A Professional Journal Published by The International Federation of Technical Analysts

IFTA Journal 08
Inside this issue

Trend Behavior, Market Depth and Time Structure of

Single Stocks within Stock Indices........................................................... page 04
Rolf Wetzer
Wyckoff Laws: A Market Test (Part B) – What has actually
happened?..................................................................................................... page 13
Henry Pruden and Benard Belletante
Dynamic MACD: Standard Deviation Embedded in the
MACD Indicator for Accurate Adjustment to Financial Market
Dynamics.......................................................................................................page 16
Gino Gandolfi, Monica Rossolini, Antonella Sabatini,
Stefano Caselli
An Exploration of the Nature of Bull Market Tops...................................page 22
Paul Desmond
Sentix: Behavioral Indices: A Behaviourally Oriented
Development of the TA Tool-Kit.................................................................page 26
Manfred Hübner

”...instead, what we can do

is create a mathematical
model that can mimic the
real thing — can mimic how
much a price varies, how
quickly it rises or falls...
And from it, you can
develop a powerful new tool
to study and work in
the market.”
Benoit B. Mandelbrot
 IFTA Journal 2008 edition

IFTA Journal
Letter from the Editor..........................................................................................................page 2

In Memoriam: John Brooks.................................................................................................page 3

by Elaine Long Knuth
Editorial
Trend Behavior, Market Depth and Time Structure of Single Stocks
Regina Meani (ATAA, STA, APTA) within Stock Indices.............................................................................................................page 4
Editor, and Chair of the Editorial Committee
[email protected]
by Rolf Wetzer

Roberto Vargas (STA) Wyckoff Laws: A Market Test (Part B) – What has actually happened?...................... page 13
Editor by Henry Pruden and Benard Belletante
[email protected]

Dr. Rolf Wetzer (VTAD)

Dynamic MACD: Standard Deviation Embedded in the MACD Indicator
Editor for Accurate Adjustment to Financial Market Dynamics .............................................. page 16
[email protected] by Gino Gandolfi, Monica Rossolini, Antonella Sabatini, Stefano Caselli

An Exploration of the Nature of Bull Market Tops .........................................................page 25

Charting Technical Assistance
by Paul Desmond
Ron William (STA, ESTA)
[email protected] Sentix: Behavioral Indices: A Behaviourally Oriented Development of the
TA Tool-Kit...........................................................................................................................page 26
Jeremy du Plessis (STA) by Manfred Hübner
[email protected]

Graphic Design and Layout

Bussolati MFTA Research Papers
design+marketing
1541 14th Street NW Using a Color Spectrum to Represent Changes in the WAKO Volume Ratio
Washington, DC 20005-3706
www.bussolati.com
on Candlestick Charts........................................................................................................page 32
by Stewart Gault
Printing
Imlay International How well do Traditional Momentum Indicators work?.................................................page 42
6312 Seven Corners Center, #303 by Cynthia Kase
Falls Church, VA 22044
703.879.2274
Harmonic Ratios as Applied To Commodity Market Technical Analysis.....................page 48
Send your queries about advertising information by George Alexander MacLean
and rates to [email protected]
Alteration of Price Movement Dynamics on a Chart via Quantization of
the Change in Closing Prices ............................................................................................page 58
by Mohammed El Saiid

Author Profiles....................................................................................................................page 70

IFTA Directory.....................................................................................................................page 72

IFTA Journal is published yearly by The International Association of Technical Analysts. P.O. Box 1347, New York, NY 10009. USA.
© 2007 The International Federation of Technical Analysts. All rights reserved. No part of this publication may be reproduced
or transmitted in any form or by any means, electronic or mechanical, including photocopying for public or private use, or by any
information storage or retrieval system, without prior permission of the publisher.
 IFTA Journal 2008 edition

Letter From The Editor

Regina Meani CFTe
STA, ATAA, APTA

Striving for excellence and the vigilant search for new and innovative ideas is what brings
together professionals in their chosen field. The International Federation of Technical
Analysts provides several forums to bring together its professionals in our common goal to
expand the field of Technical Analysis of the Financial Markets.
The 20th Annual IFTA conference, held in Sharm el Sheikh, Egypt this year, brings
together Analysts from all over the world around the 2007 theme of Middle East, Energy,
Commodities and the Globalization of the Financial Markets. Over a period of four days
attendees from all continents will meet for presentations, discussions and network with
leading Technical Analysts.
Professional Accreditation in any field is a mark of distinction. As a world-wide
Federation for Technical Analysis, IFTA sponsors the single globally-recognized
professional certifications in the field, the CFTe, and MFTA. Both the CFTe and MFTA
designations set IFTA colleagues apart with standards of professionalism increasingly
welcomed by investors, regulators and institutions.
In a two tiered process, IFTA colleagues may sit for two successive examinations which
culminate in the award and professional qualification in Technical Analysis as a Certified
Financial Technician (CFTe). The exams test not only technical skills knowledge, but also
ethics and market understanding. In the second tier, IFTA’s Master of Financial Technical
Analysis (MFTA) candidates produce an original academic-style research paper proposing
to add to the TA body of knowledge. This second tier is a severe test of professionalism in
the global arena.
This brings us to the IFTA journal which is — through its global distribution to
professionals in the field — now one of the most important forums to publish leading work
in TA. In past years we have endeavored to present the most meritorious work by MFTA
Candidates. In this 2008 edition of the IFTA Journal, we have published four MFTA research
submissions. This body of work offers fresh ways of looking at the behavior of markets and
is testament to the high standing of the MFTA designation.
As the new editor of the journal it is my pleasure to announce some changes in this
edition. Our “call for papers,” sent out earlier this year (and to be repeated every year),
produced a commendable response and we present papers on trend analysis, and market
behavior, market depth and time structure; on the dynamic MACD; and on the development
of a TA tool–kit. Professor Hank Pruden follows up on his Wyckoff paper first published in
the IFTA 2003 Journal.
Towards the end of the journal we introduce you to the writers in “Author Profiles” where
you will find brief biographies. In future editions of the journal we plan a “Books” section
where we will review and discuss ideas from recent and notable publications.
Finally, I would like to thank the Authors of the papers included in this journal, all of
whom have made a contribution to the field of Technical Analysis; I am sure readers will
benefit from your work. My grateful appreciation and thanks go to my team: Roberto Vargas
and Dr. Rolf Wetzer who have been tireless, devoting many hours of their time to edit and
produce a professional journal. For their help and advice in guiding me through my new
role, I thank Larry Lovrencic (former editor) and Elaine Long Knuth (IFTA chairperson).

