hasn’t been solved, so
allows 28.Nd7+, winning the black
more material is about to come.
queen, and 27...bxc5 28.Qh6+ Ke7
Trying to deal with the menacing g4- 29.Nxe4 is similar to the game.
knight and abandoning everything else
with 25...h5 also loses to 26.bxc5 Bxc2 28.Qh6+
27.Rd7!. The d7-rook can’t be touched
due to a fork on f6, and moving the Taking on e4 immediately would also
queen would either allow Nf6+ or be winning, but why not kick the black
Nh6+, with devastating consequences in king a little further towards the center
either case. first?

26.Nf6+ Kf8 27.bxc5 28...Ke7 29.Nxe4 Qd5 30.Bb3 Qe5

Pinning the knight, but the pin is quite

8
*+t+tL*+ easy to break.
7
O*+*+o+o 31.f4
6
*O*WoNo+
+*P*+*+*
5 Gaining a tempo, while eliminating
any possibility for cheapos associated
4 *+*+v+*+ with the weakness of my back rank.
3 +*P*Q*+*
p+b+*PpP
2 31...Qf5 32.Qh4+ Kf8
1 +*+*R*K* 32...Kd7 allows White a skewer with
a b c d e f g h 33.Ba4+.

33.Qxh7 Red8

Black to move 33...Qxf4 opens the f-file for White’s

rook and is met with the crushing
Black is hopelessly lost since he loses 34.Rf1.
34.cxb6 axb6 35.Ng5 36.Rxe6!

Pawns can be weak even when Mikhail Tal was one of the first great
protected. The e6- and f7-pawns are players to be acutely aware of his
juicy sacrificial targets. audience and of the emotional impact
his moves had on his opponents.
35...Qf6 Growing up, Mommy Dearest
Lakdawala taught kid-Cyrus that it is the
35...Rxc3 hangs a rook to 36.Qh8+, so
height of obnoxiousness to be a show-
35...Rc6 was necessary, yet it’s obvious
off and gloat over a beaten opponent.
that Black won’t survive in the long
My mother’s training paid off, since it
term.
took every ounce of willpower on my
part to avoid displaying any exaggerated
8
*+tT*L*+ gestures for the spectators.
7
+*+*+o+q
6
*O*+oWo+ 36...Rd1+
5
+*+*+*N* 36...Qxf4? is met with 37.Qh8 mate,
4 *+*+*P*+ while taking the rook with 36...fxe6
3 +bP*+*+* allows 37.Nxe6+ when 37...Ke8
38.Ba4+ only leaves Black with spite
2 p+*+*+pP blocks on b5, c6 and d7.
1 +*+*R*K* After 36...Qg7 37.Qh4!? White’s rook
a b c d e f g h remains tactically protected, and
threatening to go to f6, where it is both
indirectly protected and piling up the
pressure against the f7-pawn.
White to move
37.Bxd1 fxe6
Exercise: We feel it in our bones
that White has access to a winning
sacrifice on e6 or f7. But the
question is: which pawn should we
sacrifice for, and with which piece?

Answer: Removal of the guard/mating

attack.
 8
*+t+*L*+ 8
*+t+t+l+
7
+*+*+*+q 7
O*+*WoOo
6
*O*+oWo+ 6
*O*Ro+*+
5
+*+*+*N* 5
+*M*+*+*
4 *+*+*P*+ 4 *+*+v+n+
3 +*P*+*+* 3 +*P*Q*+*
2 p+*+*+pP 2 pPb+*PpP
1 +*+b+*K* 1 +*+*R*K*
a b c d e f g h a b c d e f g h

White to move White to move

Exercise: Yes, everything leads to a Now 1.b4? would be nonsense, and it

win, but it’s best to wrap up the would allow Black multiple ways to
game sooner rather than later. You gain an advantage, the simplest of which
only need to find one strong move would be 1...Bxc2 2.bxc5 Rxc5 with a
to end the game on the spot. What clean extra pawn. Thus, White would be
should White play? forced to go back with 1.Rd2, after
which any ideas about the initiative are
Answer: Double attack. dissipated.
By the way, this game could easily
38.Qd7! 1-0 also have been placed in the next
chapter, since all Black’s holes could be
Going to b7 would have worked as seen as a weakness on a color complex –
well. The black rook is threatened, along in this case, the dark squares.
with a deadly fork on h7.
It might be 25.b4!! that won me this The following is probably the most
game, but the deciding factor was mysterious game in the book. In the
situated on the other flank, namely the opening, Carlsen commits multiple
black g-pawn that sat on g6 instead of strategic violations, only to later outplay
g7. Had the pawn been on g7, then the his rival through a series of brilliant
position would be the one pictured strategic contortions.
below:
 Game 26 even by the Scandinavian’s standards.
Fabiano Caruana – Magnus More normal is to move the queen to
Carlsen either a5 or d6. All I can comment on
Tromsø (Olympiad) 2014 this line is: this isn’t North Korea, and
we still have the right to make offbeat
1.e4 d5!? choices in the opening. The secret power
of the Scandinavian is that it is often
Yes, Magnus plays the Scandinavian! difficult for White to open the position
When I play the Scandinavian in my to truly exploit the lead in development.
own games, I think: “Welcome to my
squalid kingdom. It may not look like 4.d4 Nf6 5.Nf3 Bg4 6.h3 Bxf3
much, but at least it’s mine.” As a guy
who has written two books on this Carlsen correctly hands over the
opening, Carlsen’s last move warms my bishop pair. Allowing White’s knight
heart. into e5 would be a mistake.

2.exd5 Qxd5 3.Nc3 Qd8!? 7.Qxf3 c6

8
tMvWlVmT
8
tM*WlV*T
7
OoO*OoOo
7
Oo+*OoOo
6
*+*+*+*+
6
*+o+*M*+
5
+*+*+*+*
5
+*+*+*+*
4 *+*+*+*+ 4 *+*P*+*+
3 +*N*+*+* 3 +*N*+q+p
2 pPpP*PpP 2 pPp+*Pp+
1 R*BqKbNr 1 R*B*Kb+r
a b c d e f g h a b c d e f g h

White to move White to move

What would Paul Morphy say? Black 8.Ne2!?

made three moves and is already grossly
Does anyone believe in developing
behind in development. This is rare,
quickly in the opening anymore? The
somewhat convoluted idea seems to be Black to move
freeing the c-pawn, either to move to c3,
protecting the d4-pawn, or maybe later 10...Nbd7!?
to c4 to grab central space. The c3-
Carlsen claims that there is no rush to
knight blocking the c2-pawn is often
swap queens, as there is nowhere good
White’s problematic piece in these
for the white queen to run. So, he keeps
structures, but this feels like a weird
the option of taking on f3 in his back
moment to deal with it.
pocket, and simultaneously urges White
Normal is 8.Be3. to take on d5 himself, improving the
black structure.
8...e6 9.g4!?
11.Qg3 Qc4!?
Caruana ambitiously grabs kingside
space. Carlsen continues to provoke and
press emotional buttons on his opponent,
9...Qd5!
as skillfully as our mothers. He threatens
This is why Magnus refused to play the c2-pawn and gives his opponent a
...Nbd7 earlier. A queen swap benefits choice: allow Black’s queen to remain
Black due to White’s extra space and active on c4, or offer to swap it by
slight lead in development. moving White’s queen to b3?
Objectively, 11...Bd6! has to be better,
10.Bg2 as it develops a piece with tempo.
However, I think that if Carlsen had
approached the game with more
8
tM*+lV*T objectivity, he’d have played the
7
Oo+*+oOo opening altogether quite differently. His
6
*+o+oM*+ aim seems to be to provoke his elite
5
+*+w+*+* opponent to the point of eventually
getting some winning chances.
4 *+*P*+p+
3 +*+*+q+p 12.Qb3
2 pPp+nPb+ Played after 25 minutes of thought. If
1 R*B*K*+r nothing else, this shows that at least
a b c d e f g h Carlsen’s eccentric choices are making
Caruana lose some time on the clock.
Moreover, the move played is in
violation of the Principle: The side with will need to open the position.
more space should avoid trades. 2) The main available pawn break for
However, Black’s queen was White is d4-d5.
annoyingly active on c4. 3) To play d4-d5, White will need to
The engine doesn’t understand the support that move with c2-c4.
word ‘annoying’ and suggests 12.c3! 4) If the white c-pawn goes to c4, then
Be7 13.b3! Qa6 14.c4, which indeed the b4-square becomes a hole.
looks like a bigger advantage than the In general, with the pawns doubled on
one White got in the game.
the b-file, it becomes harder for White to
push the pawns together as a unit. Does
8
t+*+lV*T this sound far-fetched? It is not. It’s
7
Oo+m+oOo schematic thinking at its finest, and
6
*+o+oM*+ that’s how we should be thinking when
making decisions that have long-term
5
+*+*+*+* implications. Of course, the above
4 *+wP*+p+ points don’t mean that Black is better or
3 +q+*+*+p anything, as White is still the one
2 pPp+nPb+ pushing. But the change in structure
favors Black, and in the end, just so that
1 R*B*K*+r Carlsen seems like even more of a
a b c d e f g h genius, White lost because of these
doubled pawns.
I believe that virtually any other chess
Black to move player would opt for 12...Nb6, not
helping White to develop the a1-rook.
12...Qxb3!!
13.axb3 Bd6 14.c4
Played instantly. Most players would
consider opening the a-file for White’s White needs to play this sooner or
rook a serious concession, but Carlsen later.
has his heart set on violating as many
14...a6
strategic norms as possible in a single
game! In fact, Carlsen’s decision is This way, Black’s a8-rook no longer
positionally brilliant and even approved needs to babysit the a-pawn.
by the engine. Let’s see why that is:
1) White has the bishop pair, which
means that to increase his advantage, he
 The knight is heading for g6, from
8
t+*+l+*T where it takes aim at the slightly
7
+o+m+oOo weakened dark squares on White’s

o+oVoM*+
6 kingside.
However, starting with 17...h6! would
5
+*+*+*+* again be more precise.
4 *+pP*+p+
3 +p+*+*+p 18.Bf3
2 *P*+nPb+ Preparing h4-h5.
1 R*B*K*+r White looks clearly better after
a b c d e f g h expanding with 18.g5!, intending h3-h4
next, putting as many pawns as possible
on the opposite color of the extra g2-
bishop.
White to move
18...Ng6 19.h4
15.Be3?!

This is seemingly natural, but I don’t

like it. It’s important for White to push
8
*+lTt+*+
the queenside pawns; the bishop could
7
+o+*+oOo
do better work if it were supporting the
6
o+oVoMm+
b3-b4 push. 5
+*+*+*+*
In that regard, it was better to go for 4 *+pP*+pP
+p+*BbN*
15.Bd2!. 3
15...0-0-0 16.0-0-0 2 *P*+*P*+
Since the bishop isn’t helping, then it
1 +*Kr+*+r
might make sense to try with the king
a b c d e f g h
and play 16.Kd2!?, intending Kc3 and
b3-b4.
Black to move
16...Rhe8
19...Bf4
More accurate was to toss in 16...h6.
Principle: If your opponent owns the
17.Ng3 Nf8?! bishop pair, try to get rid of it via a
swap. However, this move does come The position is deceptive. On paper,
with a concession: when Black takes on White’s portfolio looks impressive:
e3, White’s center is strengthened and 1) White owns more space in all three
the f-file is opened. sectors of the board.
2) The f-file is opened for White’s
20.h5 Bxe3+ 21.fxe3 Ne7 22.e4?!
rooks.
Losing almost all the advantage. 3) White threatens e4-e5, followed by
Caruana wants to play e4-e5, hitting the Ne4 and then occupying the hole with
f6-knight, gaining space, and creating Nd6+.
the idea of Ng3-e4-d6. However, this Yet Stockfish assesses the position as
weakens the d4-square and allows Black even! The reasons may be:
to advantageously fix the kingside 1) White must deal with a hole on g5.
structure. Black can occupy it with ...Nh7!
White would have retained a followed by ...Ng5.
positional advantage in a rather one- 2) Shockingly, White’s knight is not
sided position with 22.g5! Nd7 23.Bg4!. particularly powerful once it reaches d6.
Black’s counterplay is restricted to a Just watch.
minimum and a rook is coming to f1
next. 23.e5 Nh7! 24.Ne4 Rf8 25.Nd6+ Kc7
26.Bg2 Ng5 27.Rhf1 f6!
8
*+lTt+*+
7
+o+*MoOo 8
*+*T*T*+
6
o+o+oM*+ 7
+oL*M*O*
5
+*+*+*+p 6
o+oNoO*O
4 *+pPp+p+ 5
+*+*P*Mp
3 +p+*+bN* 4 *+pP*+p+
2 *P*+*+*+ 3 +p+*+*+*
1 +*Kr+*+r 2 *P*+*+b+
a b c d e f g h 1 +*Kr+r+*
a b c d e f g h

Black to move

22...h6! White to move

 It’s in the nature of life that what we g5 and not allowing Black outposts on
desire rarely comes to us in the form we b5 and d5.
expect. Visually, it appears that White
must have a substantial advantage due to 8
*+mT*T*+
the knight on d6. But is this really the
case?
7
+oL*+*O*
1) Sure, White’s knight is entrenched
6
o+oNo+*O
on d6 – but so what? It doesn’t matter if
5
+*P*P*Mp
Black can simply play around it. 4 *+*+*+p+
2) Black’s monster g5-knight, now 3 +p+*+*+*
occupying a plump hole on g5, reminds
us of the 1950s sci-fi movie where
2 *Pk+*+b+
nuclear radiation created a giant mutant
1 +*+r+r+*
insect. That monster now threatens to a b c d e f g h
devour the city.
3) White’s pressure down the f-file
was an illusion. Black just played ...f7- Black to move
f6, and White could never capture
without surrendering his biggest 30...Ne7!
advantage: the knight on d6.
4) Black’s worst-placed piece is the A typical motif. To the untrained eye
knight on c7, currently dominated by the it might look as if the knight just likes
white pawns and unable to do anything wasting time. But the two tempi going
active. However, even this seemingly back and forth are well spent, since the
bad piece can perform a crucial task in white pawn was lured to c5 and the e7-
fighting against the d6-knight with knight, a self-made hero, comes back
...Nc8. and looks to occupy the newly created
hole on d5.
28.Kc2 fxe5 29.dxe5 Nc8 30.c5!?
31.b4?!
When playing this ambitious move,
consolidating the knight but conceding And just like that, Black becomes the
the d5-square, Caruana might have still one pushing. This creates a target on b4,
been under the spell of his previous while doing nothing about Black’s threat
advantage. to play ...Nd5 and ...Ne3+.
Safer would have been 30.Ne4, A much safer way to play would have
challenging Black’s powerful knight on been 31.Kd3 Nd5 32.Kd4, but it still
seems more pleasant to be Black. It’s both ...Nf3 and ...Rf3.
hard for White to keep together a 2) White’s queenside light squares are
position with so many holes: for weak, and Black’s king may soon slip
example, he should always be ready for into c6. After that, White’s knight will
an invasion on the f4-square. be tied down to d6, as it must protect
against ...Kb5.
31...Nd5!
Despite the engine’s equal assessment,
Once again, played instantly. Black White is clearly on the defensive. For
has a double threat on b4 and e3, forcing humans, this means a clear advantage
White to part with the bishop for the for Black.
knight. Look back at the diagram from
33.b5
only four moves ago, on move 27. Who
could have guessed that this seemingly Caruana realizes that his knight is
terrible knight on e7 would manage to doing nothing on d6, adopts a correct
trade itself off for the g2-bishop? defensive mindset, and tries to liquidate
some pawns on the queenside.
32.Bxd5 cxd5!
33...axb5 34.Nxb5+ Kc6 35.Nd6?
8
*+*T*T*+ The lure of d6 is too much for
7
+oL*+*O* Caruana. Principle: Despite their good
6
o+*No+*O looks, pieces deeply entrenched within

+*PoP*Mp
5 the opponent’s camp become completely
useless if the opponent can play around
4 *P*+*+p+ them; they might as well be considered
3 +*+*+*+* trapped.
2 *Pk+*+*+ White should activate his knight by

+*+r+r+*
1 offering a pawn with 35.Nd4+! Kxc5
36.Kd3 Kb6 37.Ke3 when he would
a b c d e f g h have good chances to save the game,
despite Black’s extra pawn. For a
human, accepting this clean pawn-down
White to move position instead of planting the knight
on d6 is tremendously difficult.
Let’s assess again:
1) The e5-pawn is no longer easy to 35...Nf3?
defend, since White must watch out for
 A rare occurrence – Carlsen misses the 37...Rxa1 38.Rxa1 Nxe5
mark. This allows White a trick
mentioned below in the note to move 37. Black wins a pawn, and it looks like
35...Rf3! gives Black a winning the white g-pawn is about to follow suit.
position. White is essentially paralyzed.
39.Ra7
36.b4! Ra8
39.Re1? is met with 39...Nf3!, holding
on to the e6-pawn due to the possible
8
t+*+*T*+ fork on d4.
7
+o+*+*O* 39...Rb8 40.Ra3 b6!
6
*+lNo+*O Undermining the defender of the d6-
5
+*PoP*+p knight yet another time.
4 *P*+*+p+
3 +*+*+m+* 41.Ra7 bxc5 42.Ra6+ Kc7 43.bxc5
Nd7 44.Ra7+ Kc6 45.g5 Nxc5
2 *+k+*+*+
1 +*+r+r+*
a b c d e f g h
8
*T*+*+*+
7
R*+*+*O*
6
*+lNo+*O
White to move 5
+*Mo+*Pp
37.Ra1?
4 *+*+*+*+
3 +*+*+*+*
Caruana misses his final chance to 2 *+k+*+*+
+*+*+*+*
save this game. 1
When playing 35...Nf3? Carlsen
a b c d e f g h
probably missed that at this moment
White has 37.Nf5!!, rerouting the
useless d6-knight, coordinating, and
saving the game. If the f3-knight moves, White to move
then Nd4+ will restore White’s
harmony, and the rook endgames A second pawn falls, and the game is
stemming from 37...exf5 38.Rxf3 are all effectively over.
drawn.
46.Nf7 d4 47.Ne5+ Kd5 48.Nd7 d3+
49.Kc1 Nxd7 50.Rxd7+ Ke4 0-1
8
*+vW*Tl+
O*+*+oOo
Many things happened in this game, 7
but essentially, it was a tale of two
holes: d6 and g5. Optically, the hole on
6
*T*O*+m+
d6 at first appeared to be the more 5
+o+p+*+*
impressive one. Then, slowly but surely, 4 *+o+pP*+
P*+mN*N*
we began to realize that White’s knight 3
was ornamental on d6, while the g5-
knight created very real threats.
2 *P*Q*+pP
That’s an instructive point if I ever 1 R*+b+r+k
heard one. It doesn’t only matter if a a b c d e f g h
hole exists and gets occupied; it also
matters how effective the piece standing
there actually is. For example, if in the
position after 32...cxd5 the d6-knight White to move
could magically appear on d4, then
22.f5!
White would have an indisputable
advantage, of almost decisive This is an instructive moment. Who
proportions. The knight on d4 would among us is unwilling to commit
perform the tasks of keeping the black strategic high treason if we believe that
king at bay by controlling c6, we will be rewarded with mate? Sure,
threatening the black pawn on e6, White’s last move is a strategic
blockading the d5-pawn, and stopping monstrosity, yet it could be a good call,
...Nf3 all at once. From d6, it did depending on how much of an attack it
nothing of the sort – it was essentially manages to generate. It’s tough to make
useless. Not all holes are created equal! a correct assessment, but we shall try
Let’s continue with an example that anyway:
will probably divide you. Some of you 1) White bleeds on the dark squares,
will love Nakamura’s next move, and with Black controlling huge holes on c5
some of you will hate it. and e5.
Game 27 2) White’s bishop is now officially a
Hikaru Nakamura – Magnus bad one, as the d5-, e4- and f5-pawns
Carlsen severely restrict the bishop’s mobility.
Shamkir 2014 However, the same exact pawns restrict
the black bishop as well.
3) The white knights are relatively
active and thus not terrible in since Black cannot realistically occupy
comparison to their black counterparts, them.
despite the holes on e5 and c5. 22...Nge5? is natural, but wrong. If
4) White has an ominous build-up of Black had another move, he would use it
both pieces and pawns on the kingside. to play ...Qg5 or ...Qh4, bringing the
Moves like f5-f6 are already in the air, queen to the kingside, stabilizing and
destroying the shelter of the black king. gaining a winning positional advantage.
5) There is no way to correctly assess As things stand, 23.f6! gives White an
such situations without accurate easily winning attack.
calculation. And if we calculate a little,
23.a4!
we will see that Black cannot play
...Nge5, which means that the major
Playing on both sides. Nakamura pays
hole White created on e5 will not even
attention to his worst-placed piece,
be occupied in the immediate future.
which is currently the rook on a1.
Conclusion: White’s attacking chances
are more dangerous than all of Black’s 23...a6
strategic portfolio.

22...Ngf4! 8
*+vW*Tl+
Carlsen finds the only move to
7
+*+*+oOo
survive. Black’s last move achieves two
6
oT*O*+*+
defensive functions: 5
+o+p+p+*
1) The knight on f4 blocks White from 4 p+o+pM*+
+*+mN*N*
playing f5-f6, since it interferes with the 3
white rook’s coverage of the f6-square.
2) The c1-h6 diagonal was the main
2 *P*Q*+pP
route for White’s queen to reach the 1 R*+b+r+k
kingside, and the f4-knight interferes a b c d e f g h
with that route as well. On the severe
downside to this, with the knight on f4
instead of e5, the two knights are
tethered together, making them both White to move
dysfunctional, and creating long-term 24.Ng4?!
coordination problems for the whole
Black army. The holes on c5 and e5 A concrete approach, threatening to
become pretty much inconsequential, take the knight on f4. It seems extremely
promising, since the f4-knight cannot
move and, as it turns out, cannot even be 8
*+vW*Tl+
protected! However, there is a third 7
+*+*+oO*
oT*O*+*+
route when one of our pieces is 6
threatened.
The simple positional route with
5
+o+p+p+o
24.Bc2! is stronger. White will eliminate 4 p+o+pM*+
one of the black knights for the 3 +*+m+*N*
*P*Q*NpP
atrocious c2-bishop, leave Black with a 2
terrible bishop on c8, and retain the
good knights, space advantage, and
1 R*+b+r+k
attacking chances. a b c d e f g h

24...h5!

Counterattacking is the least Black to move

appreciated method of reacting to a
threat. 25...Qf6!
24...Qg5?? walks into Nakamura’s Carlsen is desperate to remove queens
devilish trap: 25.Nh5!!, putting pressure from the board and seeks to play
on the pinned piece. Black can’t touch ...Qxb2.
the h5-knight, and he will still lose a
piece after 25...Nh3 26.Qa5!. 26.Nxd3!?

25.Nf2! There was a clean materialistic

solution available: 26.Nxh5! Nxh5
This is, at least partly, Nakamura 27.Nxd3! cxd3 28.a5! Rb7 29.Bxh5 Rc7
admitting his mistake from the previous 30.Rac1 Rxc1 31.Rxc1 Re8 32.Bf3.
move. He keeps the status quo, claims Black is down a pawn and close to
that the pawn on h5 is a weakness, and losing a second one. The engine
tries to mount a ferocious attack on the manages to just about keep Black in the
kingside. game, but it is still promising for White.
After 25.Rxf4!? hxg4! 26.Rxg4 Black
loses a pawn, but he also finally gets to 26...Nxd3 27.Qe3!
put a knight on e5 and achieve some
positional superiority. White retains an Gaining a useful tempo on the b6-
advantage, but it starts requiring more rook.
and more accuracy to keep it alive.
27...Rb7 28.Nxh5?! but 29.Qd4?! then 29...bxa4! offers
Black good play on the queenside and
Surprisingly, failing to include the down the b-file. That’s why White
capture on b5 allows Black good should have tossed in axb5 earlier.
chances to stay in the game. After
28.axb5! axb5 29.Nxh5 Qh6 30.Qd4! 29...gxh6 30.axb5 axb5!
White is in control and up a pawn.
30...Rxb5? is well met with 31.b3!.
8
*+v+*Tl+ After 31...cxb3 32.Be2 b2 33.Rab1 Rb3
34.Bxd3 Rxd3 35.Rxb2 Black’s
7
+t+*+oO* horrendous bishop is even more of a
6
o+*O*W*+ problem than the missing pawn.
5
+o+p+p+n When we find ourselves down a pawn
4 p+o+p+*+ in an ending, our world is usually not a

+*+mQ*+*
3 happy place. Black’s position is an
exception, and he has full compensation
2 *P*+*+pP for the missing pawn, since his
1 R*+b+r+k opponent’s position is riddled with a
a b c d e f g h vast array of impediments:
1) We go back to a Sesame Street
segment: “One of these things is not like
the other.” Black’s queenside majority
Black to move looks fast and may later turn into a
dangerous passed pawn on the
28...Qh6!!
queenside, while White’s majority is
Carlsen correctly realizes that his best sluggishly unimpressive.
shot at survival will be in an endgame, 2) Black’s knight has access to
despite White’s extra pawn. amazing holes on e5 and c5.
The greedy and predictable 28...Qxb2? 3) White’s e-pawn may later turn into
gives White a winning attack after a target on the open file.
29.f6!. A picture of an enraged coach 4) White has a bad bishop, with three
Bobby Knight comes to mind, tossing a central pawns fixed on the same/wrong
chair on to the basketball court. color.
5) The b2-pawn hangs, and White
29.Qxh6
must decide if he wants to protect it
White does not want to swap queens passively or let it go.
 so White seeks to return it to relevance.
8
*+v+*Tl+
7
+t+*+o+* 8
*+vT*+l+
6
*+*O*+*O 7
+t+*+o+*
5
+o+p+p+n 6
r+*O*+*O
4 *+o+p+*+ 5
+o+pMp+*
3 +*+m+*+* 4 *+o+p+*+
2 *P*+*+pP 3 +*+*+*N*
1 R*+b+r+k 2 *Pb+*+pP
a b c d e f g h 1 +*+*+r+k
a b c d e f g h

White to move
Black to move
31.Bc2!? Ne5!?
33...Rb8!
This slightly over-caffeinated response
indicates that Carlsen, like some White’s rook is chased off the sixth
medieval king who rules by divine right, rank.
is playing for the full point, from a
position where most players would be 34.Ra7 b4
happy to draw – especially against a
world-class player like Nakamura. Principle: Pawn majorities should be
pushed.
Objectively best is 31...Nxb2, but it
allows: 32.Ra6 Rd8 33.e5! with 35.Ne2 Bd7 36.Rfa1?!
excellent counterplay for White. The
engine works this out to a draw. The engine indicates that this is a
mistake, but it looks extremely natural
32.Ra6 to us humans. When normal-looking,
logical moves hand over an advantage to
White goes after Black’s main
our opponents, it means our position
weakness, the d6-pawn.
was already unpleasant.
32...Rd8 33.Ng3

The knight wasn’t doing much on h5,

 38.Kh2?
8
*T*T*+l+
R*+v+o+*
7 There are no heroes in time pressure.
Nakamura solves the problem presented
6
*+*O*+*O in the note above by removing the king
5
+*+pMp+* from the first rank and thus away from
4 *Oo+p+*+ any possible ...Rc1+ attempts, but that

+*+*+*+*
3 wasn’t Black’s only threat.
Only 38.Rb1!! was correct, but good
2 *Pb+n+pP luck finding that!
1 R*+*+*+k 38...c3 39.Nd4 cxb2 40.Rb1
a b c d e f g h
With the king on the first rank, the
issue was ...b4-b3 followed by ...Rc1+.
Black to move With the king on h2, that doesn’t work,
but there is another problem with the
36...Bb5! white position.

*Tt+*+l+
Black has managed to activate his bad 8
bishop and stands better.
7
R*+*+o+*
37.h3! 6
*+*O*+*O
The only move that does not lose by
5
+v+pMp+*
force. 4 *O*Np+*+
37...Rdc8!
3 +*+*+*+p
2 *Ob+*+pK
Preparing ...c4-c3. 1 +r+*+*+*
The immediate 37...c3?! is answered a b c d e f g h
with 38.Nd4! and after 38...cxb2 39.Rb1
b3!? 40.Bxb3! White holds it together. If
h2-h3 wasn’t played on the previous
move, 40...Rdc8 would decide here Black to move
because of the back-rank threat. That’s
40...Rc4!
why both 37.h3! and 37...Rdc8!
deserved exclamation marks. Loose pieces drop off.
41.Nxb5 Rxc2! Black to move

Apart from the knight on b5, Black 45...Nf2!

also threatens ...Nd3 followed by ...Rc1.
Creating dual threats:
42.Nd4 Rd2 1) Transferring the knight to c3 via d1.
2) Bringing the e8-rook in with
Not without reason, winning a tempo
...Rxe4, followed by doubling rooks on
in such sharp situations is almost a no-
the second rank.
brainer for a human. The engine calls for
42...b3!. 46.Ra2 Nd1! 47.Rxd1

43.Nc6 Re8 44.Ra4 This desperate exchange sacrifice

won’t save the game.
44.Nxe5 also loses: 44...Rxe5 45.Ra2
Kg7! 46.Raxb2 Rxb2 47.Rxb2 Rxe4 47...Rxd1 48.Rxb2 Rxe4
48.g4 Rd4. Black’s king will walk up
the central dark squares and pick off the White’s remaining pawns are weak
d5-pawn. The white king stepping on to and at least one more will fall.
the third rank would always be met with
a rook check on d3, followed by ...b4- 49.Nc6 Kg7!
b3.
Black’s king is safe from attempted
44...Nd3 45.Nxb4 perpetual check nets.

50.f6+ Kxf6 51.Rf2+ Kg6 52.Nd8

8
*+*+t+l+
7
+*+*+o+*
6
*+*O*+*O
5
+*+p+p+*
4 rN*+p+*+
3 +*+m+*+p
2 *O*T*+pK
1 +r+*+*+*
a b c d e f g h
 tethered to one another after 22...Ngf4!.
8
*+*N*+*+ Despite Nakamura’s decision being
7
+*+*+o+* correct, the further course of the game
6
*+*O*+lO serves to highlight a different instructive
point: when you make the committal
5
+*+p+*+* decision to destroy your structural
4 *+*+t+*+ integrity in favor of dynamics, you bear
3 +*+*+*+p a huge burden to continue accordingly!

*+*+*RpK
2 A few inaccurate moves, and the game
was dragged into an endgame where
1 +*+t+*+* Carlsen, despite being a pawn down, felt
a b c d e f g h able to play for a win due to the
innumerable holes all around the white
structure. He rightly felt that with a
stallion like the one on e5, nothing bad
Black to move
could happen to him – and he turned out
52...Re8! 0-1 to be right.
Conclusion
White can’t even take the f7-pawn, This chapter featured a few examples
since it is tactically defended, for concerning holes: how to create them,
example: 53.Nxf7?? Rxd5! 54.Rf3 Kg7! how to avoid them, how to exploit them,
and there is no remedy to the coming and, finally, when allowing them
...Rf8, trapping Black’s knight. wouldn’t be a mistake. One thing we
An extremely high-level and double- should keep in mind is that holes
edged encounter between two of the best occupied by a strong piece, most
players of the modern era. The most commonly a knight, can almost single-
notable moment came at our starting handedly decide a game. We should be
diagram, with Nakamura’s extremely very wary of creating them and swift in
committal decision on the 22nd move to our actions to occupy them.
put yet another one of his pawns on light
Commonly, holes and weak squares in
squares, completely killing his light-
general come in waves of more than
squared bishop and giving Black a one. For example, Nakamura’s 22.f5!
massive hole on e5. In exchange, did not only create a hole on e5, but also
Nakamura got tremendous attacking weakened White’s control of g5, made
chances against the black king and
f6 a safe square for the black pieces due
momentarily destabilized the black
to the absence of the e4-e5 resource, and
knights, who were forced to become
even exposed the dark squares behind
White’s pawns, namely e3, g3, f4, and
h4. Those squares became accessible to
Black’s queen and generally easier to
Chapter Five
occupy. This is what we would call a Exploiting Weak Color
weakness on a color complex. It is a Complexes
sister theme to holes and weak squares,
and shall be the subject of the next
To my mind, this is the most important
chapter. Onwards!
chapter of the book. It’s also the largest.
One of the most important skills we can
acquire in chess is to develop an eye for
a weakness arising on a color complex,
even before it occurs on the board.
Weak color complexes tend to be a
permanent factor, since they’re decided
by the structure and the assortment of
pieces that remains on the board.
However, that does not mean that such
weaknesses should make us regard a
position as static and go about exploiting
them in a patient, slow, positional
manner. On the contrary, our opponent
being weak on a color complex often
means that we hold some relative
weakness on the opposite color –
otherwise, we’re just dominating! Such
positions are extremely double-edged,
and who holds the initiative becomes the
deciding factor.
Let’s look first at a game where the
presence of opposite-colored bishops
accentuates the weakness on a color
complex. This time, it will be Black’s
light squares that are insufficiently
controlled. Kasparov exploits that
weakness in the most dynamic way
imaginable. Taking advantage of Black’s knight
Game 28 not going to e5 and choosing a more
Garry Kasparov – Judit Polgar active post for the bishop than e2.
Wijk aan Zee 2000
11...Qb6!
1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4
This is considered Black’s best
Nf6 5.Nc3 a6 6.Be3 Ng4 7.Bg5 h6
response, even today. White’s dark-
8.Bh4
squared bishop is committed to the
kingside, meaning that White’s control
8
tMvWlV*T over the central and queenside dark
squares are softened, and Black’s move
7
+o+*OoO* highlights that weakness in the best way
6
o+*O*+*O possible. White must watch out for both
5
+*+*+*+* ...Qxb2 and a discovered attack on the
4 *+*Np+mB d4-knight with ...Nxe4.
3 +*N*+*+* 8
tMv+l+*T
2 pPp+*PpP 7
+o+*OoV*
1 R*+qKb+r 6
oW*O*M*O
a b c d e f g h 5
+*+*+*O*
4 *+bNp+*+
Black to move
3 +*N*+*Bp
2 pPp+*Pp+
8...g5! 1 R*+qK*+r
Black has played ‘A’, so she has to a b c d e f g h
play ‘B’. Is this a tempo-gaining move
that grabs dark squares, or is it a
potential weakening of Black’s White to move
kingside? The answer is yes and yes!
12.0-0!
9.Bg3 Bg7 10.h3 Nf6
Kasparov boldly sacrifices a pawn for
10...Ne5 is more popular. a development lead, rejecting 12.Nb3
which protects against both threats, but
11.Bc4!?
is a more passive approach.
12.Bb3 Nxe4! was the main tactical 8
tMv+*Tl+
point behind Black’s previous move. 7
+o+*OoV*
12...0-0!
6
o+*O*M*O
5
+*+*+*O*
Polgar wisely avoids the temptation:
12...Nxe4? is way too much, and after
4 *+*+p+*+
13.Nxe4 Qxd4 14.Nxd6+! exd6
3 WbN*+*Bp
15.Qe2+ White was winning in 2 p+p+nPp+
Ivanchuk-Shirov, Wijk aan Zee 2001.
On the other hand, 12...Qxb2!? is
1 R*+q+rK*
a b c d e f g h
playable but extremely risky.

13.Nde2!?
White to move
Some people just don’t care about
money. I haven’t decided if pig-headed 15.f4!
obstinacy is a character flaw, since the
pig-headed tend to get their way in life. Principle: Create confrontation and
Kasparov refuses to defend his b-pawn! open the game when leading in
13.Nb3 was the prudent approach. development. Does White have enough
for the pawn?
13...Qxb2!? 1) White enjoys a substantial lead in
development.
You can’t wave a red flag in front of a
2) Black’s king cover was weakened
bull and expect anything else. Judit
when Polgar tossed in ...g7-g5.
Polgar, one of the most uncompromising
3) White’s queenside dark squares are
players to ever sit behind a chessboard,
punctured, and he is down a pawn. This
drops any pretense of meekness, forcing
means that if White’s attack fails, he is
Kasparov to prove that his pawn
in danger of being left in an unpleasant
sacrifice is sound.
endgame.
14.Bb3 Qa3! 4) White’s superiority on the light
squares is not uncontested, but is clearly
visible. The pawns on d6, e7, g5 and h6
are the culprits behind the weakening of
Black’s light squares.
 Conclusion: White’s pawn sacrifice is
a single-crop economy. If his attack 8
*+t+*Tl+
fails, so will his position. The engine 7
+o+*OoV*
o+mOvM*+
rates it at ‘0.00’, which translates in 6
human-speak to ‘unclear’.
5
+*+n+*O*
15...Nc6 16.Kh1! 4 *+*+p+*+
This is a wise investment of a tempo,
3 Wb+q+*Bp
since White doesn’t want his king on the 2 p+p+n+p+
open g1-a7 diagonal and on a dark
square, no less. At some point, someone
1 R*+*+r+k
a b c d e f g h
would eventually target the king on g1,
and he would have to move. By doing it
pre-emptively, on his own accord,
Kasparov doesn’t need to keep guessing Black to move
what’s going to hit him.
19...Rfe8?!
16...Be6!?
Black is in possession of major
A nice and natural solution to the weaknesses, most of which are
problem of the annoying pressure associated with the safety of her king. In
exerted on Black’s kingside by the b3- that regard, slow play is just not going to
bishop. White can almost never take on cut it. A dynamic solution is what Black
e6, as that would include the f8-rook in should be hoping for.
the game and, more importantly, it Due to the above, it is understandable
would help Black better control the that Black’s best path to dynamic
central light squares, including the equality lies in: 19...Nxd5! (19...Bxd5!,
currently weak d5- and f5-squares. with the same trick in mind, is also
possible) 19...Nxd5! 20.exd5 Nb4
17.Qd3 Rac8 18.fxg5!? 21.Qd2 Nxd5!, demolishing White’s
center by using some tactics. The main
Kasparov decides to open the f-file,
idea is that 22.Bxd5?! Bxd5 23.Qxd5
even if it means granting Black access to
runs into 23...Bxa1 24.Rxa1 Qe3! when
a hole on e5. In exchange, he gets to
the g5-pawn is protected, the white
further activate his f1-rook and leave
pieces are a bit clumsy, and the c8-rook
Black with a weak pawn on g5.
is coming to c5, activating the black
18...hxg5 19.Nd5 pieces while gaining more tempi.
20.Rad1! Black is busted, as the knight on d5
cannot be protected nor can it move due
Removing the rook from the influence to the possible Bxf7+, which would
of the g7-bishop and increasing control capture the queen on a3.
over d5.

20...Nb4 21.Qf3?
8
*+t+t+l+
7
+o+*OoV*
Sharp positions can force even
monsters like Kasparov to go astray.
6
o+*OvM*+
21.Qe3! attacking g5, the main black
5
+*+n+*O*
weakness, would guarantee a large
4 *M*+p+*+
advantage. Kasparov might have been 3 Wb+*+qBp
afraid of losing a pawn on d5, but any 2 p+p+n+p+
+*+r+r+k
attempt by Black to grab that pawn loses 1
by force: 21...Nbxd5 22.exd5.
a b c d e f g h
8
*+t+t+l+
7
+o+*OoV*
6
o+*OvM*+ Black to move
5
+*+p+*O* 21...Nbxd5!
4 *+*+*+*+ A wise move, even though it loses a
3 Wb+*Q*Bp tempo. Polgar recognizes the potential
2 p+p+n+p+ weakness on her light squares and
1 +*+r+r+k retains the bishop that controls them.
a b c d e f g h Instead, 21...Bxd5? 22.exd5 Nxc2 is a
terrible position for Black, and after
23.Rd2 Nb4 24.Bf2! White gets a
winning attack with Bd4 and Ng3,
Black to move intending to hop into f5.

Now 22...Nxd5 23.Qxg5 Nf6 loses to 22.exd5 Bd7

24.Bxe6! fxe6 25.Bxd6! exd6 26.Rxf6.
Even quicker is the refutation of
22...Bxd5, which gets met by the
spectacular 23.Rxd5! Nxd5 24.Qf3! and
 23...a5!
8
*+t+t+l+
+o+vOoV*
7 The engine fully supports Polgar’s
move, claiming that Black’s best chance
6
o+*O*M*+ is to go for the greedy approach,
5
+*+p+*O* unbalancing the white queenside, taking
4 *+*+*+*+ the a2-pawn, and then distracting the

Wb+*+qBp
3 white army with her newly created
outside passed pawn.
2 p+p+n+p+ However, the power of hindsight
1 +*+r+r+k allows us to slightly doubt the engine’s
a b c d e f g h assessment and suggest that perhaps
more prudent was 23...Bb5, stopping the
queen from accessing d3.

White to move 24.Qd3 a4 25.Bc2 Qc5!?

23.c3! Intending either ...Bb5 or ...Qc4.

The engine respects White’s attack
One of Kasparov’s finest traits was
and calls for 25...Kf8!, continuing to
that he virtually never allowed his pieces
prefer Black.
to malfunction. Like the greediest of
corporations, he made his employees
work long hours for low pay.
8
*+t+t+l+
1) The b3-bishop wasn’t doing much
7
+o+vOoV*
on its square. 6
*+*O*M*+
2) Thus Kasparov prepares to switch it 5
+*Wp+*O*
to c2, followed by Qd3, with a
queen/bishop battery on the b1-h7
4 o+*+*+*+
diagonal.
3 +*Pq+*Bp
3) When this happens, Black will need
2 p+b+n+p+
to be on constant lookout for exchange 1 +*+r+r+k
sacrifices on f6, followed by a queen a b c d e f g h
infiltration to h7.
The engine says Black has more than
enough resources to defend. For a
human, Black’s position isn’t easy. White to move
 How should White react?
8
*+t+tL*+
+o+vOo+q
26.Rxf6!! 7
A malignant spirit arises. The rook 6
*+*O*V*+
destroys the defender of the infiltration 5
+*Wp+*O*
o+*+*+*+
point on h7. The engine insists on 4
calling this even, but as I wrote before,
such evaluations are based upon the
3 +*P*+*Bp
assumption that Black will defend with 2 p+b+n+p+
perfect precision, which no human could
accomplish.
1 +*+r+*+k
a b c d e f g h
26...exf6!