page 2 IFTA.org
 IFTA Journal 2008 edition

In Memoriam: John C. Brooks

Elaine Long Knuth

For all of us, there are some unique people we will never forget. John Brooks was such an
individual. When he passed away this June, IFTA lost not only one of its founding members,
but an exceptional person, a dear friend and teacher to many.
John was a veteran of Wall Street and of life; and he was neither fooled nor fool-able.
Anyone who met John understood this about him immediately. John had values. He never
put himself out front, but stood back and observed. He had a way of looking at the world
and immediately taking its measure for what it was. Just like that — quick, sharp and with
a glint in his eye. His lightening fast wit was always spot on, yet never with malice. This is
what made him the clear perceptive analyst, and the special person.
In this spirit he was the honest friend and supporter we at IFTA came to know. I was
lucky to know and be associated with John who would always — no matter how busy he was
— answer his phone with a warm and welcoming “Brooklyneze” “How are ‘ya Deaaer…”
I can hear his voice now.
John worked at a number of firms in the canyons of Wall Street: Dominick and Dominick,
Edwards & Hanley, Robinson Humphrey, and more recently at Lowry Research Corporation,
where he served as Senior Vice President and Senior Analyst. John was absolutely dedicated
to the professionalism of TA. He was the co-founder and Past-President of the Market
Technicians Association (MTA). In 1985, he co-founded the International Federation of
Technical Analysts (IFTA) that today includes thousands of members worldwide and
continues to grow. He served as IFTA’s Chairman from 1996-1998. As the long-time
Chairman of the MTA Educational Foundation, John was instrumental in establishing
accredited college-level courses in Technical Securities Analysis. His dedication to TA
and to his colleagues did not stop there. In recent years, John co-Founded the American
Association of Professional Technical Analysts (AAPTA), a member society of IFTA. And
in 2005 John authored the wonderful read, Mastering Technical Analysis published by
McGraw-Hill.
John had a number of awards, and most recently — and one for which he was particularly
proud — he was named a Fellow of the British Society of Technical Analysts. This award is
not lightly given and was in recognition of his outstanding support of technical securities
analysis throughout the world.
Let me end here, not with a greater list of John Brooks professional accomplishments,
but with his own words sent to colleagues at the start of the year — good natured, lively and
ever supportive:

“Hi All: As the year comes to a close I know a lot of

folks will be busy doing other things. I wanted to
take a moment to publicly say thank you and wish
the best. I also would like to wish everyone a great
holiday season. It’s going to be a great 2007 for
IFTA.” John C. Brooks

IFTA.org page 3
 IFTA Journal 2008 edition

Trend Behaviour, Market Depth and Time Structure of

Single Stocks within Stock Indices
by Rolf Wetzer

Abstract In figure 1, the HDAX exhibits an Trend Structure for

The following work is a technical study of extended uptrend from 2003. It recently Single Stocks
markets using the example of the German formed new all time highs but, at the The first step is to describe the trend
HDAX index. The paper attempts to answer same time, appears to be overextended. behavior for every single stock by an
questions on the stability and robustness The question is if this current trend is in indicator, and to plot it in a single graph.
of the prevailing trend using different jeopardy or not. This serves as an overview of the market
technical tools. It takes advantage of a The analysis is based on data from July and allows us to discern which, and how
bottom-up approach to add a perspective 12, 2007 and is created exclusively on MS many shares, support the current trend of
from the point of view of single stocks ExcelTM which can be easily reproduced. the market.
within the total market. Classic technical Additional charts illustrating different To do this, two classic ideas from
tools are used as well as the new concept of points in time are added to clarify or technical analysis are borrowed and
the time structure. Emphasis is applied to expand concepts presented. The study combined together: Donchian’s four week
the graphic compression of information for limits itself to technical data, i.e. prices, rule and the standard inputs using the
single stocks. monthly and annual highs or lows, with Stochastic indicator. According to the
the corresponding dates when these rule of Donchian, a share that trades at
Introduction extreme values were traded in the its four week high should be purchased
Investors are generally interested in the market. Moreover, the market value of the while a share that trades at its four-week
strength of the current trend in the market respective share was processed. low should be sold. The %k measurement
if it has large market depth, and how up-to- As 110 single values are analysed, the of the Stochastic indicator describes the
date the price-driving factors are. analysis does not refer to the movement current market price as a percentage of the
Most of the time, commentary on the of a single share in time, but instead distance between the current price, and its
stock market is based on the analysis illustrates the technical constitution of all four week low relative to the total range
of a particular index. This “top-down” shares as a cross-section on a particular traded within the last four weeks.
analysis rescues the investor from the point in time.
time-consuming work of analyzing The analysis consists of three parts: %k = Ct - Lt-20
individual stocks. He only needs to look at trend behavior (structure), market depth Ht-20 - Lt-20
one aggregate figure, the index. Through (breadth) and time structure. C : = closing price
this kind of analysis, however, valuable L : = lowest price
information on individual shares is lost. H : = highest price
t : = daily time index
Methods and Background
In contrast, this paper looks at the stock
market with a “bottom-up approach,” Figure 1: HDAX price chart
examining the individual shares. It
describes the behavior of all the single
stocks within the index and bundles
this information to deliver an additional
dimension on the total market. The paper
analyses the HDAX index, an aggregated
German index containing 110 stocks of the
most popular German indices: DAX, MDAX
and TecDax. It is a broad market index,
covering the main stocks of the German
Prime Standard. The challenge of using the
HDAX was to extract information from the
110 single stocks into a few easy to read, yet
informative graphics. The classic line chart
of the HDAX is shown in figure 1.

page 4 IFTA.org
 IFTA Journal 2008 edition

This four-week %k coefficient is then % k < 0.75). The existing up-trend of the astonishing is that within the group of
determined for each stock within the HDAX is therefore supported by a quarter down trending stocks, the majority traded
market index. A value of zero means that of the index members and the downward at their relative low, with few trading at
the share has just formed a new low price, trending stocks are not sufficient to turn their relative high. This may be a sign that
while a value of one signals a new high the ship around. Many neutral shares there is downward pressure on the Index.
price. A value of 0.23 means that the share are stabilizing the trend, but their future The great advantage of the Range
price has recovered 23% from its lowest low behavior remains unclear. Distribution is its simplified representation
price of the last 20 trading days, relative to For another example, we show the Range of the trend behavior of all single values in
its trading range. Distribution of the Nikkei 225 for July 12, the selected index. In addittion, concrete
If we plot this coefficient for each single 2007 (figure 3). If we compare it with the ideas and analysis for single stocks may be
stock in the index, we create a Range HDAX, the Nikkei seems to be more distinct derived from this tool.
Distribution chart. An example of a Range in its extreme values, with more green and It is simple to screen the %k coefficients
Distribution chart for the HDAX is shown in red stocks at the top and bottom relative to of the individual shares in order to look for
figure 2. the HDAX. Therefore, the yellow (neutral) simple buy and sell candidates. Following
How should we read this chart? Every stocks are not so cumulative. What is the idea of combining Donchian and %k,
single point on the range distribution
represents a share of the HDAX. The %k
values of the respective stocks are shown Figure 2: HDAX range distribution (July 12, 2007)
on the y-axis. These values change daily.
The chart does not consider time, but
instead a cross-section of all shares in the
index at a single point in time.
If this chart shows many points (stocks)
at 0.75 or above, a variety of shares are
approaching new highs. These shares are
marked green in the chart as a symbol
of their uptrend. If, on the other hand,
the chart shows many stocks at 0.25
or lower, many shares are approaching
new lows. These shares are marked red,
representing a downtrend. Shares with
a %k between 0.25 and 0.75 swing to and
fro unenthusiastically in a trading range,
exhibit a neutral trend behavior, and are
marked in yellow. Shares that tend to go
up or down in a directionless manner, are
considered the motor for future trend
behavior. If many of the yellow stocks
become green or red stocks in the future,
they will add to the ongoing trend. Usually, Figure 3. Nikkei 225 range distribution (July 12, 2007)
stocks that are already red or green will
fall back into the yellow range after some
time, but tend not to cross the whole area
in one step.
In summary, the Range Distribution
graph resembles a traffic light reflecting
the trends of single stocks within an index.
From this, it is possible to gain, at a glance,
additional insight to the trend for the
overall market. It contains all the trend
information for the last trading month,
not in absolute terms, but relative to each
stock’s trading history.
At the time of this analysis, 26 of 110
shares exhibit a %k over 0.75 in the trend
mode, whereas 38 stocks trend downwards
(%k less than 0.25). The remaining 46
shares are in a neutral condition (0.25 <