Polgar finds the only move, which

White to move
gives her the following defensive
advantages: After 28.Nf4!! (threatening all sorts of
1) The pawn on f6 offers Black’s king invasions on e6 and g6)
more cover than if she recaptured with 28...gxf4 (forced) 29.Bxf4 e5! (again
the bishop. forced due to the threat of Bh6+)
2) By recapturing with the e7-pawn, 30.Bg6!! Ke7! (another only move)
Black opens the e-file for the rook on 31.Qxf7+ Kd8 32.Qxf6+ Kc7 33.Bxe8
e8. Rxe8 34.Bg5 White is winning, but
3) By recapturing on f6 with the e- Black is at least still kicking.
pawn, Black’s king may later escape to
the queenside via e7 and d8. 27.Qh7+ Kf8 28.Nd4!
26...Bxf6? 27.Qh7+ Kf8 is natural,
but incorrect.
 rook sounds overoptimistic for White’s
8
*+t+tL*+ dark-squared bishop.
7
+o+v+oVq Alternatively, 29.Bf5 Rxf5 30.Nxf5
6
*+*O*O*+ Bxf5 31.Qxf5 Rd8 is OK for Black.
5
+*Wp+*O* 29...fxe5?
4 o+*N*+*+ It feels almost unfair that Polgar’s first
3 +*P*+*Bp real mistake is already the decisive one.
2 p+b+*+p+ Could anyone have played it better with
1 +*+r+*+k Black? For the younger readers – if you
didn’t already know this – that’s how it
a b c d e f g h
felt watching Kasparov play: almost
unfair for his opposition.
The alternative with 29...dxe5! is
Black to move extremely ugly, as it hems in the g7-
bishop even further. But, in the game,
Let’s assess:
the absence of the pawn on f6 is the
1) The hole on f5 is a serious concern
deciding factor. Somewhat surprisingly,
for Black.
the engine thinks that after 30.Nf5 Bxf5
2) Apart from Bf5, White may even be 31.Qxf5 Rd8 32.Bxa4 b5 33.Bb3 Qxc3
threatening Bxa4 and if Black accepts, Black is completely equal. Of course,
Nf5 follows with decisive effect. it’s still a bit unpleasant for humans to
3) The d6-pawn is a serious weakness play with Black, but it is holdable.
for Black, keeping his queen tied down
to its defense.

28...Re5!

The only move. Black fights for

control over the f5-square, while also
covering the d6-pawn. Moves that solve
multiple issues at once are almost
always good.

29.Bxe5

Hoping for more in life than taking a

 Black’s king position is hopeless:
8
*+t+*L*+ 1) White threatens e6-e7+, followed
7
+o+v+oVq by Qxg7.
6
*+*O*+*+ 2) Black can’t afford to take the e6-

+*WpO*O*
5 pawn, since that would fatally open the
f-file for White’s rook.
4 o+*N*+*+ 3) Black is bleeding on the light
3 +*P*+*+p squares around her king. The opposite-
2 p+b+*+p+ colored bishops are amplifying the

+*+r+*+k
1 problem.

a b c d e f g h 31...Rc7

White to move
8
*+*+*L*+
7
+oT*+oVq
Exercise: Kasparov found a shot 6
*+*Op+*+
which rips apart the light squares
around the black king. How would
5
+*W*O*O*
you continue?
4 o+*+*+*+
3 +*P*+*+p
Answer: Overload the defence/open 2 p+b+*+p+
lines. 1 +*+r+*+k
a b c d e f g h
30.Ne6+!!

The f7-pawn is both controlling key

light squares and blocking the f-file. If it White to move
is removed, the black king is hopelessly
exposed. 32.Bxa4!?

30...Bxe6 Of course, it’s not the pawn that

Kasparov wants, even though it
The alternative 30...fxe6? 31.Rf1+ obviously doesn’t hurt. By taking the
Ke7 32.Qxg7+ Kd8 33.dxe6 is pawn he provides access to the b3-
obviously a disaster. square for his bishop, where it is both
stable and targets the weakened light
31.dxe6
squares on Black’s kingside. Moreover, namely through the b6-square.
Bd7 becomes an idea, blocking the c7-
rook’s connection to the kingside. 35...fxe6
The engine finds the even stronger and
It might look suicidal to open the f-
quite thematic 32.Bg6!, insisting on
file, but there wasn’t anything better.
eliminating the f7-pawn.
After 35...Qc4 36.Rf1 f6 37.Qc1! Ke7
32...d5 38.Qa3+! Rc5 39.Rb1 the white queen
will enter the black camp via a8, and the
Hopeless for Black is 32...fxe6 33.Qg6 black king’s agony will soon be over.
Rf7 34.Qxe6 due to the massive light-
square weaknesses. 36.Bxe6!

33.Qf5 Qc4 34.Bd7 Qf4 Playing for the attack on the light
squares. That is what got White so far. If
it ain’t broke, don’t fix it!
8
*+*+*L*+ It would be foolish to sell out with
7
+oTb+oV* 36.Rf1? Rxd7 37.Rxf4+ gxf4 and,
6
*+*+p+*+ suddenly, Black stands no worse.
5
+*+oOqO* 36...Ke7 37.Bxd5 Rd7 38.c4!
4 *+*+*W*+ Kasparov reinforces his powerful
3 +*P*+*+p bishop in the center, preferring to attack
2 p+*+*+p+ over a winning endgame after 38.Rf1
1 +*+r+*+k Qxf1+ 39.Qxf1 Rxd5.
a b c d e f g h
38...Qe3 39.Qh7! Kd8

White to move

35.Qb1!

We don’t normally associate retreats

with attacking ideas. In this case,
Kasparov seizes control over f1 for his
rook and gives his queen access to
infiltration paths on the queenside,
 Rd1, pinning Black’s queen.
8
*+*L*+*+ When the subject of opposite-colored
7
+o+t+*Vq bishops arises, most players think
6
*+*+*+*+ “draw”. When attacking chances remain
though, opposite-colored bishops tend to
5
+*+bO*O* be a nightmare for the defending side.
4 *+p+*+*+ For example, when Kasparov was
3 +*+*W*+p attacking on the light squares in this

p+*+*+p+
2 game (h7, g6, e6, f7 and so many
others), the g7-bishop was just a
1 +*+r+*+k bystander, unable to influence the
a b c d e f g h situation in the least.
As we said in the chapter introduction,
the true mastery lies in realizing the
White to move potential weakness on a color complex
before it fully materializes. In this case,
Black’s king is lost in the wilderness, the black kingside structure had allowed
unable to cope with the endless threats. light-square weaknesses to be created
The worst thing is the hopelessness from very early on, with the moves
surrounding this whole ordeal – could 7...h6 and 8...g5. It took time until
the king ever find safety were White to Kasparov managed to make that
go wrong? The answer is a resounding weakness felt. However, going through
“no”. The light squares will always be his moves one gets the feeling that he
way too weak, and the presence of always aimed to highlight his light-
opposite-colored bishops makes it so square superiority, right from move 11
White is essentially attacking with an when he chose 11.Bc4!?. Another
extra piece. noteworthy moment was 26.Rxf6!!.
Sure, tactics were involved, but the
40.Rb1 Qf4 41.Be6 Re7 42.Bg4 Rf7 sacrifice also signifies something else:
43.Qd3+ Qd4 the best players won’t allow anything to
stop them from achieving their
43...Kc7 44.Qd5 is also decisive.
positional goals.
44.Qg6! 1-0
Great players make chess look so
White’s queen is Annie Oakley, who simple. In the following strategic
never misses her target. There is no masterpiece, Anand does it all. He
remedy to the dual threats of Qxf7 and secures space, punctures the opponent
on a single color, hands the opponent a Black intends ...Bxb5, ridding himself
bad minor piece, threatens a decisive of the bad French bishop.
attack that induces Black into entering a
lost ending, and, finally, skillfully 9.Nxc7+
breaks down his opponent’s attempt to
White’s main line, picking up the
secure a fortress draw.
bishop pair and a potential for dark-
Game 29 square dominance.
Viswanathan Anand – Smbat
Lputian 9...Qxc7 10.Nf3 a6
Wijk aan Zee 2000
Insisting on his efforts to rid himself
1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.e5 c5 5.a3 of the bad light-squared bishop, Lputian
Ba5 prepares ...Bb5.

8
tM*+l+mT
8
tMvWl+mT 7
+oWv+oOo
7
Oo+*+oOo 6
o+*+o+*+
6
*+*+o+*+ 5
+*+oP*+*
5
V*OoP*+* 4 *P*O*P*+
4 *+*P*+*+ 3 P*+*+n+*
3 P*N*+*+* 2 *+p+*+pP
2 *Pp+*PpP 1 R*BqKb+r
1 R*BqKbNr a b c d e f g h
a b c d e f g h

White to move
White to move
11.a4!?
There are numerous sects in chess
openings, all of which differ in their Gaining queenside space and stopping
doctrines. This line is GM Lputian’s the black bishop from accessing b5 or
specialty, yet still risky against a known a4. As also indicated by his next few
French-killer like Anand. moves, Anand was in no rush to regain
the pawn on d4.
6.b4! cxd4 7.Nb5 Bc7 8.f4 Bd7
11...Ne7 12.Bd3 Nf5 13.0-0 Nc6 19.h3
14.Qe1 Qb6
Anand calmly creates luft for his king,
Lputian plans to make Anand sweat to ignoring Black’s threat to take on b5.
regain the pawn, but Anand has shown
that he doesn’t mind missing it. All he 19...Nf5
wants in return is to dominate the dark
The b-pawn was poisoned:
squares.
19...Bxb5?? 20.Qb4 Ra5 21.Nxd4!!
15.Rb1?! Nc6 (21...Qxd4 22.Bxc4 wins the rook
on a5) 22.Nxc6 bxc6.
15.Ba3! Ne3 16.b5! is winning for
White. 8
*+*+l+*T
15...Nce7 16.b5
7
+*+*+oOo
6
*Wo+o+*+
White gains space and deprives Black
of access to the c6-square, restricting the
5
Tv+oP*+*
mobility of both the knight on e7 and the
4 *Qm+*P*+
bishop on d7. 3 +*+b+*+p
16...axb5 17.axb5 Ne3 18.Rf2 Nc4
2 *+p+*Rp+
1 +rB*+*K*
a b c d e f g h
8
t+*+l+*T
7
+o+vMoOo
6
*W*+o+*+ White to move
5
+p+oP*+* Black is up a pawn, and everything is
4 *+mO*P*+ seemingly held together. But the
3 +*+b+n+* construction is more fragile than
2 *+p+*RpP appearances indicate: 23.Bxc4! dxc4

+rB*Q*K*
1 24.Be3! Qc7 25.Rd2 when Black is
hopelessly lost. His king is trapped in
a b c d e f g h the center, and the opposite-colored
bishops only serve to ensure White’s
triumph. A cute trick would be: 25...Ra4
White to move 26.Qc5 Qe7 27.Qxc6+!, reminding
Black of his back-rank weakness. pawn was the only pawn on the entire
black kingside holding the dark squares.
20.Re2 h5 21.Qf2 Ra4 22.Re1 g6
23...exf5 24.Nxd4

8
*+*+l+*T
7
+o+v+o+*
8
*+*+l+*T
6
*W*+o+o+
7
+o+v+o+*
5
+p+oPm+o
6
*W*+*+o+
4 t+mO*P*+
5
+p+oPo+o
3 +*+b+n+p
4 t+mN*P*+
2 *+p+*Qp+
3 +*+*+*+p
1 +rB*R*K*
2 *+p+*Qp+
a b c d e f g h 1 +rB*R*K*
a b c d e f g h

White to move
Black to move
23.Bxf5!
White has a lot to be happy about and
Why now? This option for White has holds a serious edge, for the following
existed since move 12. As fate would reasons:
have it, it was Lputian’s last move, 1) White finally regained his sacrificed
overprotecting the knight on f5, that pawn.
changed things enough for Anand to
2) White owns more space across the
decide to take the knight. Behind that
board.
counterintuitive connection (why take
3) White’s knight sits in an
the knight when it is protected more
times?) lies an ingenious positional unassailable spot on d4.
realization. 4) Black’s kingside dark squares are
Taking the knight leads the game to a punctured, and the toxic waste on that
color continues to grow.
position with opposite-colored bishops,
and the g-pawn being on g6 instead of 5) Due to number 4, Black’s king will
g7 seriously enhances White’s potential be perpetually unsafe.
to dominate on the dark squares. The g7-
24...0-0 25.c3 Rc8 26.Be3!
 Introducing ideas of Nxf5. improving moves like this one.

26...Nxe3 28...Ra3

This swap accentuates the good versus Black’s only hope is to try and tie his
bad minor piece dynamic. opponent down to the defense of the c3-
One could advocate for a defensive pawn.
set-up that doesn’t involve taking on e3,
such as 26...Qd8 27.Kh2 Qe7. 29.Rc1 Rc4
However, White can smartly regroup
Going immediately after the pawn on
with: 28.Re2 Be6 29.Qe1! Rca8 30.Bf2!.
c3 with 29...Qa5 can be answered by
The dark-squared bishop is about to
simply ignoring it: 30.Qh4!. Now taking
come to h4 and invade the black camp
on c3 loses in every tactical way
through the dark squares, showing its
possible and 30...Qd8 is the only move
true strength.
that keeps Black in the game, but after
27.Rxe3 Be6 31.Qf6! it still looks like pure misery.

30.Rc2 Kh7 31.Rf3 Qc7?!

8
*+t+*+l+ Black shouldn’t allow White’s queen
7
+o+*+o+* entry to the kingside dark squares.
6
*W*+v+o+ Correct was 31...Qd8!.
5
+p+oPo+o
4 t+*N*P*+
8
*+*+*+*+
3 +*P*R*+p
7
+oW*+o+l
2 *+*+*Qp+
6
*+*+v+o+
1 +r+*+*K*
5
+p+oPo+o
a b c d e f g h
4 *+tN*P*+
3 T*P*+r+p
2 *+r+*QpK
White to move 1 +*+*+*+*
a b c d e f g h
28.Kh2

White’s advantage is permanent, and

there is time for multiple small White to move
32.Qh4! but after 35.Rd2 Qxf6 36.exf6 Black is
busted anyway. The pawn moving from
The walls around the black king start e5 to f6 helps restrict the black king and
to close in on him. clears the e-file for the white rooks.
32...Qd7 8
*+*+t+*+
Once again, the c3-pawn is poisoned: 7
+o+w+o+l
32...Raxc3? 33.Nxe6 fxe6 34.Rfxc3
Rxc3 35.Rxc3 Qxc3 36.Qe7+ Kh8
6
*+*+vQo+
37.Qf8+ Kh7 38.Qf7+ Kh6 39.Qg8!
5
+p+oPo+o
Qc7 40.Qh8+ Qh7 41.Qe8 and Black is 4 *+tN*P*+
busted in the queen ending after the e6- 3 +*P*+*Rp
pawn falls. 2 *+r+*+pK
+*+*+*+*
Best is 32...Raa4!, threatening to chop 1
the knight on the next move. After
33.Rd2 Ra8 it isn’t clear how White can a b c d e f g h
improve his position.

33.Rg3!
White to move
Attackers begin to gather menacingly
35.Ra2!
around the black king.
The final white piece is activated, and
33...Ra8 34.Qf6
the situation starts becoming
It’s a frightening thing when the insufferable. Black is faced with a
opponent’s pieces begin to cross the “damned if you do, damned if you
demilitarized zone into our territory. don’t” situation: if he retains queens on
Alarm bells must have been ringing for the board, he risks getting mated; if he
Lputian, since White threatens a swaps queens, the ending might be lost,
combination of Nxe6 and Rxg6 (in and simplified positions usually benefit
either order), crashing through. the side holding a permanent advantage.

34...Re8 35...Qe7!?

Protecting against White’s threat and One concern (White’s threat of a

hinting at ...Qe7 next. winning attack) is eased, while another
arises to fret about (a lost ending for
34...Qd8?! manages to trade queens,
Black). Without the queens, even the
white king could start contemplating 8
*+*+t+*+
entering through the dark squares. 7
+o+*+o+l
Keeping the queens on the board with
a move like 35...Rg8 looks equally
6
*P*+v+o+
depressing. White will play 36.Ra7,
5
+*+oPo+o
keeping an eye on one extra weakness 4 *+tN*P*+
inside the black camp, and then start 3 +*P*+*Rp
r+*+*+pK
thinking how to go about breaking 2
through.
1 +*+*+*+*
36.Qxe7 a b c d e f g h

Keeping the tension between the

queens was also perfectly sensible.
White to move
36...Rxe7 37.b6!
38.h4!!
The b7-pawn becomes a fixed target.
A technical detail of immense
37...Re8? importance. By fixing all his pawns on
dark squares, Anand not only fixes his
It was Black’s last chance to include opponent’s pawns on light squares, but
37...h4, and keep some instability inside also guarantees the g3-square for his
the white camp. rook to target g6, and the possibility for
the white structure to get stabilized with
g2-g3. In an objectively perfect world,
where no mistakes are made, this would
be the differentiator between White
winning or drawing this game.

38...Kg7 39.Ra7 Rb8 40.Kg1!

 Black had ...Bd7-e8 against this trick,
8
*T*+*+*+ keeping the kingside together.
7
Ro+*+oL* 44.Re3!
6
*P*+v+o+
5
+*+oPo+o The rook is no longer needed on the

*+tN*P*P
4 kingside.

3 +*P*+*R* 44...Ke7 45.Re1 Bc6 46.Rea1

2 *+*+*+p+ Heading towards a5 and c5, further
1 +*+*+*K* underscoring the inability of the black
a b c d e f g h army to control a single relevant dark
square.

46...Bd7 47.R1a5!
Black to move
Tying the bishop down to defense of
Principle: The king is a fighting piece d5.
in the ending. Anand plans to transfer
him to the queenside. 47...Be6 48.g3 Kd8 49.Kc2 Ke7
50.Kb2 Kd8
40...Bd7 41.Kf2 Kf8 42.Ke2?

Oops! Correct was starting with

42.Rd3 so that 42...Ra4 can be met with
8
*TtL*+*+
43.Rd1 Rxa7 44.bxa7 Ra8 45.Ra1, and
7
Ro+*+o+*
White wins. 6
*P*+v+o+
42...Ke8?
5
R*+oPo+o
4 *+*N*P*P
OK, no one noticed. 3 +*P*+*P*
The cruel engine points out 42...Ra4!, 2 *K*+*+*+
+*+*+*+*
breaking White’s construction. 1
43.Kd2 Rcc8 a b c d e f g h

Now 43...Ra4 fails to 44.e6! and the

g6-pawn falls, with decisive effect. On
the previous move, with the king on f8, White to move
 Exercise: We clearly sense the Rcc8.
wretchedness of Black’s position,
especially the pale, scrawny bishop, 8
*T*L*+*+
since all of Lputian’s pieces are tied
down to passive defense. Black is
7
+o+*+o+*
clearly playing for a fortress draw.
6
*P*+v+o+
Come up with a clear plan for 5
R*+oPo+o
White to make progress: 4 *+tN*P*P
3 +rP*+*P*
Answer: Step 1: Double rooks on the
fifth rank, intending to swap off a pair of
2 *K*+*+*+
rooks via c5. If this happens, Black’s
1 +*+*+*+*
chances of a fortress draw fall to zero, as a b c d e f g h
the white king gains the ability to
infiltrate the black camp.