IFTA.org page 5
 IFTA Journal 2008 edition

we can generate a simple list of all stocks climbing, while most issues are falling. The
Table 1. “Green” HDAX single which are green in the chart, i.e. with a %k entire market represented by capitalization
stock list (July 12, 2007) greater than 0.75. Below in Table 1 is a daily is trading at its lows.
list for the HDAX. A week later, however, the situation
The same list can easily be generated dramatically changed as the Nikkei climbed
for stocks in a downtrend, exhibiting a %k by 1000 points from 15273 to 16248. What
of 0.25 or less. These stocks are the “red” is even more interesting is that nearly all
stocks in the Range Distribution. See below issues recovered from their lows and the
in Table 2. majority of the market cap had changed
With the Range Distribution and the into a trend-less state (0.25 ≤ %k ≤ 0.75) as
corresponding screening lists, it is possible shown in figure 6.
to quickly judge the health of the market. As figure 6 illustrates, analyzing only
the trend behavior of a stock may not be
Market Breadth of the Current sufficient, and that we may also need to
Index Trend consider the relative capitalization of the
The second part of the analysis is devoted stock to capture a better understanding of
to market breadth. So far, we have analyzed market events.
the current trend behavior of single The effect of market capitalization itself
stocks. Often, however, the question is if can be measured by a new indicator called,
the trend in the index is broadly based, Market Cap Trends (MCT). This indicator
or if it is only supported by a few highly explains the percentage of the market
capitalized shares. Technicians will ask capitalization exhibiting a certain kind of
whether the trend is supported by market trend behavior. It answers the question
breadth or not. A well-known example of a of which part of the market cap is in an
trend supported by only a few stocks was uptrend, and which part is in a downtrend.
Table 2: “RED” HDAX single the “tech bubble” in 2000-2003, when In turn, the MCT measures the trend
stock list (July 12, 2007) the German market was driven mainly by behavior in terms of the %k value of single
Telecom stocks. stocks. From this value we ascertain where
In the example of the HDAX Range the stock in question is trading relative to
Distribution, we observed more down its own four week trading range. Again, we
trending stocks (red) than up trending can add the relative market cap of the issue
stocks (green). As the market is in an to the trend information, and note that the
uptrend, either the market value of the relative market cap is nothing more than
rising issues must be more important than the index weight of the issue in question.
the market value of the falling issues, or If one organizes all the shares of an
the trading range of the falling stocks are index which are rising due to its %k values
tighter relative to the ranges of the rising and subsequently sum up the relative
issues. To include trading size into trend market capitalization of the individual
analysis, we can use the relative market index members, we arrive at the MCT. It is
capitalization of the single shares within an empirical distribution function, which
the index, i.e. the index weight, as a proxy starts at zero in the left part of the chart,
for volume. and ends at one at the right upper corner
For example, if we plot the range of the chart. From this indicator we can
distribution as a bubble chart where the read the percentage of the total market
market capitalization defines the size of the capitalization, which is exhibiting certain
single bubble, the picture of what is driving trend behavior (as defined by %k): for
the market trend becomes clear (figure 4). example finding out how much market cap
It can easily be seen that small issues has a %k coefficient of 60 (or less). This
exhibit bullish behavior, while the larger information can be extracted from the MCT
capitalized stocks are either trend-less or chart as shown in figure 7.
already in a downtrend. This illustrates that The MCT is a now useful indicator.
the risk of a market shift to the downside is However, it shares the typical problem of
a likely scenario. all statistical distribution functions—it is
To give more examples, we can compare neither read nor understood intuitively. The
the Nikkei on August 17 and 23, 2007. indicator becomes easier to read and more
On August 17 the Nikkei was in a severe intuitive to interpretation if represented as
downtrend. This situation is exhibited in a histogram.
distribution chart (figure 5). One issue is To do this, we first built a number of

page 6 IFTA.org
 IFTA Journal 2008 edition

Figure 4. Size Weighted Range Distribution of 110 HDAX stocks

(July 12, 2007)

Figure 5. Size Weighted Range Distribution of 225 Nikkei

stocks (August 17, 2007)

Figure 6. Size Weighted Range Distribution of 225 Nikkei

stocks (August 24, 2007)

IFTA.org page 7
 IFTA Journal 2008 edition

Figure 7. Market Cap Trends of the HDAX index (July 12, 2007) quantiles for the %k values. In our example,
the relative market capitalization of the
HDAX is divided up into groups of %k
values with 5% intervals each, i.e. 0, 0-5%,
5%-10%, etc. This leads to the following
empirical density function (figure 8).
The message from both curves is
identical: we can see that a huge part of the
market cap is at the very low end of its own
trading range. This tendency is balanced by
the larger capitalized stocks with ranges in
the middle of the scale. Stocks with a %k of
0.8 or more account only for about 20% of
the total market cap. Because the market
is still heading north, it would imply that
these stocks are driving the market. This
indicator, however, also indicates that the
trend is, indeed, rather weak as it is not
supported by large volume.
Overall, we cannot say that the market
Figure 8. Market Cap Trends histogram of the HDAX index is overheating as if that were the case, we
(July 12, 2007) would see much more of the market cap
moving to the right of the indicator. The
MCT histogram might rather be interpreted
as an early warning signal: more market
cap may slip to the left and, therefore,
create downward pressure on the index. For
the moment, the market appears to be still
in a balanced state.
To complete this part, we show the two
MCT-histograms for the Nikkei. On August
17, the market was down and the trend sup-
ported by most of its market cap (figure 9).
It is evident that not all issues could
continue to fall at the same time ad
infinitum and, therefore, a recovery
seemed to be more likely than not, as the
subsequent MCT histogram for the next
week confirmed.
Figure 10 demonstrates that most of the
Figure 9. Market Cap Tends histogram of the Nikkei225 index market cap shifted from the left side of the
(August 17, 2007) scale to the center. This indicated strong
market recovery, supported by market
breadth.