51.Rb5! Rc4 52.Raa5! Rbc8 White to move

54.Rb4!
Black cannot allow the intended swap.
Soon, Black learns that a fate postponed
Proposing a trade forces Black’s
is not the same thing as a fate changed.
reaction.
To illustrate, 52...Kd7? loses without a
fight: 53.Rc5 Rc8 54.Rxc4 Rxc4 55.Ra8 54...Rcc8 55.Rbb5!
Rc8 56.Rxc8 Kxc8 57.Ka3 Kd7 58.Kb4
Ke7 59.Kc5 Kd7 60.Nb5. The white But now there is no stopping the Rc5
king-and-knight duo will pick apart the plan. White wins.
black pawns one by one with apparent
55...Ke7 56.Rc5! Kd7 57.Rxc8 Rxc8
ease.
58.Ra7 Rb8 59.Kb3 Kd8 60.Kb4 Kd7
53.Rb3 61.Ra5! 1-0

Since the trade isn’t allowed to happen White’s threat to infiltrate c7 with his
on c5, let’s do it on a8. White’s threat is rook forces the rooks off the board,
Rba3 followed by Ra8. which, as we have seen, is an easy win
for White, so Black resigned.
53...Rb8 This game could have also easily gone
into the weak piece chapter. If we look
Black gets ready to answer Rba3 with
closely, the root cause of Black’s defeat
was the chronic weakness on the dark Nepomniachtchi
squares. A clear understanding of the Yekaterinburg 2020
source of our misery is not very useful if
there is no solution. We get the feeling 1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.e5 c5 5.a3
that GM Lputian was perfectly aware of Bxc3+ 6.bxc3 Ne7 7.h4 Qc7 8.h5 h6
his strategic difficulties, yet he was 9.Rb1 b6 10.Qg4 Rg8 11.Bb5+ Kf8
helpless to correct them. 12.Bd3 Ba6
Let’s go back to one of the pivotal
moments of this game: Anand’s
23.Bxf5! immediately after Black’s
8
tM*+*Lt+
22...g6. Imagine how different the game
7
O*W*MoO*
would have been if Anand had taken on 6
vO*+o+*O
f5 before the black pawn moved to g6.
Nothing would be the same. There
5
+*OoP*+p
would be no invasion through the dark
4 *+*P*+q+
squares, no queen on f6, no perpetually 3 P*Pb+*+*
weak black king – it is fascinating how, 2 *+p+*Pp+
in chess, such small moves can have so
far-reaching implications. Anand
1 +rB*K*Nr
a b c d e f g h
immediately realized that the pawn on
g6 created a perpetual weakness on the
dark squares for Black, and he seized the
opportunity to exploit it to the fullest. White to move

In the following game, MVL’s 13.dxc5!

opening play would be difficult to
Confusion is one of the unpleasant
explain to a player who just started
side effects of this line. Here are some
chess, since he first refused to develop
of the reasons behind White’s radical
his pieces, preferring to push pawns, yet
decision:
White still had a slight advantage. What
can I say? Chess is a difficult game! We 1) White already has his rook posted
could try to explain what’s going on, but on b1, which means he is happy to open
it falls outside the scope of this book, so the b-file.
let’s move on immediately to the first 2) The structural weakness may only
position that falls under our jurisdiction: be temporary for White, since Black will
likely be induced to swap on d3 soon,
Game 30
fixing White’s structure.
Maxime Vachier-Lagrave – Ian
 3) Black had a strategic threat of
...Bxd3 followed by ...c5-c4, which 8
t+*+*Lt+
would create a hole on c4. By 7
O*WmMoO*
*+*+o+*O
exchanging on c5, White avoids this 6
idea.
4) White is not afraid of ...Qxe5+,
5
+*OoP*+p
since Black is far too far behind in 4 *+*P*+q+
development to be grabbing such pawns. 3 P*P*+*+*
13...Bxd3
2 *+*+*Pp+
It seems crazy to straighten out
1 +rB*K*Nr
a b c d e f g h
White’s pawns, yet there is good reason
for it.
Trying to play against the doubled
isolated pawns would look something White to move
like this: 13...Qxc5?! (keeping the c-file
16.Qd1!?
open to highlight the weakness of the
pawns) 14.Ne2 Qc8 (protecting the a6- Doesn’t anyone believe in
bishop to facilitate the development of development anymore? The players had
the knight on b8, while also moved quickly, and both were still
safeguarding the queen against White’s clearly in their home preparation. In
easy tempo-gaining moves) 15.a4 Nbc6 terms of logic, one could claim that the
16.Qf4. Up next is Ba3, and even Bg6 queen had done all she could on the
becomes a threat. Black has nothing to kingside, and now that the play has
do other than swap on d3, but then what shifted more to the other side of the
was the point of his previous play? In board, d1 is a better square than g4.
essence, the d3-bishop is just too good a
piece to let it live for too long. 16...Qa5 17.Bd2!

14.cxd3 Nd7 15.d4 bxc5 The a3-pawn does nothing for White
except obstruct his own pieces.

17...Rb8!

Nepo takes the safer and wiser route.

17...Qxa3?! offers White excellent
compensation after 18.Ne2 Qa6 19.0-0.
18.Ne2 c4? influence over the central and kingside
dark squares.
Black can stay close to equal by 4) Black’s largest problem is the
retaining structural fluidity with awkward placement of his king. If the
18...Rxb1! 19.Qxb1 Qa6!. The dream rook on g8 could magically jump over
for Black is to one day post a knight on the king, Black would immediately
c4 instead of a pawn. solve all his problems.
5) White has access to the only real
8
*T*+*Lt+ pawn break in the position, on f5.
7
O*+mMoO* Black’s break on f6 is unlikely, since it
6
*+*+o+*O would expose the black king. This last
5
W*+oP*+p statement is even more important than it
might seem at first. Whichever side is
4 *+oP*+*+ able to control the pawn breaks is the
3 P*P*+*+* side that chooses when the two armies
2 *+*BnPp+ come into contact and that is a huge

+r+qK*+r
1 advantage, often playing the deciding
factor on its own.
a b c d e f g h
19.0-0 Rb6 20.Qc2 Rh8

The rook paces like a bored zoo

White to move animal in its enclosure.
Let’s discuss the structural shift which 20...Ra6! is a better attempt, trying to
took place with Black’s last move: suppress White’s a3-a4 plan and not
fearing the immediate Qh7. However,
1) The board is split into two well-
even with the engine’s perfect defensive
defined camps.
skills, Black’s position remains nearly
2) Black controls the light squares,
lost.
while White controls the dark squares.
3) In this case, dark-square control is 21.a4!
the more potent force, and White’s so-
called “bad” bishop may be the best The a3-square is cleared for White’s
piece on the board. That’s because bishop.
White can play for a3-a4, followed by
rerouting the bishop to a3, via c1. From 21...Ke8 22.Rb4
a3, the bishop will exert tremendous
This is somewhat unnecessary, but
isn’t bad, and it is based on a truly next move shocked the hell out of
fascinating idea by MVL. your writer. White’s b4-rook is
After the simpler 22.Rxb6! Nxb6 attacked. A brilliant idea calls out to
23.Ra1 Kd7 24.Bc1 Black is under its future creator, out of the
heavy strategic pressure. darkness. What is White’s most
potent plan in the position?
22...Nc6?!
Answer: Refuse to back down!
Necessary was 22...Nb8!, clearing d7
for Black’s king, trying to finally solve 23.f4!!
the problem of the rook that is trapped
on the kingside. “Am I having a vision, or did White’s
last move actually take place?” I
8
*+*+l+*T exclaimed, after White’s last shocking
7
O*+m+oO* move. It’s a scary thing when the
opponent harnesses the power of rage,
6
*Tm+o+*O approaching with madness in his eyes.
5
W*+oP*+p In this case, it’s a form of divine
4 pRoP*+*+ madness, since MVL’s sacrifice is
3 +*P*+*+* completely sound, highlighting exactly
why the knight shouldn’t have left e7.
2 *+qBnPp+
1 +*+*+rK* 23...Ne7!
a b c d e f g h
Chess is a game with no memory. The
past technically has no bearing on the
present or future. It’s our defective
White to move human nature that makes us believe
otherwise. Nepo realizes the danger and
In the 1950s, Japanese miners awoke admits his mistake.
Rodan (sort of a cousin to Godzilla and After 23...Nxb4?! 24.cxb4 Qa6
Mothra), who then inflicted havoc upon (24...Rxb4 25.Qc3 picks up the rook)
the world. The lesson I learned from the 25.f5 Black’s extra exchange is useless,
movie was: don’t poke around in the and White has a winning attack.
dark, or you risk awakening a terrible
evil. Black’s last move did just that. 24.Rfb1
Exercise: I watched this high-stakes
Candidates’ game live and MVL’s
 8
*+*+l+*T 8
*T*+*+*+
7
O*+mMoO* 7
O*+mMlO*
6
*T*+o+*O 6
wT*+o+*O
5
W*+oP*+p 5
+r+oPo+p
4 pRoP*P*+ 4 p+oP*P*+
3 +*P*+*+* 3 B*P*+*+*
2 *+qBn+p+ 2 *+q+n+p+
1 +r+*+*K* 1 +r+*+*K*
a b c d e f g h a b c d e f g h

Black to move White to move

24...f5! Exercise: We get the sense that

White’s position is loaded with
Nepomniachtchi again finds the unused potential. Come up with
optimal defensive idea. Black stops f4- White’s winning plan:
f5 for good and creates the f7-square for
his king.
Answer: Undermine the defense.
25.Rb5 Qa6 26.Bc1!
28.Bxe7!
The “bad” bishop is about to become
After all the trouble of shifting the
the strongest minor piece on the board.
bishop to a3, White swaps it away to
26...Kf7 27.Ba3 facilitate a breakthrough on f5.

At last. The bishop reaches the desired 28...Kxe7 29.g4!

diagonal, peeking inside the enemy
It becomes clear that Black’s position
camp.
isn’t the most impregnable of fortresses.
27...Rhb8 The f5 point collapses.

29...Rxb5

29...fxg4? 30.Qh7!? is decisive.

30.axb5 Rxb5 31.gxf5! Nepomniachtchi does his best, but
with a depleted arsenal. He just doesn’t
After 31.Rxb5? Qxb5 32.gxf5 Qb3 have enough firepower with the queen
Black is much better off than in the alone. The knight efficiently protects the
game continuation. white king from perpetual check.

31...Rxb1+ 32.Qxb1 exf5 8

*+*+*L*+
7
O*+m+*O*
8
*+*+*+*+ 6
*+*+*+*O
7
O*+mL*O* 5
+*+oP*+p
6
w+*+*+*O 4 *+oP*Pw+
5
+*+oPo+p 3 +*P*+*N*
4 *+oP*P*+ 2 *+*+*+*+
3 +*P*+*+* 1 Q*+*+*K*
2 *+*+n+*+ a b c d e f g h
1 +q+*+*K*
a b c d e f g h
White to move

37.Kg2 Qxf4 38.Qxa7 Ke7

White to move
38...Qg4 allows 39.Qa8+ followed by
33.Ng3!
Qxd5.
When you see something good, look
39.Qa3+
for better.
33.Qxf5?! Qe6 isn’t nearly as clear. It’s fitting that the weakness of the a3-
f8 diagonal and the dark squares in
33...Qb6 34.Nxf5+ Kf8 35.Qa1!
general is felt one last time.
Of course, White would never agree to
39...Kd8 40.Qd6!
a queen trade. Black’s dark squares
continue to leak, and soon his position
begins to spit teeth and blood.

35...Qe6 36.Ng3! Qg4

 jump in front of the pawn chain
8
*+*L*+*+ restricting it and into the opponent’s
7
+*+m+*O* camp, it could cease being bad and

*+*Q*+*O
6 instead become a huge asset. When the
bishop came to a3, Black’s weakness on
5
+*+oP*+p the dark squares became especially
4 *+oP*W*+ pronounced, but take note of the fact
3 +*P*+*N* that the breakthrough did not occur on a
dark square. On the contrary, it was
2 *+*+*+k+ 29.g4! (preceded by 28.Bxe7!) that
1 +*+*+*+* allowed White to break down the black
a b c d e f g h defense. That is a common occurrence.
Since White was dominating the dark
squares, Black tried to erect a fortress on
the light squares. In order for White to
Black to move
win, that fortress had to be broken.
The d5-pawn falls, and the game ends
with it. Time for more comic relief with one
of my games. My attack was flopping,
40...g5 until my student Peter allowed me to
showcase the importance of the
40...Qd2+ is useless for Black after Principle: Opposite-colored bishops
41.Kh3 when there are no more checks. favor the attacking side

41.hxg6 h5 42.g7 1-0 Game 31

Cyrus Lakdawala – Peter Bisgaard
Nepomniachtchi decided he’d seen Internet (rapid) 2021
enough.
Take note of how White’s technically 1.d4 d5 2.c4
bad bishop, with pawns fixed on the
Your writer is normally the apostle of
same color on c3, d4, e5, and f4,
peace. In a “real” game, I would, of
actually became the most potent minor
course, play the winning move 2.Bf4!!,
piece on the board once it reached a3.
prompting Black’s immediate and fully
This means that principles have their
justified resignation!
exceptions.
One such exception would be the 2...c6 3.Nc3 Nf6 4.Nf3 dxc4 5.e4!?
Principle: If a “bad” bishop manages to
Normal is 5.a4.
 of a raging tactician. It’s a sad sight to
8
tMvWlV*T watch a posturing dove putting on airs
7
Oo+*OoOo of being a hawk.
6
*+o+*M*+ 5...b5 6.e5
5
+*+*+*+*
*+oPp+*+
4 We were looking at this variation
during the lesson.
3 +*N*+n+* In top-level chess, attention has started
2 pP*+*PpP to shift towards 6.Be2!? and 6.Qc2!?,
1 R*BqKb+r keeping the white center fluid.
a b c d e f g h
6...Nd5 7.a4 e6 8.axb5 Nxc3 9.bxc3
cxb5 10.Ng5!

Black to move Threatening cheapo ideas like Nxf7

followed by Qf3+ and Qxa8, or just the
An asteroid the size of Los Angeles is immediate Qf3.
hurling towards the Earth, with an
estimated time of impact at 3:00 a.m. on 10...Bb7 11.Qh5!
Friday. It is still Thursday evening and
Inducing dark-squared weaknesses
you have only a few hours to live, so
around the black kingside.
you decide to play some online games.
Are you really going to play soundly? 11...g6 12.Qg4 Be7 13.Be2 h5 14.Qg3
Or are you going to be tempted to have Bd5!
some fun and go for a gambit?
Apparently, the Lakdawalas are a 14...Nd7?? allows 15.Nxe6! when,
courageous clan, which is news to me. suddenly, Black is busted.
I’m normally a safety-first, non-gambit
guy, but this was a lesson with my
student Peter, who wanted to learn the
white side of the Geller Gambit, the
Slav’s no-account, disreputable cousin.
Personally, every time I try to attack, I
feel dishonest, since I am a natural
strategist and in complex positions I
have an aptitude for hanging queens!
Yet here I am, trying to play in the style
 This is too slow, but I didn’t know
8
tM*Wl+*T what else to do. At least, I managed to
7
O*+*Vo+* come up with a plan to threaten
something. My intention was Bc2
6
*+*+o+o+ followed by Nxf7.
5
+o+vP*No
4 *+oP*+*+ 17...b4 18.Bc2
3 +*P*+*Q*
2 *+*+bPpP 8
t+*Wl+*T
1 R*B*K*+r 7
+*+mVo+*
a b c d e f g h 6
*+*+o+o+
5
O*+vP*No
4 *OoP*+*P
White to move 3 +*P*+*Q*
We reach an old tabiya of the Geller
2 *+b+*Pp+
Gambit, which objectively favors Black, 1 R*B*+rK*
but allows White some practical a b c d e f g h
chances.

15.0-0!? Black to move

15.Bf3 is the main line. Exercise: Like George Washington
and Abe Lincoln before me, I
15...a5! cannot tell a lie. White’s
compensation for the sacrificed
Principle: Push forward your pawn pawn is just not enough, and Black
majority as quickly as possible. should be winning if he finds the
optimal defensive plan. What is
16.h4
Black’s best move?
An unintentional novelty and playing a
slow move was not the most logical Answer: Black should hand over his
choice, but I didn’t believe in my precious dark-squared bishop,
position to begin with. eliminating my dangerous knight. You
may ask: doesn’t that weaken Black’s
16...Nd7 17.Bd1!? dark squares? It does, yet White’s
initiative is just not enough. With 26...Ra7! the rook participates in
the defense laterally. After 27.gxh5
18...Rg8?! Qd8! (a queen trade would ensure
Black’s win because of the far-advanced
This is still in Black’s favor, but, as
queenside passers) 28.Qf4 Qf8!
we shall see, it is just prolonging the
(controlling the h6-square to prevent any
inevitable.
tricks associated with Qh6+) 29.h6 f6!
Black will at some point have to take (clearing the seventh rank and clogging
that knight with his bishop, so he might the dark squares so that the white queen
as well do it immediately: 18...Bxg5! can’t cross through; the main threat is
(it’s psychologically difficult to make a ...a4-a3) 30.exf6 Qb8!, with a double
move like this, since it weakens the dark threat of a queen trade and ...b3-b2,
squares around Black’s king) 19.Bxg5 Black wins.
Qc8 and the game becomes concrete.
Will White be able to drum up an attack 19.Ba4!
before the black pawns on the queenside
reach the first rank? Approximately best The g6-pawn is too well protected, so
play for both sides would be: 20.Ba4! b3 I switch the light-squared bishop to a
21.Qf4 Kf8! 22.Bxd7 Qxd7 23.Qf6 more active diagonal before Black shuts
Kg8 24.f3 a4 25.Bc1 (stopping ...a4-a3) it down with ...b4-b3.
25...Kh7 26.g4.
19...Kf8
8
t+*+*+*T The engine doesn’t mind this, but I’m
7
+*+w+o+l not sure if it’s the most practical
6
*+*+oQo+ approach.
5
+*+vP*+o 19...Rg7!? is an interesting try to
abstain from the capture on g5, but the
4 o+oP*+pP rook looks a bit ridiculous on g7.
3 +oP*+p+*
2 *+*+*+*+
1 R*B*+rK*
a b c d e f g h

Black to move
 21...Kg7 22.Nf6
8
t+*W*Lt+
7
+*+mVo+* 8
t+*+*+t+
6
*+*+o+o+ 7
+*+wVoL*
5
O*+vP*No 6
*+*+oNo+
4 bOoP*+*P 5
O*+vP*+o
3 +*P*+*Q* 4 *OoP*+*P
2 *+*+*Pp+ 3 +*P*+*Q*
1 R*B*+rK* 2 *+*+*Pp+
a b c d e f g h
1 R*B*+rK*
a b c d e f g h

White to move

20.Bxd7! Black to move

My only hope of an attack rests on the Exercise: Should Black accept

dark squares, and my light-squared White’s pawn offer, or should he
bishop cannot help in that mission. The decline?
best it can do is trade itself for the
knight, a crucial dark-square defender Answer: He should decline, since
for Black. accepting the pawn allows White a
wicked attack based on the presence of
20...Qxd7?!
opposite-colored bishops.
Peter had to make the tough call of
22...Bxf6??
inserting 20...Bxg5! before taking on d7.
It looks crazy to do so, but the game The worst possible moment to take
move allows White to create concrete that knight. The pawn that advances to
threats. f6 is too far advanced, and the black
king gets suffocated if the pawn
21.Nh7+!
remains, but if Black recaptures on f6,
I take advantage of Black’s failure to the king can never get back behind the
toss in ...Rg7 or take the knight on g5. pawn chain.
Black should offer the exchange with
22...Qd8!, or another queen move. The
d8-square is best, but it really doesn’t 8
t+*+*+t+
matter, as long as you don’t take that 7
+*+w+o+*
*+*+oLo+
poisoned knight! I don’t believe in 6
White’s position at all. Grabbing the
exchange on g8 leads to a dead-lost
5
O*+v+*+o
position for White. The engine comes up 4 *OoP*+*P
with a mind-boggling solution to keep
White in the game: 23.cxb4!! axb4
3 +*P*+*Q*
24.Nxh5+! Kf8 25.Nf6!! Bxf6 26.Rxa8
2 *+*+*Pp+
Qxa8 27.exf6 Rh8 28.Bd2!! c3 29.Bxc3! 1 R*B*+rK*
bxc3 30.Rb1!. The line continues for all a b c d e f g h
eternity and, in the end, Black keeps a
small advantage. What bearing does that
have on human chess? Nothing at all.
White to move
23.exf6+ Kxf6
24.Qg5+!
Peter intended 23...Kf8 and then
The only winning move.
realized that it allowed a shot that
exploits the weakness of his back rank. During the game I thought that
After 24.Rxa5! White wins. Such is the 24.Qe5+? Ke7 25.Bg5+ was also
nature of rapid chess: you are going to winning, as stepping backwards with the
miss some tactics. However, it could be king allows Rxa5, again exploiting
argued that all the positional Black’s back-rank problems. However, I
preconditions are there for that tactic to preferred to give the check on g5
exist, due to Black’s extreme weakness instead, as, at least intuitively, it looks
on the dark squares. better to push the black king forward,
away from his camp, instead of
backwards, towards the back rank. Of
course, I missed that here 25...f6!!
26.Qxf6+ Kd6 gets the black king to
safety on c6 and is even losing for
White!

24...Kg7 25.Qh6+ Kf6 26.Bg5+ Kf5

 29.Re5+!
8
t+*+*+t+
+*+w+o+*
7 The best attacks are those with a touch
of madness. The f6-pawn is overworked,
6
*+*+o+oQ as taking the rook allows the white
5
O*+v+lBo queen to access the g5-square. It’s a
4 *OoP*+*P forced mate in 14, and to be honest, I

+*P*+*+*
3 was too lazy to work it out to the end.
Our eyes simply tell us that Black’s king
2 *+*+*Pp+ will not be able to survive alone against
1 R*+*+rK* a white queen, rook and bishop.
a b c d e f g h
29...fxe5 30.Qg5+ Ke4

White to move
8
t+*+*+t+
7
+*+*+*+*
*+*Wo+o+
Step by queasy step, the black king is 6
forced to march to his own funeral.
5
O*+vO*Qo
27.Rae1 4 *OoPl+*P
It doesn’t matter too much, but this
3 +*P*+*+*
rook is the most logical of the two to 2 *+*B*Pp+
use. There is an outside chance of the
f1-rook becoming useful in the
1 +*+*+rK*
a b c d e f g h
checkmate process, while there was no
such chance for the rook on a1. The
obvious threat is Re5+, pushing the
black king further up the board. White to move

27...f6 28.Bd2 31.Re1+

The new threat is Qf4 mate. This move, while still forcing
checkmate, is in violation of the
28...Qd6 Principle: When hunting the enemy
king, don’t chase him; instead, cut off
Also lost for Black is 28...e5 29.dxe5
escape routes.
Ke6 30.exf6+ Kd6 31.cxb4.
In this case, 31.Qg3! cuts the black
king off from d3 and threatens Qf3 sometimes finds the acorn lying around
mate, as well as Re1+ followed by Qg5 on the ground. I finally saw a forced
mate. The only way to stop that would mate in four.
be 31...Kf5, but then 32.Bg5! cuts off
the f6 escape route. There is no defense 34.Bf4+ Kb3 35.Rb1+ Ka4 36.Qc2+ 1-
to the threat of Qh3+ followed by Re1 0
mate or Qf3 mate. Of course, it doesn’t
Peter resigned before I could execute
really matter how you win, as one
36...Ka3 37.Bc1 mate.
cannot win with a score larger than ‘1-0’
in chess, and all wins count the same, Black’s game went from winning to
but still discussing technique shouldn’t losing when he traded off his important
be taboo. dark-squared bishop for my knight on
f6. This game is a good illustration of a
31...Kd3 32.Qe3+ Kc2 33.Qe2! well-known Principle: In positions with
opposite-colored bishops, an attack
Controlling some light squares so that against the opponent’s king will
the king doesn’t accidentally escape completely overshadow any other
through the route c2-b3-a4-b5. positional aspect.
Were my light squares in this game
33...bxc3
weak? Sure. Would I lose without a
fight because of the missing pawns if we
8
t+*+*+t+ gave Black a tempo to safeguard his
king? Absolutely. However, nobody
7
+*+*+*+* would ever notice the ‘weakness’ on
6
*+*Wo+o+ White’s light squares if the black king is
5
O*+vO*+o on f5.
4 *+oP*+*P
+*O*+*+*
3 When you get old, you live in a
universe of non-bewilderment, where
2 *+lBqPp+ you feel like you have seen it all and
1 +*+*R*K* there are no more surprises for you. The
a b c d e f g h following ending was an exception, full
of anomaly and geometric shock!
Game 32
Fabiano Caruana – Teimour
White to move
Radjabov
Even the elderly, blind chipmunk Madrid 2022
 soon as possible.
8
*L*+*+*+
+oT*ToO*
7 29...f6?

6
o+*+*+*O How can it be a mistake to reinforce
5
+*+*Mp+* the knight? Well, Black, on his last

*+*+b+*+
4 move, just created a hole on g6 and
weakened the kingside light squares.
3 +*P*+p+* “Which white piece is going to exploit
2 p+*R*KpP the hole?” you may ask. It is White’s
1 +*R*+*+* king, who can walk into g6 via g3, h4,
and h5. That single-handedly changes
a b c d e f g h
everything, since from an almost useless
bystander, the king becomes a heroic
protagonist, allowing White to attack
Black to move with a full extra piece.
After 29...Ng4+! 30.Kg3 Nf6 it’s
We first gather data:
unlikely that Caruana would have been
1) White is up a pawn.
able to make use of his extra pawn.
2) The extra pawn isn’t all that great,
since it’s doubled. White will not be 30.Kg3! Ka7 31.Rd4 Rc5
able to create a passed pawn on the
kingside with that hobbled majority.
3) White’s queenside pawns are split,
8
*+*+*+*+
and there is potential for them to 7
Lo+*T*O*
become weak. 6
o+*+*O*O
4) White owns a bishop versus a
knight, yet we sense that the knight
5
+*T*Mp+*
should hold its own since the structure is
4 *+*Rb+*+
mostly stable. 3 +*P*+pK*
Exercise: Black should hold the
2 p+*+*+pP
game, but only if he finds the 1 +*R*+*+*
correct plan. How would you a b c d e f g h
continue for Black?

Answer: Black’s knight is optimally

White to move
posted on f6 and should get there as
32.Kh4!
8
*+*+*+*+
L*T*+*O*
White’s king is a blade, just looking 7
for some juicy ribs it can jam itself into.
6
o+*+*O*O
32...Rec7 33.Kh5 b5 5
+oT*Mp+k
33...Rxc3?? isn’t much of a
4 *+*Rb+*+
consideration for Black. After 34.Rxc3 3 +*P*+p+*
Rxc3 35.f4 Nc6 36.Rd7 Na5 37.Rxg7
Ra3 38.Bd5 White wins prosaically.
2 p+*+*+pP
Trying to exploit the connection
1 +*R*+*+*
a b c d e f g h
between the c5-rook and the king on h5
also fails: 33...g6+ 34.fxg6! (34.Kxh6 is
what Black was hoping for, as after
34...gxf5 the king on h6 is in a bit of White to move
trouble: for example, 35.Bxf5 Nf7+
36.Kg6 Nh8+ 37.Kxf6 Rf7+; Black 34.f4!
wins a piece for the white pawns and
should be able to draw the game) The defender of the g6-square receives
34...Nd3+ (34...Nxf3+ is met with its eviction notice.
35.Rd5!) 35.Kxh6 Nxc1 36.g7 Rg5
34...Nd7 35.Rcd1 Nb6 36.R1d3 Na4
37.Bg6! Rc8 38.g4!?. Other ways exist,
but I like that White doesn’t need to rush 36...Rxc3? loses time compared to
things. Next up are h2-h4, Rd6, Rxf6, Radjabov’s approach and is thus much
and Rf8. The armada of white pawns on easier to refute. The obvious moves do
the kingside cooperates perfectly with the trick: 37.Rxc3 Rxc3 38.Kg6 Rc7
the pieces. Black is busted. 39.Rd8 a5 40.Rg8 b4 41.Rxg7 Rxg7+
42.Kxg7 a4 43.Bc2 b3 44.Bxb3! axb3
45.axb3. Black’s puny king and knight
have no chance of holding off the white
pawns.

37.Kg6 Nxc3 38.Bf3!

Caruana realizes that Black’s two

connected queenside passers are actually
slower than White’s future passed
pawns on the kingside. Should White be alarmed? His entire
queenside was wiped out of pawns. Our
38...Nxa2 eyes can sometimes lie to us. Believe it
or not, White promotes first, for the
By using his knight to take on c3,
following reasons:
Radjabov has managed to cling to his
1) Even though White has yet to create
g7-pawn for a good while longer, while
a passed pawn on the kingside, both the
also removing the pawn from a2. A
g7- and f6-pawns will fall to the white
good fighting try, but his position
king.
remains lost anyway.
2) Then, White’s forward f-pawn is
8
*+*+*+*+ the fastest runner in the promotion race.
3) The white bishop could potentially
7
L*T*+*O* use its long-range abilities to help both
6
o+*+*OkO causes of pushing the white pawns
5
+oT*+p+* forward and stopping the black pawns,
4 *+*R*P*+ while Black’s knight can do no such
thing.
3 +*+r+b+*
2 m+*+*+pP 39...Nc3 40.Kxg7 b4 41.Kxf6 b3
1 +*+*+*+* Without analysis, we feel in our bones
a b c d e f g h that Black is faster. Our bones are
wrong!

42.Rd2!
White to move
After this disarmingly simple move,
39.Rd7!
Black experiences a difficult time in his
As a kid, I was a master of shortcuts. I attempt to promote.
wrote a fifth-grade book report on H.G.
42...a5
Wells’ The Time Machine and proudly
scored an A-grade. Except I was too With 42...Na4 43.Kg6 b2 44.Be4 Nc3
lazy to read the actual book and 45.Rxc7+ Rxc7 46.Bd3 Black will win
substituted it with the Classics the white bishop for the b-pawn, but will
Illustrated comic-book version! lose much more than just a piece to stop
Promotion races work exactly the same all of White’s passers on the kingside.
way. Efficiency and speed matter.
 47.f8=Q
8
*+*+*+*+
7
L*Tr+*+* 8
*+*+*Q*+
6
*+*+*K*O 7
+*+r+*+*
5
O*T*+p+* 6
l+*+k+*O
4 *+*+*P*+ 5
O*T*+*+*
3 +oM*+b+* 4 *+*+*P*+
2 *+*R*+pP 3 +*M*+b+*
1 +*+*+*+* 2 *+*+*+pP
a b c d e f g h
1 +w+*+*+*
a b c d e f g h

White to move

43.Ke6! Black to move

43.Kg6?! is also winning, but it’s not The position is an analytical

logical to go to g6 when Black’s b-pawn nightmare. Unfortunately for Radjabov,
is about to queen, as the newly promoted Caruana may be the most accurate pure
queen could potentially be born with calculator in the world. This is a rare
check. case of the side promoting first losing
anyway. Normally, there is the
43...Rxd7 Principle: In a mutual promotion race,
the side which promotes a new queen
43...a4 is also too slow: 44.f6 a3 45.f7
first wins by means of a direct attack. In
b2 46.f8=Q b1=Q 47.Qxc5+ and Black
this case, White’s king is surprisingly
can resign.
safe, and it is the second f-pawn that
44.Rxd7+ Ka6 45.f6 b2 46.f7! b1=Q wins the game.

It turns out that the king being on e6 47...Qg6+

instead of g6 is quite the relevant detail.
After 47...Qb3+ 48.Kf6! it is the black
46...Rc8? loses to: 47.Bb7+ Kb6 king who won’t survive, despite the fact
48.Bxc8 b1=Q 49.Rb7+. White wins that it is Black’s move.
Black’s new queen and the game.
48.Qf6 Qe8+ 49.Kd6! Rc4
 can win in a multitude of ways from
8
*+*+w+*+ here – the attack against Black’s king
7
+*+r+*+* combined with the far-advanced f-pawn

l+*K*Q*O
6 will be more than enough.

5
O*+*+*+* 51.Qe5+!
4 *+t+*P*+ Welcome to Promotion Race, Part II.
3 +*M*+b+* White’s passed e-pawn will be faster
2 *+*+*+pP than Black’s a-pawn.
1 +*+*+*+* 51...Qxe5+ 52.fxe5 Rxc7 53.Kxc7 Kc5
a b c d e f g h
After 53...a4 54.e6 a3 55.e7 a2
56.e8=Q+ White promotes with check
and wins.
White to move
54.e6 Nb5+ 55.Kd7 a4 56.e7 1-0
50.Rc7!!
Radjabov resigned in view of:
Caruana is a master of distorting
56...Nd6 57.Bd1 a3 58.Bb3 h5 59.h4!.
reality. What the...! Has this crazy move
Zugzwang!
been peer-reviewed? White’s rook walks
directly into a knight fork. This game teaches us two things:
1) Even the creation of an
50...Kb5 imperceptible weakness, like the one
Black created on his kingside light
What can we say about 50...Nb5+? squares with 29...f6?, can later come
Sometimes the most literal interpretation back to haunt us. The white king used
is not the best one. This allegedly those squares to infiltrate the opponent’s
“winning” fork, in reality, loses instantly camp. When His Majesty reached g6, it
for Black: 51.Kd5+! White meets a meant that White was attacking with a
check with a check, and Black is mated. full extra piece.
Instead, 50...Ne4+! comes quite close, 2) The position on move 39 is an
but closer analysis reveals that it also optical illusion. To the naked eye,
fails for Black: 51.Bxe4 Qxe4 52.f5 a4 without analysis, it may appear that
53.Qe6! Qd4+ 54.Ke7+ is check! Black is faster, since Black owns two
White’s rook isn’t hanging at all. There connected passed pawns, while White
is still some work to be done, but White lacks even a single passer. What
analysis shows is that the g7- and f6-
pawns fall in rapid succession, and then 8
t+*W*Tl+
White’s f-pawn is faster than both of 7
+*O*+oOo
o+m+v+*+
Black’s queenside passed pawns. Don’t 6
trust first impressions! Calculate and
judge for yourself.
5
+oVoP*+*
4 *+*+m+*+
Strategists and tacticians are 3 +*P*+n+*
pPbN*PpP
hereditary enemies. In those long-past 2
days when I would play blitz against the
13-year-old IM Hikaru (somehow, it’s
1 R*Bq+rK*
not easy to visualize Nakamura as a a b c d e f g h
mere IM!), he was already an
extraordinarily gifted tactician.
Fortunately for me, he was also a poor
Black to move
positional player (for an IM). Of course,
that has changed with age, and the 11...Nxf2
following game is a good example of
how Nakamura developed into one of This is just the start of this line. Black
the strongest pure strategists in the gives up two pieces for a rook and
world. Just watch how he subtly exploits pawn, which, materially, isn’t enough.
Caruana’s weak dark squares. The compensation lies in White’s
Game 33 exposed king position and the fact that
Hikaru Nakamura – Fabiano Black will soon open the f-file with ...f7-
Caruana f6.
Madrid 2022
12.Rxf2 Bxf2+ 13.Kxf2 f6 14.Nf1!
1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6
Without prior knowledge, this would
5.0-0 Nxe4 6.d4 b5 7.Bb3 d5 8.dxe5
be a tough decision. A human wouldn’t
Be6 9.c3 Bc5 10.Nbd2 0-0 11.Bc2
like to give Black two strong pawns in
the center on d5 and e5. However, the
engine taught us that the time saved by
not dealing with the tension allows
White to finish development.
14.exf6 Qxf6 is known to offer Black
decent compensation for the slight
material sacrifice. squares.
After 19...Nxb3? 20.axb3 Rf7?!
14...fxe5 15.Kg1 Qd6 21.Bc5 Qf6 22.Qg3 Re8 23.Re1 White
had a winning position in Malicka-Mei,

t+*+*Tl+
8 San Jose 2022.
7
+*O*+*Oo 20.Re1 c5!?
6
o+mWv+*+
5
+o+oO*+* 8
*+*T*Tl+
4 *+*+*+*+ 7
+m+*+*Oo
3 +*P*+n+* 6
o+*W*+*+
2 pPb+*+pP 5
+oOoOv+*
1 R*Bq+nK* 4 *+*+*+*+
a b c d e f g h 3 +bP*Bn+*
2 pP*+*QpP
White to move
1 +*+*RnK*
a b c d e f g h
16.Be3!

The engine’s top choice, over 16.Bg5.

White to move
I feel that e3 is the more accurate post,
since the bishop can shift over to g3 Caruana activates his central majority
anyway, via f2. Also, on e3 it covers c5 with this theoretical novelty. Some of
and discourages Black’s ...d5-d4. my students watch a few online videos
and then believe they are the world’s
16...Bf5 17.Bb3
leading experts on a line. That’s the
An interesting alternative would be same as assuming they are great military
17.Bxf5!? Rxf5 18.Ng3 Rff8 19.b4!, experts because they watched all the
trying to control some more dark Rambo movies. At the top level,
squares. exceptional memory is a required
survival skill. It’s quite common for
17...Rad8 18.Qe1 Na5 19.Qf2 Nb7! players to bang out a novelty between
moves 20 and 30!
Black needs to fight for the dark
21.Ng3 from the center and keeping the dreams
alive for a future ...d5-d4.
The engine suggests 21.Bg5! c4!
22.Bxd8 cxb3 23.Bh4 bxa2 24.Ra1 Bb1, 23.Bd1!
assessing it as a slight advantage for
White. However, claiming a “slight Principle: When you own the bishop
advantage” for anyone here makes no pair, try to avoid swapping one of the
sense for a human, and it’s sensible to bishops off.
avoid stuff like this when playing
23...Rd7 24.Bf2
against a prepared Fabiano Caruana.
The idea is to shift the bishop over to
21...Bd3 22.Qd2
g3 to provoke Black into pushing his e-
pawn forward.
8
*+*T*Tl+ 24...Rdf7
7
+m+*+*Oo
6
o+*W*+*+ Threatening ...e5-e4, winning a piece.
5
+oOoO*+*
4 *+*+*+*+
8
*+*+*Tl+
3 +bPvBnN*
7
+m+*+tOo
2 pP*Q*+pP
6
o+*W*+*+
1 +*+*R*K*
5
+o+oO*+*
a b c d e f g h
4 *+o+*+*+
3 +*Pv+nN*
2 pP*Q*BpP
Black to move
1 +*+bR*K*
a b c d e f g h
22...c4?!

This offers Black some stability and

tempts our human intuition. However, White to move
stability isn’t what Black wants.
Stability benefits minor pieces; rooks 25.Nh1!!
crave chaos and open lines.
Planning is the art of transforming the
In that regard, it was better to go for abstract into concrete. What a strange
22...e4!, kicking the f3-knight away
way to pick a fight! This powerful 26.Nd4
retreat is a move riddled with
contradictions, since it is drummed into
our heads to centralize our pieces.
8
*+*+*Tl+
However, stability means the world to 7
+m+*+tOo
White’s minor pieces, and Nakamura
somehow manages to hold them
6
o+*W*+*+
together. There are two main points
5
+o+o+*+*
behind this externally weird but
4 *+oNo+*+
internally beautiful move: 3 +*Pv+*+*
1) White defangs Black’s threat to 2 pP*Q*BpP
+*+bR*Kn
play ...e5-e4 by protecting his f2-bishop. 1
2) The g3-square is cleared for Bg3.
a b c d e f g h
The bishop on g3 would put too much
pressure on the e5-pawn for it to stay in
place. And if White provokes his
opponent into ...e5-e4, that will ensure Black to move
complete dark-square domination for the
minor pieces and stable weak points on The knight is excellent on d4, but
d4 and e3. White still has to solve a fair number of
problems to untangle. The f2-bishop
25...e4! cannot move because of ...Rf1 mate, and
if the bishop can’t move, then the knight
Caruana realizes that he cannot keep on h1 is also stuck in the corner.
the pawn on e5 forever and correctly
decides to push it sooner rather than 26...Qg6?!
later. For the moment, at least, White is
somewhat disorganized and not ready Caruana probably wanted to vacate d6
for direct confrontation. for his knight.
Trying to keep the pawn on e5 fails The faster 26...Nc5! is correct,
miserably after 25...Re8? 26.Bg3 Rfe7 planning ...Ne6 to trade off White’s best
27.Nf2. The white pieces are stable and piece.
active, while Black’s are all terrible,
27.h4!
confined to passive defense and stuck
watching over the weak pawn on e5. Grabbing space, aiming to harass the
White is winning. black queen, and creating luft for the
white king.
27...Nc5?! White to move

27...h5 was better. 32.Ne6!

28.h5! Principle: Remove your opponent’s

most active piece from the board. In this
Black’s queen is pushed off the g-file, case, that piece is Black’s knight.
and even more squares are controlled for Furthermore, White wins the d5-pawn in
the white pieces to jump to. exchange for his less valuable h-pawn.

28...Qd6 29.Bg4! 32...Nxe6

Nakamura begins to unravel the knot. 32...Rxg4? 33.Nxf8 wins.

29...h6 30.Qe3! Qf4 33.Bxe6+ Kh7 34.Bxd5 R8f5 35.Bc6

Rxh5 36.Bd4 Rhf5 37.Nf2!
The endgame is bad, but keeping the
queens on wouldn’t be much better. The consequence of losing the d5-
After 30...Qf6 31.Qg3! a5 32.Be3 b4 pawn becomes clear. Black is unable to
33.Nf2 White maintains his dark- hang on to his e-pawn as well. White
squaree bind while slowly activating his has more pieces at his disposal. So, if
whole army. they all manage to coordinate and target
the same point in Black’s camp, there’s
31.Qxf4 Rxf4
no chance of protecting it.

37...Rf7
8
*+*+*Tl+
7
+*+*+*O*
6
o+*+*+*O
5
+oMo+*+p
4 *+oNoTb+
3 +*Pv+*+*
2 pP*+*Bp+
1 +*+*R*Kn
a b c d e f g h
 42.Ra1!
8
*+*+*+*+
7
+*+*+tOl 8
*+*+*+*+
6
o+b+*+*O 7
+*+*+tOl
5
+o+*+*+* 6
o+*+*+*+
4 *+oBoT*+ 5
+*+*+t+*
3 +*Pv+*+* 4 bPo+o+*O
2 pP*+*Np+ 3 +*PvB*+*
1 +*+*R*K* 2 *+*+*Np+
a b c d e f g h
1 R*+*+*K*
a b c d e f g h

White to move

38.b4! Black to move

Black is stuck in a bind, so why rush We can see that:

to grab a pawn that will always be there? 1) Black is weak on the dark squares
Nakamura takes the correct technical and his rooks are rendered impotent by
route of improving his position to the White’s dark-squared bishop, whose
maximum before delivering the final blockade on e3 can’t be broken.
blow. 2) Black’s a6- and e4-pawns are both
targets. White can slowly pick Black
38...h5? 39.a4?
apart by using the principle of two
Neither side notices that Black hung weaknesses.
his h-pawn.
42...h3!
White can attack it with 39.Be8, and
Black must allow the pawn to fall since Caruana perfectly understands the
39...R7f5? only makes matters worse by point made above, so he goes to extreme
hanging the exchange after 40.Be3!. lengths to create as much immediate
Trying 40...Rh4 doesn’t help due to confrontation as possible. Staying put
41.g3, winning a full exchange and would lose easily. His last move
leaving White with a clean extra piece. threatens ...h3-h2+ followed by ...Rxf2,
so White needs to react.
39...bxa4 40.Bxa4 h4 41.Be3 R4f5
43.Ra2!
8
*+*+*+*+
+*+*+*Ol
Hikaru deftly dodges a pair of traps: 7
a) 43.gxh3? allows 43...Rf3 44.Re1
Rg3+ 45.Kh2 Rxe3! 46.Rxe3 Rxf2+
6
o+*+*+t+
when White will be lucky to hold. 5
+*+t+*+*
b) 43.Nxh3? is met with 43...Rf1+!. 4 *PoBoKb+
Trading a pair of rooks is always good 3 +*Pv+*+*
r+*+*N*+
for the side fighting against minor 2
pieces, no matter whether we’re
examining situations with two pieces for
1 +*+*+*+*
a rook or exchange sacrifices. a b c d e f g h
Concretely in this case, after 44.Rxf1
Rxf1+ 45.Kh2 Re1! Black gets serious
counterplay with the passed e-pawn. The
Black to move
minor pieces now look lost without the
rook that was keeping them company. What a harmonious set-up for the
white army. Everyone is working their
43...hxg2
hardest, and everyone is protected.
Nakamura’s next few moves are pure 1) White’s king is a working piece,
perfection. while Black’s is out of play.
2) White continues to dominate the
44.Bd1! central dark squares.
3) The a6- and e4-pawns remain
Alertly preventing ...Rf3.
targets for White’s pieces.
44...R7f6 45.Bg4 Rd5 46.Kxg2 Rg6
49...Kg8 50.Bf5 Rh6 51.Ng4! Rhd6
47.Kg3 Bf1 48.Bd4 Bd3 49.Kf4!
52.Ne3 Rb5 53.Bc5

The unrelenting harassment of Black’s

rooks continues.