Temporal Movement Analysis

The last part of this analysis examines the
time structure of trends in single stocks
within the indices. For example, during the
period of a year, the members of an index
drift up and down marking new yearly
highs or lows. If the stock is in a trend,
these extreme values will pop up at the left
end of the chart and their positions may be
defended for a period. In a range market,
stock prices typically form new highs or
lows but then fall back into the old trading
range, i.e. into the mass of trend-less

page 8 IFTA.org
 IFTA Journal 2008 edition

stocks. The typical characteristics of trend- Figure 10. Market Cap Tends histogram of the Nikkei225 index
less markets are marked by extreme values (August 23, 2007)
far from the current date.
A new method, Temporal Movement
Analysis, is based on the characteristics
of the temporal occurrence of yearly high
and low values. This is a representation
in contrast to point & figure analysis, as it
does not measure price movement itself,
but instead measures how much time
has passed since the prices reached their
extreme values. This temporal behavior is
analyzed for each single stock within the
selected index. It is then aggregated and
transformed into the Temporal Movement
Analysis indicator.
The first step is to collect the dates
when each single stock within the index
traded at its highest high or lowest low
within the annual period. Second, we
measure which extreme value was reached
last and how many weeks have passed
since this event. In the third step, these
distances are then divided into different
classes, which are indicated by characters, Figure 11. Temporal Extreme Value Distribution (static) of the
along the following lines: HDAX on July 12, 2007
A: = the last annual high was within the
past 2 weeks
B: = the last annual high was between 3
and 4 weeks
C: = the last annual high or annual low is
older than 4 weeks
D: = the last annual low was between 3
and 4 weeks
E: = the last annual low was within the
past 2 weeks
In a final step, one counts how many
of the 110 stocks of the HDAX index
are allotted to the individual segments
above. Plotting this information into a
chart provides the following result, called
Movement Potential (See figure 11).
In contrast to the two previous
analyses, where we discussed if a stock is
in a trend and where the market cap was
trending, Temporal Movement Analysis
focuses exclusively on the temporal
dimension (or time limited dimension) of
the market. Figure 11 shows the temporal
development of the extreme prices and,
with that, the trend behavior of all 110
shares over a year. If currently all stocks
are forming their annual highs or lows,
there may be less likelihood or potential
for the trend to continue.
In the sample above, about 40 percent of
all shares in the HDAX had reached a new
annual high within the past two weeks.

IFTA.org page 9
 IFTA Journal 2008 edition

This indicated a recent robust movement

Figure 12. Temporal Extreme Value Distribution (dynamic) of of the index. Additionally, only five shares
the HDAX on July 12, 2007 (about 5%) made their annual high three
to four weeks previously. This implied
either that new highs were defended, and
were therefore followed by higher highs;
or that the market was exhibiting a new
uptrend impulse move. This also illustrated
that new stocks were making new highs,
with only a small portion of the market
in a prior uptrend. These stocks could not
achieve new highs and are labelled “B” in
the indicator.
More than half of all stocks in the index
were trend-less within the last four weeks
and did not reach new annual highs or lows,
with only a very small portion touching
new lows. A market with this characteristic
appears stable, as there is no downward
pressure, and there remains a big pool
of stocks, which might form new highs,
allowing room for stock or sector rotation.
Additionally, a small portion of the market
is “trendy” and moving the index. This
analysis is conducted without looking at
prices, only at time, and looks at the market
from a bottom-up point of view through a
section of all stocks.
So far, the Temporal Extreme Value
Distribution is a valuable, yet static tool,
presenting the market condition of one
section at a very specific point in time. It
does not show whether or not a temporal
rotation of a group is about to take place.
This static quality can be overcome by
comparing two times periods. If we do
this, a Dynamic Movement Potential arises,
which is shown for the HDAX in Figure 12.
This is a rather abstract graph and may,
upon first glance, not be easy to read. This
graph is a portrait of the HDAX stock index,
illustrating that the time structure of the
index changed from last week to this week.
For example, when adding up the letters
in vertical axis “to this week” (i.e. for the
letter A it is 14+29) you will arrive at the
static diagram shown in Figure 12.
What does this tell us? Only fourteen
stocks could defend their A position in the
HDAX from one week to the next (they were
A last week and remain A the following
week). This could mean that there was
either a new annual high or that the annual
high was only a week-old in the previous
week. Twenty nine stocks (out of 110)
formed a new high during the last week,
moving from the C group to A. There were
no new B or C stocks during this period,

page 10 IFTA.org
 IFTA Journal 2008 edition

implying that the market defended its

highs. Finally, the three stocks which were
losing value in the previous week held their
position as E.
With this graph one can vividly see the
rotation taking place by stocks within
the index. In this case, with new stocks
moving to new annual highs, and no stocks
weakening, the picture is of an overall
stable condition of the stock market. In
addition, during this period there is no
downward pressure from new annual
lows, while a large portion of the market is
trend-less (C stocks), indicating room for
development of new annual highs.

Conclusion
The analyses above examined the HDAX
through a scrutiny of the individual shares.
The technical market information of the
single stocks was collected and bundled
in different, partly new indicators and
diagrams, at a specific point in time. In
addition to the pure analysis, the objective
was to give the Technical Analyst tools at
his fingertips with which he can quickly
attain a summary of the current market
events in full breadth.
Making use of these tools in analysing
the HDAX, we may infer here that while the
index is at high levels, it remains in a stable
trend environment. However, this condition
is not supported by the market breadth.
The bottom-up analysis demonstrated
many shares are high relative to their own
price history of the last four weeks. We
note, however, that these stocks represent
a small part of the total market cap, and the
shares exhibit strength as they establish
themselves at these highs. Nonetheless,
there are clear risks that a falling market
cap may break the current trend. IFTA

IFTA.org page 11
IFTA Certification Programs

CFTe
Certified Financial Technician
Cost of the exams: Level I, USD $500; Level II, USD $800

The CFTe consists of Levels I and II, which together constitute a complete
professional program. The two examinations culminate in the award of the
internationally recognized professional qualification in Technical Analysis.
The exams test not only technical skills, but also ethics and international
market knowledge.
What is IFTA?
Level I This multiple-choice exam covers a wide range of technical knowledge IFTA offers certification to Professional Technical
and understanding of the principals of Technical Analysis, usually not involving Analysts. The International Federation of Technical
actual experience. Analysts was incorporated in 1986 and is a global
organization of societies and associations. IFTA is
Level II In this you should demonstrate a depth of knowledge and experience a non-profit professional organization with member
in applying various methods of technical analysis. The exam provides a societies in more than 26 countries.
number of current charts covering one specific market (often an Equity), to be
analyzed as though for a Fund Manager. What is Technical Analysis?
Technical Analysis is the systematic method of
analyzing financial instruments, including securities,
futures and interest rate products, with only market
delivered information such as price, volume, volatility
MFTA and open interest. The tools of technical analysis are
measurements and derivatives of price, for example
Master of Financial Technical Analysis on-balance volume, price oscillators, momentum
measurements and pattern recognition. A Technical
Technician Cost of the exam: USD $900 Analyst applies such tools for forecasting and timing the
trading and investing in financial instruments. Technical
Analysis is a universal discipline.
The MFTA requires an academic style research paper. It is intended to be a
rigorous demonstration of professionalism in the global arena of Technical When and where are the examinations held?
Analysis. It is anticipated that most candidates for the MFTA will have some Examinations are held twice annually and can be given
anywhere in the world. Traditionally, examinations
academic background. This should convey the high standard against which
are given in Frankfurt, Geneva, Paris, Lugano/ Milan,
these papers will be judged. London, Hong Kong, Madrid, Singapore and at the
Annual IFTA Conference.
Your paper must meet the following criteria:
> It must be original. How long does it take to complete the programs?
There is no fixed time limit, although candidates
> It must develop a reasoned and logical argument and lead to a are encouraged to complete the Certified Financial
sound conclusion supported by the tests, studies and analysis contained Technician program in two years, and the Master of
in the paper. Financial Technical Analysis in one.

> The subject matter should be of practical application. Do I have to be an IFTA Member to take one
> It should add to the body of knowledge in the discipline of International of these programs?
You must be a member of a national IFTA society.
Technical Analysis.
Check the list of IFTA societies on the IFTA homepage.
If there is no IFTA member society in your country, most
societies will welcome you as a nonresident member.