53...Rf6 54.Ke5!

White’s pieces continue to creep

forward, growing in power.

54...Kf7 55.Nd5!
 8
*+*+*+*+ 8
*T*+*+*+
7
+*+*+lO* 7
+*+*+lO*
6
o+*+*T*+ 6
o+*+*+*+
5
+tBnKb+* 5
+*+nK*+*
4 *Po+o+*+ 4 *Po+*+*+
3 +*Pv+*+* 3 +*Pv+*+*
2 r+*+*+*+ 2 r+*+oB*+
1 +*+*+*+* 1 +*+*+*+*
a b c d e f g h a b c d e f g h

Black to move White to move

Take out your cameras. White’s 59.Be1?!

domination is complete.
Nakamura has played perfectly up to
55...Rxf5+ this point, but the endgame is more
complicated to convert than it first
Equally hopeless is 55...Rh6 56.Bg4! seems, and a few unnecessary
when Rf2+ is a deadly threat. adventures occurred. We will not go into
great depth, as it falls outside our
56.Kxf5 e3+ 57.Ke5 e2
jurisdiction. The engine finds 59.Rxa6!!,
Yes, the e-pawn is far advanced. which upon first sight appears to be a
Black’s problem throughout this game beginner’s error. But it’s not: 59...Re8+
was the lack of control over the dark 60.Kd4 e1=Q 61.Bxe1 Rxe1 62.Ra7+
squares. As fate would have it, the e1- Kg6 63.b5 and White’s b-pawn races
square is a dark square as well. toward promotion, costing Black the
game.
58.Bf2 Rb8
59...Re8+ 60.Kf4?

This is the wrong direction for the

king. Correct is 60.Kd4!.

60...g5+ 61.Kg3 Re6?!

 61...Re5! might have held a destination, and the rest of White’s army
miraculous draw. is free to hunt down the black pawns.
The immediate threat is Kf3 and Be3,
62.Kf2 Rh6 63.Ke3 Re6+ 64.Kf2 Rh6 corralling the g-pawn.
65.Ne3 Rf6+ 66.Kg3 Rf1 67.Ng2 Rf6
70...Bf5 71.Nf3!

8
*+*+*+*+ Now the bishop is pinned, and Rxf5 is
7
+*+*+l+* a hidden threat.
6
o+*+*T*+ 71...Rd6 72.Nd4! Bd3 73.Re5! Kf6
5
+*+*+*O* 74.Nf3! 1-0
4 *Po+*+*+ Controlling e1 a million times, and
3 +*Pv+*K* getting ready to finally grab some
2 r+*+o+n+ pawns. White threatens Re3, followed
1 +*+*B*+* by Nd4 and Nxe2. If Black plays ...Re6,
then White picks off Black’s g-pawn.
a b c d e f g h
Understandably, Caruana saw no point
in continuing the struggle.
When fighting for control over a
White to move single square, two pieces are better than
one, no matter the power of any of those
68.Bf2! three pieces. In that regard, when
playing with two minor pieces against a
White unravels, while retaining his
rook, the two pieces should usually be
extra piece and the iron blockade on e1.
able to create a dominance on a color
The knight is by far the weakest piece in
complex, usually the same color as the
White’s army, especially in this
unopposed bishop. For example, in this
situation, and is typically considered the
game, White had a dark-squared bishop
best blockader. Now we understand
and a knight against a black rook. That
what Nakamura’s last few moves were
meant that the fight for any dark square
all about. He methodically brought his
was bound to be won by White, who
knight back, knowing that it was the
could attack that square twice, while
correct piece to perform the blockade.
Black would only be able to control it
68...Kg6 69.Ra5! Re6 70.Ne1! once (if we disregard every other piece
on the board). Of course, that’s an
The knight finally reaches its oversimplification: rooks can also be
better than minor pieces. But to flex
their true strength, they need chaos, 8
t+vL*+*T
unprotected pieces and pawns, open files 7
OoO*VoOo
*+o+*+*+
and attacks. 6
That’s why the stunningly original
maneuver with 25.Nh1!! held such
5
+*+*Pm+*
power. Nakamura managed to stabilize 4 *+*+*+*+
the situation in his camp, there was 3 +*+*+n+p
pPp+*Pp+
nothing for the black rooks to attack, 2
and he was then able to slowly but
surely conquer every dark square on the
1 RnB*+rK*
board. By the way, such moves teach us a b c d e f g h
that the beauty or ugliness of a move –
much like the beauty or ugliness of
humans – should be judged by its
White to move
internal idea, rather than its external
appearance. 10.Nc3
I met GM David Navara in an online
lecture, where he said that he had read White is in no rush to give a rook
my books in his youth – am I really that check on d1, since Black’s king is not
old? In the next game, he convincingly well placed in the center on d8.
dispatches none other than my buddy Moreover, White’s rook often stands
Alexei Shirov, illustrating the power of better on either e1 or its current square,
his unopposed light-squared bishop. f1. The latter is especially relevant when
White plays for f2-f4-f5.
Game 34
Alexei Shirov – David Navara 10...Nh4!?
Malmo 2022
Principle: Swaps benefit the more
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.0-0 cramped side. However, one can raise a
Nxe4 5.d4 Nd6 6.Bxc6 dxc6 7.dxe5 pair of objections to the move:
Nf5 8.Qxd8+ Kxd8 9.h3 Be7 1) A swap of White’s f3-knight means
that the way is cleared for White to
activate his kingside pawn majority with
f2-f4.
2) Black’s decision was to make a
fifth(!) move with the same knight as
part of his first 10.
Number 2 on the list is a gross 8
t+*+l+*T
violation of the Principle: Don’t move 7
OoO*+oOo
the same piece again and again during
the opening phase. Navara is, of course,
6
*+o+*+*+
perfectly aware of this principle, but he
5
+*+*Pv+*
violates it deliberately, gambling that 4 *+*+*PpV
White lacks the firepower to endanger 3 +*N*+*+p
pPp+*+*+
the black king. 2
11.Nxh4 Bxh4 12.f4 Bf5 1 R*Br+*K*
a b c d e f g h
Developing, stopping f4-f5, and
attacking the c2-pawn. However, this
walks right into g2-g4.
Black to move
13.Rd1+!
“Here I spent a lot of time as I was not
Including this check will make it sure whether the bishop should retreat to
impossible for Black to take the pawn c8 or to d7. I had a vague idea that c8
on c2. was the right square, but did not
13.g4!? is interesting but risky, since remember why,” writes Navara.
now Black can actually take the pawn
with 13...Bxc2, and White is the one in 14...Bc8!
need to prove that there is adequate
Don’t worry, modern opening theory
compensation for the missing pawn.
confuses the hell out of me as well. Such
13...Ke8 14.g4! a move doesn’t exactly give off a vibe of
“Feel my righteous rage!”, yet the
counterintuitive retro-developing move
is Black’s most accurate option. Yes,
Black’s position is lacking in
development, but Black has the bishop
pair, and White runs the risk of
overextension.
The trouble with 14...Bd7 is that the
bishop may be vulnerable later to line-
opening tricks on e6: for example, 15.f5
h5 16.Bf4 g6 17.e6! when the engine 16...g6!
says Black is OK, but it looks seriously
unpleasant from a human perspective. Breaking the kingside bind and
14...Bxc2?? is a beginner’s error. The fighting for the light squares. The
bishop is trapped after 15.Rd2 Bg6 unopposed light-squared bishop on c8
16.f5. That’s why the check on d1 was hides tremendous power and should not
included. be allowed to be passive.

15.Kg2 17.Kf3!

The white king tries to support the Shirov finds his best fighting chance.
advanced pawns on the kingside.
17...gxf5
15...h5 16.f5!?
Including 17...hxg4+!? is also good.
This is principled, but it might be a bit
18.g5!
too much for the white position to
handle. Objectively it is fine, but the The point of White’s previous play.
structure cannot maintain its integrity. Does a dream count as a plan? GM
White can also try the slightly more Alexei Shirov’s religion is
solid 16.Kf3!?. complications, and his motto is: “There
is no investment without risk.” Alexei
8
t+v+l+*T normally likes to play chess as if he is
7
OoO*+oO* on his fourth glass of champagne. In this
case, he sacrifices a pawn to corral the
6
*+o+*+*+ h4-bishop.
5
+*+*Pp+o
4 *+*+*+pV 18...Rg8 19.Be3 Be6 20.Rg1 Rd8
3 +*N*+*+p
2 pPp+*+k+
1 R*Br+*+*
a b c d e f g h

Black to move
 maneuver is impossible to implement.
8
*+*Tl+t+ Black has access to hidden counterplay
7
OoO*+o+* against every white try.
6
*+o+v+*+ Conclusion: Black would be winning
if it wasn’t for the bishop on h4. With
5
+*+*PoPo the bishop trapped there, the position
4 *+*+*+*V becomes extremely concrete. The engine
3 +*N*Bk+p assesses the position as even.
2 pPp+*+*+ 21.Ne2?!
1 R*+*+*R* For the ambitious, there is no right or
a b c d e f g h
wrong. Shirov plays to trap Black’s
bishop with Kf4, Nd4, and Nf3, which
is not a bad idea. The problem is that his
White to move implementation is premature.
Correct is a more restrictive approach
David wrote that this position was still
like 21.b4, but it’s hard for a human to
part of his home preparation! When both
play that way, since it looks a bit
sides believe they stand better, one of
purposeless.
them must be wrong. How can we get a
handle on this unwieldy mess? Let’s try 21...Bc4! 22.Nd4
and work through the details:
1) Black is up a pawn.
2) Black owns the bishop pair in an
8
*+*Tl+t+
open position. 7
OoO*+o+*
3) Black’s unopposed light-squared 6
*+o+*+*+
+*+*PoPo
bishop is a powerful piece. 5
4) Black’s kingside structure is broken
with three isolated pawns – two of them
4 *+vN*+*V
doubled – and may later be vulnerable to 3 +*+*Bk+p
an attack from White’s pieces. 2 pPp+*+*+
5) Black’s bishop is stranded on h4,
and White can try to win it with the
1 R*+*+*R*
a b c d e f g h
maneuver Ne2, Kf4, Nd4, and Nf3.
However, White’s initial euphoria
begins to fade upon the realization this
 Black to move

22...Bxg5!!
8
*+*Tl+t+
7
OoO*+o+*
Blood is repaid in blood. It takes a 6
*+o+*+*+
prodigious level of calculation to see
that Black isn’t losing material, but
5
+*+*P*Vo
David’s shot turns out to be correct.
4 *+vN*O*P
3 +*+*+k+*
23.h4? 2 pPp+*B*+
White had a final chance to bail out 1 R*+*+*R*
with: 23.Rxg5! Rxg5 24.Nxc6! Bd5+ a b c d e f g h
25.Kf4 Bxc6 26.Kxg5 and Black can
press, but with a few more accurate
moves, White should hold.
Black to move
23...Bd5+ 24.Ke2 Bc4+ 25.Kf3
26...c5!
It looks like the game should end in a
Again going on the counterattack.
draw by threefold repetition.
27.hxg5
25...f4!
Shirov decides that enough is enough
But Black can play for more.
and accepts an unpleasant opposite-
26.Bf2 colored-bishop endgame.
The alternative, 27.Nb3, loses to:
Also difficult for White is 26.Rxg5 27...Bd5+ 28.Ke2 f3+ 29.Kf1 Bc4+
Rxg5 27.hxg5 fxe3 28.Kxe3 c5 when 30.Ke1 Bxb3!. White is busted, since
Black has the more active rook, the Black will unpin the g5-bishop with a
better minor piece and, at least check on d2.
temporarily, an extra pawn.
27...cxd4 28.Kxf4 Be2!
 This weakens White’s light squares
8
*+*Tl+t+ further, but it serves the important
7
OoO*+o+* purpose of preventing ...c7-c5.
6
*+*+*+*+ 30...Rg6!
5
+*+*P*Po
*+*O*K*+
4 Activating the g8-rook and getting
ready to exploit the numerous
3 +*+*+*+* weaknesses on the white queenside.
2 pPp+vB*+ 31.Re4 Rc6 32.Bxd4 Rxc2 33.a4 b6
1 R*+*+*R* 34.a5 Kd7! 35.Be3 Ke6 36.Rc1
a b c d e f g h
Black’s rook was too active and had to
be swapped off.
White to move 36...Rxc1 37.Bxc1 Rd3!
White is in deep trouble: Threatening checkmate on f3.
1) Black is up a pawn and has the
potential to create a second passed pawn 38.Re3
on the queenside.
2) Black’s bishop is about to move to
its optimal post on g4.
8
*+*+*+*+
3) Black’s king may later be able to
7
O*O*+o+*
walk up to e6, endangering White’s e- 6
*O*+l+*+
pawn. White’s g-pawn will also require 5
P*+*P*Po
*P*+*Kv+
protection. 4
4) White is chronically weak on the
light squares.
3 +*+tR*+*
5) The presence of opposite-colored
2 *+*+*+*+
bishops will enhance, rather than dull, 1 +*B*+*+*
White’s pain. Black can use his pawns a b c d e f g h
and pieces to control enough dark
squares, while White is unable to do the
opposite.
Black to move
29.Rae1 Bg4 30.b4!?
38...Rd4+
 David, understanding there are no Bxa7 tricks.
deep thinkers in time pressure, repeats a
few times to reach the time control on 45.Be3+ Kd5 0-1
the 40th move.
This must be the same feeling as when
39.Re4 Rd3 40.Re3 Rd4+ 41.Re4 Kd5! we die and begin walking through
heaven’s gates. White is unable to fend
With the time control reached, David off the two black passed pawns. To
is able to calculate the committal illustrate: 46.Bd2 c5 47.Be1 Bd7!
transition to entering the pure opposite- 48.Bf2 c4 49.Be1 Be6 50.Bd2 h4
colored-bishop endgame. 51.Bc3 h3 52.Kg3 Ke4. Up next is
...Kd3, and the c-pawn costs White his
42.Rxd4+ Kxd4 43.Bd2 bishop.
Practically, the game was won by

*+*+*+*+
8 David’s superb opening preparation and
excellent calculation. On the other hand,
7
O*O*+o+* positionally, the game was won due to
6
*O*+*+*+ Black’s dominant display on the light
5
P*+*P*Po squares. Yet again, it was the power of

*P*L*Kv+
4 the unopposed bishop that did the trick.
Let’s move on to a game that was won
3 +*+*+*+* due to a superiority on a color complex,
2 *+*B*+*+ but not because of an unopposed bishop.
1 +*+*+*+* The irony of this game was that Black,
a b c d e f g h in the opening, radically gave up his
good light-squared bishop for White’s
knight. Such a decision normally risks
weakness on the light squares. However,
Black to move Giri cleverly sacrificed a pawn to flip
the narrative and, in the end, it was
43...bxa5!
Praggnanandhaa who lost due to his
Black creates a fatal second passed weakness on the dark squares. Let’s see
pawn. what I mean.
Game 35
44.bxa5 a6! Rameshbabu Praggnanandhaa –
Anish Giri
Securing the a-pawn from Be3+ and Oslo (rapid) 2022
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6
5.Be2 0-0 6.Nf3 e5 7.0-0 Bg4 8
tM*W*T*L
7
+oO*+oVo
8
tM*W*Tl+
6
*+*O*Mo+
7
OoO*+oVo
5
O*+pO*+*
6
*+*O*Mo+
4 *+p+p+*+
5
+*+*O*+*
3 +*N*+b+p
4 *+pPp+v+ 2 pP*+*Pp+
3 +*N*+n+* 1 R*Bq+rK*
2 pP*+bPpP a b c d e f g h
1 R*Bq+rK*
a b c d e f g h
White to move

This is clearly a King’s Indian with an

White to move identity crisis. Giri’s idea is to play
...Ng8 and ...Bh6, and then swap off
Who among us hasn’t, at some point, White’s good dark-squared bishop, after
cheated on our main opening by which White is left with a bad remaining
committing symbolic adultery with a bishop. Of course, there are a few
sleazy other? We are all in search of that problems with this plan. For one, it
unanalyzed hybrid, whose venom is seems a bit slow. Moreover, White can
especially virulent. Welcome to the possibly prevent it if he sees it coming.
latest King’s Indian fad.
11.Bg5!
8.d5!
The idea is to provoke ...h7-h6 to
Pragg is booked up, even on this prevent the ...Ng8 and ...Bh6 plan.
sideline, and chooses the best white
approach. 11...h6 12.Be3 Nbd7 13.a3?!

8...a5 9.h3 Bxf3 10.Bxf3 Kh8!? Ignoring Black’s play on the kingside
isn’t the way to go.
The more circumspect 13.Qd2 can be
met with 13...Nh7!, which is similar to
the pawn sacrificed in the game.
However, White at least hasn’t remaining bishop. 14...h5?! 15.Bd1!
committed a tempo to the somewhat leaves White clearly better and would
useless a2-a3. fully justify Pragg’s approach.
A restrictive approach like 13.Qc2!
Nh7 14.Be2 f5 15.exf5! gxf5 16.f4! is 15.Bxh6
also both typical and strong. Especially
Everyone likes free stuff, but,
in hindsight, it looks like a massively
honestly, in hindsight of the game
better practical choice.
continuation, maybe this is not White’s
13...Nh7 14.Qd2 best path, since it leads to chronic
weakness on the dark squares and offers
Threatening the pawn on h6. But this Black a kingside attack. It would take
is inconsistent with White’s previous superhuman restraint to play 15.Qc2.
move. But White has lost some time, and
Black’s initiative on the kingside is at
8
t+*W*T*L least enough for equality after
something like 15...f4 16.Bd2 Ng5.
7
+oOm+oVm
6
*+*O*+oO 15...Qh4! 16.Bxg7+
5
O*+pO*+* 16.Be3?? would be a delightful gift for
4 *+p+p+*+ Black after 16...f4.
3 P*N*Bb+p
*P*Q*Pp+
2 16...Kxg7

1 R*+*+rK*
a b c d e f g h
8
t+*+*T*+
7
+oOm+*Lm
6
*+*O*+o+
Black to move 5
O*+pOo+*
14...f5!
4 *+p+p+*W
3 P*N*+b+p
Running is sometimes not an option,
since it is even more risky to turn our
2 *P*Q*Pp+
back upon an opponent, rather than face
1 R*+*+rK*
him head on. Giri is willing to give up a a b c d e f g h
pawn to leave his opponent with a bad
 White to move been 19...Nc5!?, threatening a fork on
b3. After 20.Qc2 a4 21.f3 Nf6 It’s not
The scenery is completely altered obvious how Black will make progress,
compared to the previous diagram. but the game will be completely one-
Black’s missing pawn isn’t felt at all, sided.
while the dark-squared dominance is
obvious, and the white king feels a bit 20.f3 Ng5
lonely on the kingside, where Black’s
army has started gathering.
8
t+*+*T*+
17.exf5 7
+oO*+*L*
This opens the g-file for Black’s 6
*+*O*M*+
rooks, which in turn will endanger the 5
O*+pO*M*
white king. 4 *+p+*ObW
P*N*+p+p
The alternative is to try 17.Bd1 f4 3
18.f3, blocking any attempts at a quick
...f4-f3. Then again, this weakens even
2 *P*Q*+p+
more kingside dark squares. The engine 1 R*+*+rK*
claims it’s playable, but it hurts a a b c d e f g h
human’s eyes just looking at it from the
white side.

17...gxf5 18.Bd1! f4! White to move

Preventing f2-f4. Giri correctly Black received more than full

declines the offer to take back a pawn. compensation for the sacrificed pawn:
After 18...Qxc4? 19.Rc1 the game’s 1) Black’s power on the kingside dark
focus shifts to the queenside, where squares must be terrifying for the white
White is the boss. king.
2) White’s bishop serves the merely
19.Bg4 Ndf6!?
defensive purpose of blocking the g-file.
This move indicates that Giri isn’t all Otherwise, it just shoots off into empty
that interested in strategic gains on the space.
queenside. Instead, he plays directly for 3) A rook is coming to h8 to further
mate. enhance Black’s attacking chances.
An equally sensible path would have 4) White’s control over the e4-hole is
merely cosmetic, since Black will
immediately seal the e4-square with a Both sides miss the idea 23...Nxe6!
pawn should White ever move his 24.dxe6 d5! when Black gets a huge
knight to e4. center and an almost winning position.