I have been a Technical Analyst working in the

markets for many years, can I directly go on
to the MFTA?
International Federation of In general, there are three categories of applicants for
Technical Analysts the Alternate Path. It is open to IFTA Colleagues who:
• Have technical analysis certification such as CMT,
STA Diploma, PLUS three years experience as a
technician
• Have financial certification such as CFA, CPA, MBA,
Register online at PLUS five years experience as a technician
www.ifta.org/certifications • Have experience only (eight years minimum)
 IFTA Journal 2008 edition

Wyckoff Laws: A Market Test (Part B) — What Has

Actually Happened?
By Henry (Hank) Pruden and Benard Belletante

The Original 2003 Prediction: educational enterprise in the early 1930s chart (figure 1) indicated 72 boxes between
“… the point-and-figure chart reveals a the Wyckoff Method has stood the test the right inverse-shoulder and the left
base of accumulation for a potential rise of time. inverse-shoulder. Each box had a value of
of 7,200 points, when added to the low of 100 Dow points. Thus, the point-and-figure
7,200 the price projects upward to 14,400… A Market Test chart revealed a base of accumulation for a
before the onset of the next bear market. The authors, as academics, have been potential rise of 7,200 points. When added
If the Dow…comes within – or + 10% of the intrigued by the natural laboratory to the low of 7,200 the price projected
projected 7,200 points, we will accept the conditions provided by a stock market. A
prediction as having been positive”. prediction study is the sine quo non of a Figure 1.
Pruden and Belletante, IFTA Journal 2004 good laboratory experiment. The Wyckoff
Law of Cause and Effect seemed to us to
The 2007 Result: provide an unusually fine instrument for
On June 1 and June 4, 2007 the DJIA conducting such an experiment, a “forward
recorded a print high in price of over test”. Parenthetically, it has been our
13,680. By reaching 13,680 the DJIA met feeling, shared by academics in general,
the lower price objective that was predicted that technicians have focused too heavily
(i.e. 14,400-720 = 13,680). upon “backtesting” and not sufficiently
upon real experimentation. The time
Conclusion: series and metric nature of the market
The Wyckoff Law of Cause and Effect data allow for “forward testing”. Forward
successfully passed its market test of testing necessitates making a quantitative
DJIA of 7,200 ±10% or 14,400 ±5%. prediction then followed by the empirical
test of the prediction with market data that
The authors noted in their article (Part A) tell what actually happened. Thus, the first
that Wyckoff was a name gaining celebrity article that appeared in the 2004 issue of Price and Time Projections
status in the world of Technical Analysis the IFTA Journal commenced a series of In summary, U.S. equities are in bull market
and Trading. Richard D. Wyckoff, the man, articles devoted to study Wyckoff and use with a potential to rise to Dow Jones 14,400. The
worked in New York City during a “golden the Law of Cause and Effect and Point-and- anticipation is for the continuance of this powerful
age” for technical analysis that existed figure charts to answer the question bull market in the Dow Industrial Average of the
in the early decades of the 20th Century. of “how-far” will the 2003 onward bull U.S.A. through 2004. This market forecast is the
Wyckoff was a contemporary of Edwin market rise. “test” to which the Wyckoff Method of Technical
Lefevre who wrote The Reminiscences of Analysis is being subjected.
A Stock Operator. Like Lefevre, Wyckoff Wyckoff’s Law of Cause and Effect – Part (B) of “Wyckoff Laws: A Market Test” will be a
was a keen observer and reporter who postulates that in order to have an effect report in year 2005 about “What Actually Happened”.
codified the best practices of the celebrated you must first have a cause and that effect As with classical laboratory experiments, the results
stock and commodity operators of that will be in proportion to the cause. The law’s will be recorded, interpreted and appraised. This
era. Around 1930 the results of Richard operation can be seen working as the force sequel will invite a critical appraisal of the Wyckoff
Wyckoff’s observations, personal of accumulation or distribution builds up Laws and in particular a critical appraisal of the
experience and journalism became known within a trading range and works itself out Wyckoff Law of Effort vs. Result. The quality of the
as the Wyckoff Method of Technical in the subsequent move out of the trading author’s application of the Wyckoff Laws will also
Analysis and Stock Speculation. In 2007, range. Point-and-figure chart counts can be undergo a critique. From these investigations and
Pruden, a co-author of these articles, used to measure this cause and project the appraisals, we shall strive to extract lessons for
wrote a book on The Three Skills Of Top extent of its effect. the improvement of technical market analyses.
Trading that spot-lighted the principles and Irrespective of the outcomes of this market test, we
procedures of the Wyckoff Method. Using the Inverse-Head-and-Shoulders are confident that the appreciation of the Wyckoff
Wyckoff is a practical, straight forward formation as the base of accumulation Method of Technical Market Analysis will advance
bar chart and point-and-figure chart from which to take a measurement of the and that the stature of Mr. Richard D. Wyckoff will
pattern recognition method. Since the “cause” built during the accumulation not diminish.
founding of the Wyckoff and Associates phase of 2002-03, the point-and-figure From Pruden and Belletante, IFTA Journal 2004

IFTA.org page 13
 IFTA Journal 2008 edition

upward to 14,400. Hence, the prediction in price or time. A price retracement would DJIA is surpassed (14,400 ± 5%), that event
was for the Dow Industrials to continue normally entail a one-half correction of would open the door for re-visiting the
to rise to 14,400 before the onset of the previous uncorrected advance. Judging 2002-03 accumulation base in order to
distribution and the commencement of the from the Chart (figure 2), the DJIA could make a larger base count from a higher Last
next bear market. retrace to approximately 12,000 under this Point of Support. (Please refer to figure 3).
If the Dow during 2004-2005 comes scenario. A second possibility envisioned Because the price breakout above the price
within + or – 10% of the projected 7,200 under the Scenario One correction would be resistance line in March 2003 was a sign
points we will accept the prediction as a sideways trading range of reaccumulation of strength, a higher level count would be
having been positive. after which either the bull market trend valid at point 4.
could re-assert itself or the sideways could The Wyckoff Count Guide says that a
Interim Report: Date Wrong, turn out to be distribution in preparation more important Sign of Strength (SOS)
Price Right for a primary bear market. followed by a more important Last Point of
Calendar dates and cycle time frames were The second scenario once again calls Support (LPS) is valid for taking a count. On
never included by Wyckoff in the Wyckoff upon the Wyckoff Count Guide. If 15,120 the 2002-03 point-and-figure chart of the
Method of technical analysis. Essentially,
Point-and-Figure Chart projections are
valid until a market by its own action The Wyckoff Count Guide: Up Count
indicates that the direction of the price After seeing a Sign of Strength, locate the Last Point of Support on a Reaction and
trend has reversed. Pruden’s definition count from right to left.
of a bear market would be confirmed
mechanically by the major equity averages Detailed Count Guide – Up Count
(DJIA, SPX and NASDQ) falling and After having identified a Sign of Strength (SOS) on the vertical line chart, locate the
remaining under their respective declining last point at which support was met on a reaction – the Last Point of Support – (LPS).
200-day moving averages. A good source Locate this point on your figure chart also and count from right-to-left, taking your most
for these indications can be found in the conservative count first and moving further to the left as the move progresses.
Investor’s Business Daily newspaper. In moving to the left, turn to your vertical line chart and divide the area of accumulation
The authors’ anticipation that the DJIA into phases, adding one complete phase at a time. Never add only part of a phase to your
14,400 target zone would be reached by count. Volume action will usually show where the phase began and ended.
2005 was clearly wrong. The DJIA did not As the move progresses, you will often see a lateral move forming at a higher level. Very
hit the 14,400 target during 2005 nor often such a move will become a “Stepping Stone Confirming Count” of the original count.
did the U.S. Equities roll over into a bear Thus, as such a level forms; you can often get a timing indication by watching the action of
market during 2005. Our time prediction the stock as the potential count begins to confirm the original count. A resumption could
was clearly inaccurate: The market did not begin at such a point.
reach the bull market price target near For longer-term counts one should add his count to the exact low, or a point about
14,400 until 2007. Moreover, the major one-half way between the low and the count line. You will thus be certain that the most
equity indices remain above rising 200-day conservative count is being used.
moving averages. Counts are only points of Stop, Look and Listen, and should never be looked upon as
An honest test of the Wyckoff Law of exact points of stopping or turning. Use them as projected points where a turn could occur,
Cause and Effect would be based upon price and use the vertical line chart to show the action as these points are approached.
behavior and only price behavior. In that In case of a longer-term count often the Last Point of Support (LPS) comes at the original
regard, the Wyckoff Method has once again level of climax, and this level should be looked at first in studying the longer-term count.
proven its utility by reaching the minimum The climax itself indicated a reversal, with the subsequent action being the forming of
price objective predicted in 2003. the cause for the next effect. For the Last Point of Support (LPS) to come at such a level of
climax usually makes it a more valid count. Very often the climax is preceded by preliminary
Possible Future Trend of support and the Last Point of Support often occurs at the same level as the preliminary
the DJIA support.
We anticipate that the DJIA will continue A no.3 Spring or the Secondary-test of a no.2 Spring, quite often constitutes the Sign
to conform to the “Wyckoff Count Guide” of Strength and the Last Point of Support in the same action which is reached at the same
for Point-and-Figure charts and the law of point and at the same time. Usually a Spring will be followed by a more important Sign
Cause and Effect. Please read the following of Strength and the reaction following that Sign of Strength is also a valid Last Point of
Box for a statement of the Wyckoff Count Support.
Guide. For the trader or investor, mid-2007 Frequently, long term counts on three and five point charts are confirmed by subsequent
for the DJIA is at what the Count Guide minor counts on the one point chart as the move progresses. Watch for this confirmation
refers to as a “Stop, Look and Listen Point.” very carefully as it often indicates when a move will be resumed.
From the present position, three In the case of three point or five point charts, the same count line should be used as for
different scenarios for the future price the one point chart.
trend of the DJIA are possible. The first Source: Stock Market Institute, Basic Lecture no. 8
scenario would be an extensive correction

page 14 IFTA.org
 IFTA Journal 2008 edition

Figure 2.
DJIA (figure 3) that SOS would have been
the break-out above the neckline resistance
level of the DJIA; the LPS would have been
at the pullback on the chart to point 4. All
nine of the Wyckoff & buying Tests would
have been passed at that juncture or point
4 on the Chart (figure 3). The Count taken
from that higher, more important P n F
price level projects upward to 19,000 to
20,800.Hence, if the DJIA should rise above
and hold above 15,120 (i.e., 14,400 plus 5%).
We would predict a continued rise until the
19,000-20,800 price range is reached by
the DJIA.
Scenario Three from the present
position envisions a reversal in the primary
price trend of the DJIA from a bull market
into a bear market. Hence, no further price
Figure 3. advances to a new high above 13,680 would
emerge. Rather, a prolonged and persistent
price decline of lower highs and lower lows
would forthwith turn the prices of the
DJIA, the NASDAQ and S&P 500 below their
respective declining 200- day
moving averages.

Appraisal of Wyckoff
Irrespective of which of the three foregoing
scenarios eventuates, it is our considered
judgment that the Wyckoff Method’s
Law of Cause and Effect passed a critical
market test to further authenticate it’s
effectiveness as a market analysis tool.
Furthermore, the forecast of 14,400 not
only provided the trader-technician with a
tool of measuring comparative reward to
risk, but it also furnished a flag, an anchor
point in the future. The flagged Count
on a point-and-figure chart is a powerful
psychological anchor that helps to keep the
trader and the technician focused on the
probable future trend of the market. The
projected target inspires confidence in the
trader to patiently let the P n F count
work out.
When the current bull market trend
of U.S. Equities does reverse into a new
primary trend bear market, the authors
REFERENCES will report in Part (C) which of the three
Definition of a Bear Market (given in “A PRIMARY TREND PROJECTOR,” by Pruden, Three
scenarios offered in this article finally does
Skills of Top Trading), Wiley 2007, pages 22-27
eventuate. IFTA
BIBLIOGRAPHY
1. Pruden, Hank, The Three Skills of Top Trading, John Wiley and Sons, 2007.
2. Pruden, Henry and Belletante Bernard, “Wyckoff Laws: A Market Test (Part A)”, IFTA
Journal, 2004-Edition, pages 34-36.
3. “The Wyckoff Count Guide”, Wyckoff/SMF Basic Lecture No. 8, Wyckoff/Stock Market
Institute.
4. Charts and data courtesy of Bloomberg and Updata.

IFTA.org page 15
 IFTA Journal 2008 edition

Dynamic MACD: Standard Deviation Embedded

In The MACD Indicator For Accurate Adjustment
To Financial Market Dynamics
by Gino Gandolfi, Monica Rossolini , Antonella Sabatini, Stefano Caselli