21.Qf2 Qh7! 22.Be6! 24.Bg4?

Praggnanandhaa recognizes that Panic. The Indian prodigy probably

...Nxg4 can’t be allowed. felt that returning the pawn by allowing
A careless move like 22.Rae1? would Black to take on e6 isn’t ideal, but
have allowed Black to crash through moving the bishop back to where it
with 22...Nxg4! 23.fxg4 (even worse is came from – essentially losing two
23.hxg4? Rh8) 23...f3!? when ...Rf4 and tempi in such a sharp position – is much
...Raf8 are coming, putting the white worse.
kingside under tremendous pressure. Correct was 24.cxd6! cxd6 25.Rfe1,
followed by bringing the other rook to
22...Rae8 23.c5?! the center and hoping that Black would
take on e6 and open up the game.
The engine prefers 23.Rfd1!.
However, this is much easier to say with
the engine running. Without it, it still
8
*+*+tT*+ looks quite scary, whether or not Black
7
+oO*+*Lw takes on e6.
6
*+*ObM*+ 24...Nxg4!
5
O*PpO*M*
4 *+*+*O*+ The structure shifts in Black’s favor.
3 P*N*+p+p 25.fxg4
2 *P*+*Qp+
R*+*+rK*
1 After 25.hxg4 Rf6! followed by ...Rh6
Black gets a decisive attack.
a b c d e f g h
25...Rf6?!

Far stronger was the thematic 25...e4!,

Black to move and if 26.h4 then 26...e3! 27.Qe1 f3!
Black to move when White collapses.

23...Kh8?! 26.cxd6 cxd6 27.h4!?

 This attempt to complicate matters Rxg4
backfires.
However, I don’t have much faith that
White would anyway survive after the
8
*+*+*+*L
correct 27.Rac1 Rg8!. 7
+*+w+*+*
6
*+*O*+*T
8
*+*+t+*L 5
Oo+pO*+p
7
+o+*+*+w 4 *+*+*OtQ
6
*+*O*T*+ 3 P*+*+*+*
5
O*+pO*M* 2 *P*+*Rp+
4 *+*+*OpP 1 +*R*+*K*
3 P*N*+*+* a b c d e f g h
2 *P*+*Qp+
1 R*+*+rK*
a b c d e f g h White to move

34.Rc7!?

Black to move Just as White’s king is threatened with

total extermination, the Indian prodigy
27...Rh6! finds a trick that temporarily flusters
Giri. “This came as a bit of a shock,”
This clever trick forces White’s
writes the Dutch supergrandmaster.
structure into frozen immobility.
Luckily for Giri, his resourceful
28.h5 opponent’s shot is ineffective.

Suddenly, White’s ugly kingside 34...Qf5!

structure assails our vision, since we can
The position’s secret is disclosed, as
see that g4 is inevitably going to fall.
Giri explained: “Fortunately all the
28...Qd7 29.Qh4 Rg8 30.Rac1 b5!? heavy pieces around my king protect it
perfectly and White’s position is just
The idea is to try and chase White’s falling apart.”
knight away from e4.
35.Qe7
31.Nd1 Ne4 32.Nf2 Nxf2 33.Rxf2
 This looks far more ominous than it higher risk of perpetual check.
actually is since Black’s king is
perfectly well protected by his major 37...f3!
pieces. In fact, White lacks even a single
Now it becomes clear that Black’s
check after Giri’s next move.
attack is more potent than White’s.
35...Rg8! 36.Qf7
38.Rc3

*+*+*+tL
8 38.Rxf3 loses to 38...Rxg2+!.
7
+*R*+q+* 38...e4
6
*+*O*+*T More attackers creep forward.
5
Oo+pOw+p
4 *+*+*O*+ 39.Re3
3 P*+*+*+*
2 *P*+*Rp+ 8
*+*+*+tL
1 +*+*+*K* 7
+*+*Q*+*
a b c d e f g h 6
*+*O*T*+
5
Oo+p+w+p
4 *+*+o+*+
Black to move 3 P*+*Ro+*
36...Rf6? 2 *P*+*Rp+
Giri misses a forced win: 36...Qb1+!
1 +*+*+*K*
a b c d e f g h
37.Kh2 Qd1! 38.Rf3 Qe2! and White
must hand over his queen to prevent
mate.
Black to move
37.Qe7?
39...Rxg2+!
Praggnanandhaa desperately needs
swaps to have any hope of surviving. And Black finally crashes through.
The correct move was 37.Qd7! Qxh5
38.Rc8!, which is not so easy at all for 40.Kf1
Black, whose king is suddenly at much
 40.Rxg2 allows 40...fxg2 41.Qe8+ Rf8 Further beatings don’t bother the
42.Qg6 Qf1+ 43.Kh2 Qh1+ 44.Kg3 already dead. White will be down two
g1=Q mate. pawns in the rook ending.

40...Qg5 41.Qe8+ Kg7 42.h6+ Kxh6 46.Rh2+ Kg4 47.Rxe4+ Rf4 48.Re6
43.Qh8+ Kg6 44.Qg8+ Kh5 Rxd5 49.Rg6+ Kf5 50.Rh5+ Ke4
51.Rhh6 Rd1+ 52.Kf2 Rd2+ 53.Kf1
Far from courting death, Black’s king Rxb2 54.Rxd6 Ke3
is actually quite safe. Moving the king to
f5 was also possible. There was no reason for White to be
playing on, except for Giri’s time
45.Qxg5+ pressure.
45.Qh7+ “was making me dizzy in the 55.Rhe6+ Re4 56.Rxe4+ Kxe4 57.Re6+
time trouble, but the win is not that hard Kd3 0-1
here, if we stop seeing ghosts,” writes
Giri. After 45...Rh6 46.Qf7+ Rg6 What happened here? After Black
47.Qh7+ Kg4 48.Rxg2+ fxg2+ 49.Kxg2 gave up his light-squared bishop on f3
Kf4+ Black wins. with 7...Bg4 and 9...Bxf3, White was
guaranteed better control of the light
45...Rxg5 squares around the whole board (the
structure also helped), while Black
gained better control of the dark squares.
8
*+*+*+*+ Which color-complex superiority is
7
+*+*+*+* more important?
6
*+*O*T*+ The answer greatly depends on which
5
Oo+p+*Tl side gets to use their superiority to

*+*+o+*+
4 infiltrate the opponent’s camp and
transform the color-complex superiority
3 P*+*Ro+* into active posts for their pieces. For
2 *P*+*R*+ White to do something like that on the
1 +*+*+k+* light squares, a pawn break needed to
take place either on c5 with c4-c5 or f4
a b c d e f g h
with f2-f4.
I urge you to go back and take a look
at the note on 13.Qc2! Nh7 14.Be2.
White to move After 14...f5 15.exf5! gxf5 16.f4! White
gets exactly what I’m talking about:
concrete play on the light squares. The Game 36
f5-pawn is tender, and the kingside light Anish Giri – Max Warmerdam
squares are weak. If Black keeps the Wijk aan Zee 2024
tension, then eventually taking on e5
would also open the f-file for the f1- 1.Nf3 d5 2.e3
rook to attack the f5-pawn. If Black
were to close the center with ...e5-e4, If the Nimzo-Indian and the Queen’s
then the hole on e5 becomes available to Indian are fully playable for Black, then
the knight via b5 and d4. Finally, if why not try them with the white pieces?
Black didn’t go for 14...f5, then there This way of playing has risen
would be no counterplay, and White tremendously in popularity in recent
could slowly prepare c4-c5. years, partly because of GM Axel
In contrast, in the game, after Giri’s Smith’s efforts in his book e3 Poison,
pawn sacrifice with 14...f5!, we ended published by Quality Chess in 2017.
up in a position where Black’s pieces Don’t let the relative quiet fool you – the
started storming the kingside via the position is destined to explode in
dark squares (all of 15...Qh4!, violence.
19...Ndf6!? and 20...Ng5 were quite
2...Nf6 3.b3 e6 4.Bb2 Bd6 5.c4 0-0
dangerous), while White’s superiority
6.Nc3 c5
on the light squares only allowed for a
defensive blockade with the bishop on
g4, serving no attacking purpose. Again, 8
tMvW*Tl+
Oo+*+oOo
the superiority on a color complex had 7
turned into a full-fledged dynamic
struggle, where who had the initiative
6
*+*VoM*+
was going to decide, and the initiative 5
+*Oo+*+*
most definitely stood with Black. 4 *+p+*+*+
The following game is another
instructive battle between White’s
3 +pN*Pn+*
power on the dark squares and Black’s
2 pB*P*PpP
power on the light squares. Honestly, 1 R*+qKb+r
while watching this game live, I a b c d e f g h
couldn’t tell which color was the more
potent one. At the highest levels, a
weakness on a single color can be a
White to move
subtle thing, which is exceedingly
difficult to discern and assess correctly.
 This is not the most popular move, but Meet the opponent’s wing attack with a
it is extremely logical. Warmerdam central counter. After 8.exd4 Black can
wants to fight for central control with either react normally with 8...cxd4 or in
everything at his disposal. even more attacking fashion with 8...e5!,
putting Giri’s gamble under serious
7.g4!? scrutiny.
7...Nxg4 is possible but risky, and
Judge not, that ye be not judged!
8.Rg1 has to be met with 8...d4! if Black
Sometimes we know what we do is
doesn’t want to come under a strong
potentially shady, yet it feels so right.
attack.
Giri goes full John Wick on his
opponent. It doesn’t take much for some 8.g5
people to boil over in fury, does it? The
quarrelsome “I don’t believe in half-
measures!” push of the g-pawn to the 8
t+vW*Tl+
unprotected fourth rank is high fashion
in all types of positions in modern chess.
7
Oo+*+oOo
Giri’s move initiates an immediate pawn
6
*+mVoM*+
storm against Black’s king, at three 5
+*Oo+*P*
potential costs: 4 *+p+*+*+
1) It is not a developing move. 3 +pN*Pn+*
2) White’s light squares are weakened,
especially on the kingside.
2 pB*P*P*P
3) Giri’s high-risk gamble sets up the
1 R*+qKb+r
potential for overextension should his a b c d e f g h
attack fail.
A more literal interpretation of
opening scripture would be to moderate Black to move
ambitions with something sounder, such
as 7.cxd5 exd5 8.d4. This will likely 8...Ne4
lead to a classical IQP position.
When the knight can answer g4-g5 by
7...Nc6 jumping actively forward, it should
usually do so. Otherwise, the g2-g4-g5
The most natural, preparing ...d5-d4. push could also be seen as a way to
The immediate 7...d4!? is a hotheaded destabilize Black’s center.
response that follows the Principle:
9.h4 Re8?!
 This is a bit too timid, not meeting the and ...d5-d4.
position’s demands. 2) It pre-emptively opens the d-file for
The engine prefers 9...f5!, with a total White’s rook about to appear on d1 after
mess, but one in which Black has good castling queenside.
and active counterplay in the center. 3) If need be, White has a second c-
pawn to challenge Black’s control of the
10.Qc2!
d5-pawn.
Giri threatens the knight on e4 and According to the engine, White’s
prepares to castle queenside, since his advantage is already close to decisive.
attack is already on the move on the
11...g6?!
kingside.
This creates weaknesses on the dark
10...Nxc3
squares around the black king.
10...Nb4 doesn’t bother White after Black’s relatively best move is
11.Qb1. 11...Ne5, but after 12.Nxe5 Bxe5 13.f4
Bd6 14.0-0-0 the position is anyway
8
t+vWt+l+ terrible.
7
Oo+*+oOo 12.h5?!
6
*+mVo+*+ This is rushed. Better was to develop
5
+*Oo+*P* with 12.0-0-0!, getting the king out of
4 *+p+*+*P the center and putting pressure on d5.
3 +pM*Pn+* After 12...Ne5 13.Nh2! followed by f2-
2 pBqP*P*+ f4, Nh2-g4 and h4-h5, White’s attack is
unstoppable.
1 R*+*Kb+r
a b c d e f g h 12...Ne5!

White to move

11.dxc3!

I like this recapture for three reasons:

1) It bolsters White’s control over key
central squares, preventing both ...Nb4
 Of course, 14...Qxg5?? hangs a piece
8
t+vWt+l+ to 15.f4.
7
Oo+*+o+o 15.f4
6
*+*Vo+o+
5
+*OoM*Pp
4 *+p+*+*+
8
t+vWt+l+
3 +pP*Pn+*
7
Oo+*+oVo
2 pBq+*P*+
6
*+*+o+o+
1 R*+*Kb+r
5
+*Oo+*Pp
a b c d e f g h
4 *+p+*P*+
3 +pP*P*+*
2 pBq+*+*+
White to move
1 +*Kr+b+r
a b c d e f g h
Black forces a trade while
simultaneously managing to switch his
dark-squared bishop to the a1-h8
diagonal, offering his king more Black to move
protection.
15...e5!
13.Nxe5
GM Warmerdam is conscientious of
In this case, 13.Nh2?! can be met with his defensive obligations. This strong
13...Qxg5!, since 14.f4? allows pawn sacrifice follows the Principle:
14...Qh4+. Meet the opponent’s wing attack with a
central counter.
13...Bxe5
16.hxg6
Now Warmerdam’s decision to start
with 11...g6?! seems brilliant rather than 16.Rxd5 Qe7 17.h6! is an interesting
weakening, as the dark-squared bishop approach, accepting that there is no
can park itself on g7 instead of being checkmate down the h-file, and instead
kicked with f2-f4. using the pawn to gradually restrict the
black king. The h6-pawn won’t play a
14.0-0-0 Bg7 big part for some time, until an endgame
is reached, where Black’s weakness on
the back rank could become crucial.
8
t+vWt+l+
Oo+*+oV*
16...hxg6 17.Bd3!? 7
Giri adopts an all-or-nothing, mate-or- 6
*+*+*+o+
bust stance. 5
+*OoO*P*
An interesting alternative would be:
17.Rxd5 Qe7.
4 *+p+*P*+
3 +pPbP*+*
8
t+v+t+l+
2 pBq+*+*+
7
Oo+*WoV* 1 +*Kr+*+r
6
*+*+*+o+ a b c d e f g h
5
+*OrO*P*
4 *+p+*P*+
+pP*P*+*
3 Black to move

2 pBq+*+*+ 17...exf4 18.exf4 Qc7?!

1 +*K*+b+r Developing with tempo, with
a b c d e f g h 18...Bg4! is more accurate. 19.Rdg1? is
nicely met with 19...dxc4!, exploiting
the rook’s absence from d1. The obvious
White to move 20.Bxg6 loses to 20...Re2.

With 18.f5! (not allowing the e5-pawn 19.Rh4 dxc4 20.Bxg6!

to disappear, activating the g7-bishop)
After 20.Bxc4? Bf5 Black seizes the
18...Bxf5 19.e4 Be6 20.Qd2! White will
initiative and stands clearly better.
get more than enough compensation for
the exchange if Black takes on d5. The
dead g7-bishop plays a big part in this
assessment.
 Now we see the reality: White has
8
t+v+t+l+ seized control of f5, and his two fluid
7
OoW*+oV* kingside pawns pose a serious threat to

*+*+*+b+
6 Black’s king.

5
+*O*+*P* 22...Qf7 23.Qd6+
4 *+o+*P*R
3 +pP*+*+* 8
t+v+tL*+
2 pBq+*+*+ 7
Oo+*+wV*
1 +*Kr+*+* 6
*+*Q*+*+
a b c d e f g h 5
+*O*+pP*
4 *+o+*+*R
3 +pP*+*+*
pB*+*+*+
Black to move 2
Sure, Giri’s move punches a hole in 1 +*Kr+*+*
the defense. However, keep in mind that a b c d e f g h
the sacrifice of a piece is a big
commitment. Which is worth more:
White’s attack or Black’s extra piece?
Black to move
20...fxg6 21.Qxg6
23...Kg8!?
White’s threat is Rdh1 followed by
Rh8 mate. The engine suggests that 23...Qe7
holds in a cleaner fashion, but White has
21...Kf8!? a few tempting options here. Even after
the most straightforward 24.Qxe7+
21...Qc6!? and 21...Qe7!? also hold. Kxe7 25.f6+ Bxf6 26.gxf6+ Kxf6 the
At first sight, 21...Qf7?? looks strong engine says ‘0.00’, but to our human
too, since it seizes control over the key eyes, the wide-open black king
square on f5. In reality, it walks into a combined with the presence of opposite-
cheapo and loses instantly to 22.Rh8+! colored bishops is a terrifying reality.
Kxh8 23.Qxf7.
24.f6 Qg6
22.f5
24...Bf5 25.fxg7 Qe6 is also OK for
Black. The resource of ...Qe3+ probably squares.
make it a bit simpler to handle compared 25...Bf8?? is way too greedy and loses
to the game. instantly to 26.f7+! Qxf7 27.Rh8+!.
Warmerdam could have maintained
25.Qf4
the precarious balance with 25...Bf5!
Heads or tails? The position has turned 26.Qxc4+ Re6! 27.fxg7 Rae8!,
into a battle between White’s attack and stabilizing the kingside. Up next is
Black’s extra material. Of course, White probably ...Qxg5+.
can just take Black’s bishop, equalizing
26.Qg3! Be6?
the material balance on the board.
Hence, the true fight is between White’s For the vast majority of players,
dark-square power and Black’s power defense is more difficult than attack.
on the light squares. Both kings are in Black puts up much greater resistance
danger, and the game is in a precarious with 26...Re5!, forcing the queens off
dynamic balance. Nevertheless, Black’s the board after 27.fxg7 Qxg5+.
task is clearly the tougher of the two.
8
t+*+t+l+
8
t+v+t+l+ 7
Oo+*+*V*
7
Oo+*+*V* 6
*+*+vP*+
6
*+*+*Pw+ 5
+*O*+wP*
5
+*O*+*P* 4 *+o+*+*R
4 *+o+*Q*R 3 +pP*+*Q*
3 +pP*+*+* 2 pB*+*+*+
2 pB*+*+*+ 1 +*Kr+*+*
1 +*Kr+*+* a b c d e f g h
a b c d e f g h

White to move
Black to move
27.fxg7
25...Qf5?
White wins back the piece, the black
After this, Black begins to lose his king is terribly exposed, and Black is
grip over his main asset, the light way too slow to create any threats
against the white king. The black queen is pushed away from
her ideal post on f5. At the same time,
27...Bf7 the white rook vacates the h-file, so that
a stronger piece – the queen – can use it.
Hoping for either ...Re2 or ...Bg6, but
Black is one tempo too slow. 28...Qe6 29.Qh2!
27...cxb3? unblocks the c3-pawn,
allowing 28.Rh8+, since taking on g7 Threatening Qh8 mate, which forces
would be met with the devastating c3- Black to take the g7-pawn and fatally
c4+, finally including the white dark- exposes Warmerdam’s king.
squared bishop in the attack. 28...Kf7
doesn’t help either. White can win in 29...Kxg7
multiple ways, the quickest being:
8
t+*+t+*+
8
t+*+t+*R 7
Oo+*+vL*
7
Oo+*+lP* 6
*+*+w+*+
6
*+*+v+*+ 5
+*O*+*P*
5
+*O*+wP* 4 *+o+*R*+
4 *+*+*+*+ 3 +pP*+*+*
3 +oP*+*Q* 2 pB*+*+*Q
2 pB*+*+*+ 1 +*Kr+*+*
1 +*Kr+*+* a b c d e f g h
a b c d e f g h

White to move
White to move
30.Rxf7+!
After 29.g6+! Qxg6 30.Qf3+! Qf5
31.g8=Q+! Rxg8 32.Qxb7+ on the next Black’s king is lacking defenders, and
move, either the a8-rook falls with the few that remain are swiftly
check, or a new white piece is eliminated.
introduced to the attack with check. The
defense has completely collapsed. 30...Qxf7

28.Rf4! 30...Kxf7?? makes it easy for White:

31.Qh7+ Kf8 32.Rf1+ mates next move. danger of becoming repetitive, when did
White’s attack actually become
31.g6! unstoppable? After 30.Rxf7+! and
31.g6!, controlling some light squares.
White’s g-pawn was the last remaining
Where was the final punch? It took
defender of the black king. It can’t be
place on d7, another light square.
captured, since opening the g-file would
Principle: When we have an undisputed
be fatal. However, if left alive, it
color-complex superiority, managing to
becomes a useful attacking unit.
seize control of the opposite complex
31...Qe6 32.Rd7+! 1-0 often has immediate decisive
consequences.
If the rook is captured, Black loses his Let’s finish our study of weak color
queen to Qh7+. There is no surviving complexes by looking at a few studies.
this, and Black understandably laid Game 37
down his arms. Bernhard Horwitz
This was another example where the Chess Monthly, 1885
two opposing color-complex
superiorities clashed with each other and
created a razor-sharp situation, in which
the first attack that hit home would
8
*L*+*V*W
probably be decisive. Despite the
7
Op+*O*O*
engine’s assessment of Warmerdam’s
6
p+k+p+pO
position as salvageable, it was extremely 5
+*+*+*+p
unpleasant to handle with Black for a
human.
4 *+b+*+*+
Special notice should be given to the
3 +*+*+*+*
moment when Giri went 22.f5. It was at
2 *+*+*+*+
that moment that White – generally 1 +*+*+*+*
playing with superiority on the dark a b c d e f g h
squares – managed to control a number
of crucial light squares. When such
things happen, it might look and feel
like (and it certainly did in this case) that White to move
White dominates both colors. That’s White to play and win
when attacks usually become virtually We should try to collect some data:
unstoppable.
1) Weirdly enough, the side that is a
Insisting on the same point despite the
queen up is the one who attempts to set
up an impenetrable zone, hoping to draw 8
*L*+*V*W
by erecting a fortress. 7
Op+*ObO*
2) Black’s pieces are hungry, out of
ammunition and shivering, since they
6
p+*+p+pO
are trapped in a muddy trench. The
5
+*+k+*+p
bishop can’t even move, and the queen’s 4 *+*+*+*+
choices are between g8, h8, and – if she 3 +*+*+*+*
*+*+*+*+
is feeling desperate – f7 and h7. 2
3) White completely dominates the
light squares. The only reason Black has
1 +*+*+*+*
a chance to hold the draw is the touch of a b c d e f g h
power on the dark squares.
Exercise: Come up with a plan for
White to overcome Black’s fortress White to move
attempt:
Step 1 has been completed. Now for
Step 2, we move our king to c6, running
Answer: We can improve White’s
Black’s king out of moves.
position by shifting our bishop to f7.
When this happens, the queen is out of 5.Kc6! Qg8!
moves, and Black is reduced to moving
his king. Oh dang! I forgot about Black’s vile
stalemate trick when I first attempted to
1.Kd5! solve this study. Black’s queen is
untouchable, at least for now.
Intending Bb5, Be8 and Bf7.
6.Kd7!
1...Qg8 2.Bb5! Kc7 3.Be8! Qh8 4.Bf7!
It’s OK that we didn’t get to capture
White has a chokehold on the position,
the queen – she’s quite useless anyway.
despite Black’s extra queen.
It is much more important that our king
4...Kb8 gets to infiltrate the opposing camp.
6.Bxg8?? is an indigestible option.
4...Qg8 loses easily to: 5.Bxg8 Kb8
6.Kc5 Kc7 7.b8=Q+! Kxb8 8.Kc6 Kc8 6...Qh8
9.Bh7! Kb8 10.Kd7 Ka8 11.Ke8.
6...Qxf7 is also no good. Any sensible
reaction wins for White: for example, Aha! We see the mating pattern. All
7.gxf7 g5 8.hxg6 h5 9.Ke8 Bg7 White needs is to transfer the bishop to
10.f8=Q Bxf8 11.Kxf8 and White’s g- c6.
pawn is way too fast.
10...Qg8!
7.Ke8! Kc7

8
l+k+*Vw+
8
*+*+kV*W 7
O*+*ObO*
7
OpL*ObO* 6
p+*+p+pO
6
p+*+p+pO 5
+*+*+*+p
5
+*+*+*+p 4 *+*+*+*+
4 *+*+*+*+ 3 +*+*+*+*
3 +*+*+*+* 2 *+*+*+*+
2 *+*+*+*+ 1 +*+*+*+*
1 +*+*+*+* a b c d e f g h
a b c d e f g h

White to move
White to move
Black’s queen is up to no good. We
Oh no, now we are the ones in can’t take her, or Black is stalemated.
zugzwang! No worries though; we have
a way out: 11.Kc7! Qh8

8.b8=Q+! 11...Qxf7 is still too slow for Black:

12.gxf7 g5 13.hxg6 h5 14.Kd8 h4
Step 3: By removing the b7-pawn 15.Ke8 h3 16.Kxf8 h2 17.Kxe7. If
from the board, White grants the black Black promotes, then we mate by
king access to the a8-square. The promoting to a queen or, better yet, a
unfortunate thing for the black monarch rook.
is that a8 is a light square, and the light
squares are White’s kingdom. 12.Be8!

8...Kxb8 9.Kd8 Ka8 10.Kc8

 8
l+*+bV*W 8
*+*+*W*+
7
O*K*O*O* 7
+rO*+*+*
6
p+*+p+pO 6
rV*+*+*+
5
+*+*+*+p 5
+*+*+*+*
4 *+*+*+*+ 4 o+*+*+*+
3 +*+*+*+* 3 L*+*+*+*
2 *+*+*+*+ 2 p+pB*+*+
1 +*+*+*+* 1 K*+*+*+*
a b c d e f g h a b c d e f g h

Black to move White to move

White to play and win
Black’s king surveys his life in a
melancholic reflection. There is no The late IM Mark Dvoretsky will
separation between church and state, and probably be forever regarded as one of
there is no stopping mate on c6. the greatest chess coaches of all time. It
is well-known that one of his favorite
12...Qg8 13.Bc6 mate coaching methods was to have his
students solve endgame studies. In a
This is an extreme example of how
book he co-authored with the legendary
total domination on a single color – in composer Oleg Pervakov called Studies
this case, coupled with a massive space
for Practical Players, Dvoretsky
advantage – can be more potent than explicitly stated his liking for the studies
even the opponent’s extra queen. of Alois Wotawa.
According to Dvoretsky, more than
Game 38 any other benefit, Wotawa’s studies
Alois Wotawa make us aware of deeply hidden mating
Wiener Schachzeitung, 1935 patterns. In my personal opinion,
Wotawa is the greatest endgame study
composer of all time. Here we take a
crack at one of his short yet brutally
difficult studies. Let’s gather some data
to orient ourselves first:
 1) Both white rooks are currently shut dysfunctional one, so we use that one to
out of the game. It’s clear the only way sacrifice the exchange, leaving us with a
to bring one of them into play is to useful rook on b7 that is mobile on the
sacrifice the exchange on b6. After the seventh rank.
exchange sacrifice on b6, White seizes
control over b4, which in turn means 1.Raxb6!
that Black must watch out for mating
1.c3?? threatens Bc1 mate, but
ideas involving Bc1.
completely neglects Black’s threat:
2) But then the question arises: which
1...Qf1+ 2.Bc1+ Qxc1 mate.
rook should we sacrifice? White’s sole
After 1.Rbxb6? cxb6 nothing really
focus is clearly going to be Black’s
makes sense as a winning attempt other
precarious king. Is our king in danger?
than 2.Rxb6. Otherwise, either ...Qf1+
Well, yes and no.
gives checkmate or the black king leaves
3) Despite our dreams of forming a
his cage on a3. But here, as mentioned
mating net around the black king, we
in the introductory text above, Black has
have a big problem. If we allow Black a
2...Qf1+ 3.Rb1 Qxb1+! 4.Kxb1 with a
free move, we get mated on our back
not-so-surprising stalemate.
rank with ...Qf1+. However, the
exchange sacrifice on b6 could possibly 1...cxb6
fix that, since our rook gains access to
b1, blocking the check on f1 and saving
our king from checkmate. 8
*+*+*W*+
4) So, what about taking twice on b6? 7
+r+*+*+*
Does that win on the spot? The plot
twists never end – it doesn’t! We
6
*O*+*+*+
suddenly notice that taking twice on b6
5
+*+*+*+*
leaves the black king with no moves, 4 o+*+*+*+
hinting at possible stalemate tricks. And 3 L*+*+*+*
the stalemate is not hard to spot: Black
can just give a check on f1 and take the
2 p+pB*+*+
rook that blocks the check on b1.
1 K*+*+*+*
a b c d e f g h
Exercise: How can White weave a
mating net around the black king,
but not allow a stalemating trick?
White to move
Answer: Step 1: Our a6-rook is the
Warning: Step 2 is going to be tough
to find. What now? We can’t take the
b6-pawn. We need to find an incredible 8
*+*+*+*+
resource. Do you see White’s shot, 7
+*+*W*+*
*O*+*+*+
which subtly uses the superiority felt on 6
the dark squares?
5
+*+*+*+*
2.Re7!! 4 o+*+*+*+
Woah! Physicists and philosophers tell
3 L*P*+*+*
us that something can never arise from 2 p+*B*+*+
nothing, yet it certainly feels like it can,
when we work on a Wotawa study! This
1 K*+*+*+*
a b c d e f g h
is the value to trying to solve such
studies. Our internal database becomes
more and more accustomed to shocking
anomalies, like White’s surreal last Black to move
move. With a single shot, White uses
pretty much every available dark square. Step 3: Suddenly, Black’s position is a
The rook steps on e7, threatens city of the dead, whose former
checkmate on e3, and becomes able to inhabitants perished from war, plague
block ...Qf1+ by retreating to e1. At the and, finally, famine. We take control
same time, the b6-pawn is left alive, and over b4 and weave a deadly mating net
that means there are no more stalemate with Bc1 mate coming up. Since the
tricks! black queen was driven to the e-file,
there is no threat to our back rank – the
2...Qxe7 e1-square is covered by our bishop. That
was the main point behind 2.Re7!!. I
Black has no choice but to accept the told you: Wotawa studies contain the
offer. power to bend and reshape our
2...Qf1+ is convincingly met with bewildered minds.
3.Re1 when Black will inevitably lose
the queen to Re3+. Trying to block the 3...Qe2
third rank by going 3...Qf6+ simply
invites 4.c3, when Bc1 mate becomes 3...Qe1+ stalls rather than prevents
virtually unstoppable. mate: 4.Bxe1 b5 5.Bd2 b4 6.Bc1 mate.

3.c3! 4.Bc1+ Qb2+

That’s the spirit. Fight on, brave

warrior. There will be no “happily ever
after” for Black’s queen, who is buying 8
*+*+*+*+
on credit and is reduced to a spite-block. 7
+*B*+oPv
5.Bxb2 mate
6
*+*+*+*L
5
+*+*K*+*
*+*O*+p+
You can perceive endgame studies as 4
merely tactical operations, where pure
calculation is the only requirement to
3 +*+*+*+*
solve them. I have to admit that people 2 *+*+*+*+
who think like that are at least partially
right. I’m sure there are many skilled
1 +*+*+*+*
a b c d e f g h
calculators out there who could spot the
winning sequence without stopping to
think about the positional concepts
[Link], in my mind, the White to move
people I’m talking about are already White to play and win
quite strong players and might
intuitively feel the positional factors at If Wotawa’s gnarled studies are
play (the reduced mobility of the black Prokofiev, then Smyslov’s are Mozart,
king and White’s superiority on the dark since the answer flows elegantly. But
squares), without having to consciously don’t even for a second get the mistaken
think about them. impression that Smyslov’s studies are
easy to crack. They certainly aren’t!
To me, especially as a coach, this
study is an enthralling story where one Let’s assess:
side exploits the weakness of a single 1) For the material count, the position
color – in this case, the dark squares – to is sparse, with only two pawns each and
weave a mating net around the enemy a pair of opposite-colored bishops. In
king. Yes, calculation is what eventually fact, we may be wondering why White
solves the problem, but the positional would bother playing on, since the
factors are what guide that calculation position appears to be a dead draw.
all the way to the final goal. 2) When we look deeper, we see that
White may be able to benefit from the
Game 39 presence of the bishops of opposing
Vassily Smyslov colors, since Black’s king is cornered
Shakhmaty v SSSR, 1987 and somewhat vulnerable due to the
weak dark squares around him.
 Exercise: How can White possibly g7-pawn from the board first, so that this
win from this barren landscape? stalemate is prevented by granting
access to h8 to the black king. Did
Answer: By sending Black’s bishop anyone notice the similarities with
and king into a panic-induced retreat. 8.b8=Q+! in the Horwitz study we saw
on page 283? The black king needs to be
1.g8=N+!! lured to a dark square – our kingdom –
in order to be checkmated.
1.g8=Q? at first appears to transpose.
However, Black need not take the queen 1...Bxg8
immediately: 1...f6+!. This zwischenzug
offers the black pieces much needed air. The only challenging response.
After 2.Kxf6 Bxg8 3.Bf4+ Kh7 4.g5 Alternatively, 1...Kg7 2.Nf6 d3 3.Ba5 is
Bb3 the game is an easy draw. Black a technical win for White. The winning
shouldn’t be allowed to get rid of the f7- process is lengthy yet certain, since
pawn. Black’s d-pawn falls, while White hangs
on to the g-pawn. 1...Kg5 loses in
Logical is 1.Kf6?, threatening Bf4
identical fashion to 1...Kg7 after 2.Nf6.
mate, but it doesn’t work: 1...Bg8
2.Bf4+ Kh7 3.Bd2 d3 4.Bf4. 2.Kf6!

8
*+*+*+v+ Stopping the f7-pawn in its tracks.
7
+*+*+oPl 2...d3
6
*+*+*K*+
5
+*+*+*+* Black’s pain is that, suddenly,
2...Bh7?? isn’t possible, since then
4 *+*+*Bp+ White would deliver mate 3.Bf4 mate.
3 +*+o+*+* 2...Kh7 transposes to the main line
2 *+*+*+*+ after 3.g5 d3 4.Bf4!.
1 +*+*+*+*
a b c d e f g h

Black to move

The problem is that 4...d2! 5.Bxd2 is

stalemate. White needs to remove the
 The lesson I learned from the movie
8
*+*+*+v+ Beneath the Planet of the Apes is: if you
7
+*B*+o+* see a dark, forbidding tunnel, you would

*+*+*K*L
6 be wise to avoid entry.
Giving up the d-pawn immediately
5
+*+*+*+* loses in easier fashion: 4...d2 5.Bxd2
4 *+*+*+p+ Kh8 6.Bc3.
3 +*+o+*+*
2 *+*+*+*+
8
*+*+*+vL
1 +*+*+*+*
7
+*+*+o+*
a b c d e f g h
6
*+*+*K*+
5
+*+*+*P*
4 *+*+*+*+
White to move 3 +*B*+*+*
3.Bf4+!!
2 *+*+*+*+
1 +*+*+*+*
It’s vitally important to lose a tempo a b c d e f g h
on the way to c3. We’re nearing a
position of mutual zugzwang, and White
needs to be careful to reach it with Black
to move. 3.Ba5? bungles the win: Black to move
3...Kh7 4.g5 Kh8 threatens ...Bh7 with
The main point is that 6...Kh7 (6...Bh7
an immediate draw. But 5.Bc3 Kh7 is
7.Kxf7 mate is the eternal problem)
drawn anyway. White wants to reach
7.Bd4! is zugzwang. The black king is
this exact position, but with Black to
forced to jump into his own grave on h8.
move.
After 7...Kh8 8.g6 White mates next
3...Kh7 4.g5! move.

It’s so easy to blow it: 4.Bd2? Kh8

5.Bc3 d2! 6.Bxd2 Bh7! 7.Kxf7 Bd3 and
Black’s king will be eternally safe on
h7.

4...Kh8
 8
*+*+*+vL 8
*+*+*+v+
7
+*+*+o+* 7
+*+*+o+l
6
*+*+*K*+ 6
*+*+*K*+
5
+*+*+*P* 5
+*+*+*P*
4 *+*+*B*+ 4 *+*+*+*+
3 +*+o+*+* 3 +*Bo+*+*
2 *+*+*+*+ 2 *+*+*+*+
1 +*+*+*+* 1 +*+*+*+*
a b c d e f g h a b c d e f g h

White to move Black to move

5.Be5! 6...d2

The only move that stops ...Bh7. This is why White needed to have the
bishop on c3, ready to take this pawn
5...Kh7 before it becomes a queen.
6...Kh8 loses, as usual, to 7.g6!.
But what about the d-pawn? 5...d2
Black’s terrified king and bishop are
6.g6! d1=Q queens the pawn, but it falls
rendered into characters in a H. P.
a tempo short: 7.Kg5+ f6+ 8.Bxf6 mate.
Lovecraft horror story. We should once
6.Bc3! again note White’s triumph on the dark
squares around the black king.
Finally, the mutual zugzwang position
mentioned in the note with 3.Ba5? is 7.Bxd2 Kh8 8.Bc3
reached, but it is Black to move. That’s
But we saw this on the note to 4...d2.
why the correct route for the bishop was
Without a black d-pawn to distract us,
Bc7-f4-e5-c3, rather than Bc7-a5-c3.
the mating process is quite simple:

8...Kh7 9.Bb2!

Zugzwang. We force Black’s king to

h8.
9...Kh8 10.g6! Warmerdam, another way to deal with
that issue is managing to control the
Black is mated on the next move. If second complex as well, which
only our lives could be as clear, immediately translates to complete
graceful, and effortless as Smyslov’s domination. In fact, it’s not unusual for
endgame studies! In this one, Smyslov that to be the final step in the process of
demonstrates an attack on one color, converting an advantage on a color
with astonishingly few attackers to complex.
conduct it.

Conclusion
In this chapter, we closely examined the
effect of a weakness on a color complex
in a position, in multiple different
circumstances. Often, this weakness is
either decided or exacerbated by the
presence of opposite-colored bishops,
but their presence isn’t necessary.
We saw several different assortments
of pieces. We went from a good knight
against a bad bishop in Anand-Lputian,
to a good bishop versus useless knights
in Vachier-Lagrave versus
Nepomniachtchi, to two pieces against a
rook in Nakamura-Caruana, and all the
way to weird and random material
imbalances in the final studies. Despite
their differences, all these positions were
decided by a weakness on a color
complex.
As we have said plenty of times
already, having superiority on a color
complex usually means being weaker on
the opposite complex. So, being quick
and aggressive in your attempt to exploit
that superiority is crucial. As we also
mentioned, most notably in Giri-
 Epilogue correct technique can be. It doesn’t
necessarily need to be our favorite thing
in the world, but we should know and
Let’s review a few points we
respect its value. In the end, what we all
examined throughout the book. We
want the most out of a chess game is a
learned:
full point, isn’t it?

1) How to exploit our opponent’s

weaknesses.
2) How to deal with a weakness or
weaknesses when they do arise in our
position.
3) How to avoid creating weaknesses.
4) How to inflict a weakness upon our
opponent’s position.
5) To correctly identify a weakness.
6) To develop an eye for a potential
hole and/or a weak color complex before
it actually arises. Once the hole is
created, it’s often too late for the
defender.

When there is no antagonist, there is

no need for a hero. That’s why we
looked at a lot of technical games, where
one side’s advantage is just barely there,
yet the player continues to induce errors
from the skillful application of steady
pressure. To many inexperienced
players, this process might seem a bit
boring or ineffective, but it is a crucial
part of the arsenal of every strong
player.
In that regard, my hope is that apart
from teaching you about weaknesses,
this book would serve the purpose of
showing how pretty and satisfying
 Index of Games and Studies Kasparov.G-Polgar.J, Wijk aan Zee
2000
Khalifman.A-Barua.D, Las Vegas
Games 1999
Adams.M-Meskovs.N, Germany 2022
Lakdawala.C-Akobian.V, Los Angeles
Anand.V-Lputian.S, Wijk aan Zee 2001
2000
Lakdawala.C-Bisgaard.P, Internet
Aronian.L-Nakamura.H, Wijk aan Zee (rapid) 2021
2014
Lakdawala.C-Griffith.K, San Diego
Carlsen.M-Giri.A, Internet (rapid) (rapid) 2013
2020
Lakdawala.C-Xia.B, San Diego (rapid)
Carlsen.M-Giri.A, Wijk aan Zee 2022 2016
Carlsen.M-Kramnik.V, Paris (rapid) Larsen.B-Kasparov.G, Niksic 1983
2016
Nakamura.H-Carlsen.M, Shamkir
Carlsen.M-Nakamura.H, Internet 2014
(rapid) 2021
Nakamura.H-Caruana.F, Madrid 2022
Caruana.F-Adams.M, London 2017
Praggnanandhaa.R-Giri.A, Oslo
Caruana.F-Carlsen.M, Tromsø (rapid) 2022
(Olympiad) 2014
Shirov.A-Navara.D, Malmo 2022
Caruana.F-Radjabov.T, Madrid 2022
So.W-Carlsen.M, Toronto (rapid) 2023
Gelfand.B-Malakhov.V, Sochi 2005
Vachier-Lagrave.M -
Giri.A-Tari.A, Stavanger 2022 Nepomniachtchi.I, Yekaterinburg 2020
Giri.A-Warmerdam.M, Wijk aan Zee
2024
Studies
Ipatov.A-Cuartas.J, Barbera del Valles
Blathy, The Chess Amateur, 1928
2008
Fontana, Schweizerische
Ivanchuk.V-Karpov.A, Tilburg (rapid) Schachzeitung, 1945
1993
Horwitz, Chess Monthly, 1885
Jakovenko.D-Fressinet.L, Plovdiv
Rusinek, The Problemist, 1993
2012
Smyslov, Shakhmaty v SSSR, 1987
Ju Wenjun-Firouzja.A, Wijk aan Zee
2024 Wotawa, Wiener Schachzeitung, 1935
Karjakin.S – Vachier-Lagrave.M, Zakhodyakin, 64, 1931
London 2017
Kasparov.G-Bacrot.E, Sarajevo 2000