Abstract Trend following indicators, or lagging moving average, named “Kaufman Adaptive
A significant limit of technical analysis indicators, are characterized by their Moving Average”. The term adaptive
tools is their capability of providing capability to perform well in trending outlines the capability of such an indicator
useful trading signals in specific market markets. In these market conditions, the to automatically adapt its effective length
conditions only. In particular, lagging trend of the market price dynamics is to current “volatility” levels. In order to
indicators perform well mainly in trending clearly either rising or decreasing. On the operate on the “volatility”, Kaufman uses
markets, whereas leading indicators other hand, during non-trending market the Efficiency Ratio, a tool capable of
are useful and function better in trading conditions, when market prices do not indicating market “volatility” levels.
market conditions. The new tool, illustrated exhibit a clear direction, often lagging The Efficiency Ratio is defined, in a
herein, denominated Dynamic MACD tools yield false signals. While leading predetermined time period, by the ratio
(Dynamic Moving Average Convergence- indicators provide useful signals during of the following quantities: the net price
Divergence), exhibits the peculiarity of trading (non-trending) market conditions, movement (determined by the difference
functioning, both in trending and trading and generally fail to perform well in between the last observation close price
markets. trending markets. viii ix x and the first observation close price in the
The latter indicator is constructed with A classical problem in technical analysis time period) over the total price movements
the difference of two Vidya (Variable Index is algorithms that function well either in (determined by the sum of all absolute
Dynamic Averages). Subsequently, the trending phases or trading range phases values of the returns in the same time
efficacy of the Dynamic MACD has been of financial markets, but not in both period). This indicator approaches zero
verified by empirical analysis on the daily situations. This problem, targeted as the when the market dynamic is non-trending,
time-series of market indices, stocks and typical and classical hurdle of technical and it approaches one during trending
exchange rates, in an eleven-year period, analysis attracts a considerable number of markets.
1996-2006. The resulting data tests confirm technicians and researchers worldwide. An Efficiency Ratio approaching zero
the higher efficiency of the Dynamic MACD In literature, it is possible to outline indicates that the market is in a trading
relative to the MACD and outline skillful two major contributions devoted to the range; the market exhibits continuous
returns. development of forecasting models, based variations of price levels both in amplitude
on the automatic adaptability to ever and in direction (up or down). In this kind
Introduction varying market “volatility”: Kama xi and of market behavior it is necessary to use
Technical analysis has been developing Vidya. xii a long moving average, in such a way that
into more and more rigorous, quantitative In this work “volatility” is defined as a the moving average is kept away from the
and scientifically based approaches and measure of trending versus trading range time series by mathematical construction,
research methods over recent years. During behavior. High “volatility” indicates a avoiding numerous and frequent cross-
this development process, the typical strong trending market. By contrast, low overs (intersections) which would provide
graphical approach, mainly based on “volatility” indicates the market is trading frequent false signals. An Efficiency
chart analysisi ii, has been largely replaced in a range. xii xiii xiv Ratio, however, approaching unity
by more computationally intensive and Kama and Vidya are two particular mov- indicates that the market has a trending
systematic methodologies, typical of the ing averages derived from the exponential behavior, it moves in a clear direction, it is
algorithmic technical analysis. However, moving average. The exponential moving characterized by low noise: prices tend to
technical analysisiii iv indicators still exhibit average formula is composed from a “vola- keep similar amplitudes and directions in
a limitation in terms of their capability tility” indicator to yield Kama and Vidya. their action. In these market conditions it
of providing useful trading signals in The functional composition allows the is advisable to use a short moving average
specific market conditions only. Technical latter indicators to automatically decrease so that profitable trades can be obtained, by
analysts categorize technical indicators their length as the market trends as well as locking in positive returns.
into two major macro groups; each group is being able to increase their length in trad- Kaufman’s idea consists of developing
characterized by similar features: lagging ing range markets. a model in which the Efficiency Ratio
indicators and leading indicators. v vii Kaufman developed xi a new model of information is embedded in his model.

page 16 IFTA.org
 IFTA Journal 2008 edition

The model encompasses the construction is defined by the difference of two expo-
of an exponential moving average able to This is the key point in support of the nential moving averages. Variable MACD is
automatically adjust its length, switching usage of the variable moving average: to defined by the difference of two Vidyas.
from a short moving average during automatically decrease the length upon The name “Variable MACD”, was chosen
trending markets to a long moving average an increment of market “volatility”, and in order to emphasize its ability to adapt
during non-trending market conditions. to automatically increase the length when to varying market “volatility” levels. K,
In addition to Kaufman’s model, “volatility” decreases. being the “volatility indicator” embedded
another contribution, appearing often For the construction of Vidya, Tushar in the Vidya expression, was defined by the
in literature, is called Vidya (Variable Chande decided to use the exponential absolute value of the CMO. Empirical tests
Index Dynamic Average). This indicator moving average as follows: have shown that Variable MACD yields a
also aims to overcome the problem of better return with a smaller number of
the limited adaptability of the commonly Vidya t = α* k* Ct + (1-α*k)*Vidyat-1 trades compared to the MACD.xiii
used technical analysis tools to trending/ The indicator should indentify a
non-trending market conditions. The Where: trending market, supporting analysts and
Vidya indicator was developed in the 90’s • C t = close at time t. traders in taking profitable positions; as
by Tushar Chande who reformulated the • Vidyat-1 = previous day Vidya. well as being able to avoid indications for
exponential moving average. xii • k is a “volatility” index that indicates unprofitable positions during trading range
Chande’s indicator by its definition and when the price action is heating up or phases of the market.xvii
construction reduces or eliminates the cooling down.xii The aim of this present work is to test
analyst’s discretionary decision process • α determines the effective length of the and verify the efficacy of another indicator,
for the defining of the length of the moving exponential moving average that needs similar in logical construction to the
average. The application of the indicator to be modulated. Chande’s definition of Variable MACD. This new indicator embeds
to historical time series has the effect α = 0.5 (Please note: the analyst can set α in its Vidyas, another “volatility” indicator,
of dramatically reducing false signals, to be equal to any value, but the quantity namely the standard deviation of the close
relative to similar action, produced by the (1-α*k) should always be positive). pricesxii xviii rather than the absolute value
traditional moving averages, especially in of the CMO. This novel indicator is called
trading range markets. The objective is to If k exhibits a small value in the Vidya Dynamic MACD.
obtain a moving average capable of self- calculation only a small portion of new data
adjusting and varying its effective length is considered and the effective length of the The Method
relative to the current market “volatility”. average increases. If k approaches large The Dynamic MACD is defined by the
Vidya is a dynamic indicator. It is values (very active market), the most recent following algorithm: a difference of
essentially an exponential moving data has a higher weight and the effective two quantities is taken, in a similar
average that adapts its length to market length decreases. This behavior allows the fashion to the MACD. The two quantities
“volatility”. Market “volatility” can be average to follow the quick price movement instead of being defined by exponential
defined and measured as follows: by at its best. When k takes on relevant values, moving averages of 12 and 26 days
taking a momentum indicator (i.e. Relative indicating high “volatility” in the market, in length are implemented by two
Strength Index [RSI]xv or the absolute value the Vidya average, automatically, takes on Vidyas of 12-days and 26-days long
of Chande Momentum Oscillator [CMO] xii shorter time periods. Therefore, more reac- respectively. The main characteristic of
ranging between 0 and 1), operating on the tivity allows the average to follow the price the indicator is its adaptability to market
standard deviation of the closexii, or via the movement more closely catching the trend- dynamics with improved returns over
R2 (ranging between 0 and 1). ing dynamics. When k approaches zero, the MACD. Whereas the MACD, defined
If a 10-day simple moving average indicating a possible trading range phase, as a trend following indicator, tends to
is calculated, market “volatility” is not Vidya automatically increases its length, perform poorly in trading range market
taken into account. However, the error separating from its time series of prices, conditions, Dynamic MACD is able to
defined by the difference of current price and reducing the number of intersections. adapt better to the market conditions.
and its moving average increases when Vidya is the basic component used in the The basic algorithm of this new indicator
the market changes direction abruptly definition and construction of another indi- is as follows:
and dramatically. By observing a 5-day cator: the Variable MACD.xiii xvi The objective Dynamic MACD = 12_day_Vidya– 26_day_
moving average and a 10-day moving of this indicator is to function as a tool able Vidya
average, the weight of each sample to provide profitable trading signals with In detail:
included into the moving average is 20% the best accuracy possible both in trend- Dynamic MACD = [α* k*Ct + ((1-α*k)*Vidya
and 10% respectively. The short moving ing and trading range phases of market (12)t-1)] – [β*k* Ct+((1-β*k)*Vidya(26)t-1)]
average responds more quickly than the dynamics. Variable MACD is essentially Where:
longer moving average because there is developed by using the same logic underly- Ct = close price at time t
a greater weight given by the new data ing the MACD. What distinguishes Variable Vidya(12)t-1= Vidya at time t-1 with time
sample entering the calculation, and the MACD from MACD is in the definition of the period (length) of 12 days
gap between the last price and its moving difference of two moving averages. MACD
average is small.xii

IFTA.org page 17
 IFTA Journal 2008 edition

Vidya(26)t-1= Vidya at time t-1 with time Crossing from the positive region down to In Table 1 the results based on the described
period (length) of 26 days the negative side indicates a descending empirical analysis are reported. Every table
Given the algorithm definition for which market. illustrates the total cumulative 11-year
n=12 and m=26 (where n indicates the return, the average yearly return and the
length of the shorter (faster) Vidya and m Tests and Empirical Results number of long and short trades, for both
indicates the length of the longer (slower) The empirical analysis has been performed MACD and the Dynamic MACD indicators.
Vidya), the parameters α and β can be to verify whether the Dynamic MACD is
computed by simple algebra: indeed more efficient than the MACD. The
analysis has been conducted by using daily
α= 2/ (n+1); with n=12 α=0.154 time series of prices over a time horizon
β= 2/ (m+1); with m=26 β=0.074 of eleven years from 1996 to 2006. In
particular, the time series used are from
Since Vidya is a dynamic indicator, the four market indices: China 1-2 Cp Index,
effective length of the average at any time Hang Seng Index, SMI, and DJIA; two
will not be constant, i.e. equal to 12 and exchange rate time series: the EUR-USD, As it can be inferred from Table 1 relative to
26. Instead, it will depend on the value and USD-JPY; and two stocks: Nokia and the tests performed on the exchange rate
of the “volatility” indicator. The effective IBM. On every analyzed time series the time series, Dynamic MACD return is better
length, at any time, indicated by the letters MACD and the Dynamic MACD have been with a much smaller number of trades;
N (shorter Vidya) and M (longer Vidya) is applied. In particular, the MACD has been similar results occur for the indices China
given by defined as the difference between the 12 1-2 Cp and SMI, with results summarized in
and 26-day exponential moving averages. It Table 2.
N = (2-k*α) / (k*α) for the 12 day Vidya. is assumed that for every “buy” signal (i.e.
M = (2- k*β) / (k*β) for the 26 day Vidya. MACD crossing from below the zero line) a
long position is entered, and for every “sell”
The improvement from the dynamic signal (i.e. MACD crossing from above the
averages is the ability to have different zero line) a short position is entered, and
lengths according to market conditions. then the return is calculated. A position
This differs from the Variable MACD,xiii is entered with the properties of stop and
where k was defined as the absolute reverse. For every closed position (trade)
value of the CMO. In this work a diverse the return is calculated. A cumulative Table 3 illustrates the test results for the
“volatility” indicator is being used, namely return value is then calculated for the Hang Seng and DJIA indices. For these two
the well known and widely used standard whole time horizon (eleven years). Then, a indices the Dynamic MACD yields a poorer
deviation. In particular, the 10-day similar analysis was performed by using the return relative to the MACD action, but
standard deviation of closing prices is Dynamic MACD. Again, the return per each the number of trades involved with the
calculated at each time t. The value is then trade was calculated and the cumulative Dynamic MACD is dramatically reduced
divided by a parameter defined by the 11-year return recorded. relative to the use of MACD. If transaction
average of the standard deviation over fees and costs were taken into account,
a 50-day period. K at time t is, therefore MACD return would have been lower than
defined as the ratio of the 10-day standard illustrated in Table 3.
deviation of the closing prices divided
by the 50-day average of the standard
deviation of the closing price.xii K outlines
the excess “volatility” relative to its Outstanding results are illustrated in
historical average value, at each time t. Table 4 (Nokia and IBM): a small number
To summarize, the Dynamic MACD’s of trades realize excellent returns for the
peculiarity is its adaptability and its ability Dynamic MACD.
to vary the velocity of the trading signals
according to the different market phases.
The averages become longer (and slower)
during trading range markets and become
shorter (faster) in trending markets.
The interpretation of the Dynamic
MACD, similarly to the MACD, consists
in analyzing the crossing of the indicator
values from negative values to positive
values and vice versa. Crossing from
below the zero line indicates an uptrend.

page 18 IFTA.org
 IFTA Journal 2008 edition

Table 1 EUR-USD USD-JPY

Cumulative Average Number Cumulative Average Number

Return Annual Return of Trades Return Annual Return of Trades
1996-2006 1996-2006
MACD -19.12% -2.10% 1434 19.55% 1.80% 1442
Dynamic MACD 39.46% 3.38% 86 60.69% 4.86% 114

China 1-2 Cp Index SMI

Table 2
Cumulative Average Number Cumulative Average Number
Return Annual Return of Trades Return Annual Return of Trades
1996-2006 1996-2006
MACD 44.51% 4.18% 1049 60.74% 4.86% 1396
Dynamic MACD 178.58% 12.06% 58 152.06% 9.69% 91

Table 3 Hang Seng Index DJIA

Cumulative Average Number of Cumulative Average Annual Number

Return Annual Return Trades Return Return of Trades
1996-2006 1996-2006
MACD 283.71% 14.39% 1384 15.07% 1.41% 1379
Dynamic MACD 112.48% 7.83% 99 -26.56% -3.04% 106

Table 4 Nokia IBM

Cumulative Average Number Cumulative Average Number

Return Annual Return of Trades Return Annual Return of Trades
1996-2006 1996-2006
MACD -65.33% -10.05% 1193 89.56% 6.60% 1241
Dynamic MACD 28.25% 2.52% 96 194.29% 11.40% 91

IFTA.org page 19
 IFTA Journal 2008 edition

The following charts illustrate some

examples of trading range markets, Figure 1. China 1-2 cp Index Time series
comparing the return realized from
MACD and Dynamic MACD, denoting
the importance of being able to manage
trading range periods in order to improve
overall return. Figure 1, China 1-2 Cp
Index, highlights the period October
2001-September 2004 as a trading range
period. In this specific period the Dynamic
MACD yields a return of 44.51% whereas
the MACD returns 20.48%. These results
indicate that within a trading range, the
Vidya are able to flatten more efficiently
relative to the exponential moving
averages, minimizing the number of false
signals.

Figure 2. IBM time series

In figure 2 a similar analysis is performed
on IBM. In this case, during the long trading
range period (June 2000- June 2004) the
Dynamic MACD produces a better return
relative to the MACD (56.40% and -70.89%
respectively).

Figure 3. DJIA time series

By analyzing the time series of the DJIA in
figure 3, it is easy to note a trading range
in the period from September 2003 to
December 2004. During this time period
the return from the MACD is -16.18%
whereas the Dynamic MACD produces a
positive result of +10.88%, which is very
peculiar. In fact, during the whole 11-year
period the Dynamic MACD return is worse
than the MACD, but during trading range
periods, the Dynamic MACD has performed
better than the MACD.

page 20 IFTA.org
 IFTA Journal 2008 edition

Conclusions and REFERENCES The term “trading range” defines those

xvii

i
Pring M. J., Guida completa all’analisi market dynamics in which prices move
Future Research grafica, Trading Library, Cinisello Balsamo. in short and frequent oscillations in both
The empirical result presented in this 2004 directions, without exhibiting a definite up
work outlines the differences between or down trend.
the use of the MACD and the Dynamic ii
Wilder W. J., New concepts in technical
MACD applied to various financial time trading systems, Trend Research, New York. xviii
The calculation is different from the
1978 statistical concept of volatility used in
series. The Dynamic MACD is constructed
finance and calculated by using standard
similarly to the MACD as the difference iii
Murphy J. J., Analisi tecnica dei mercati deviation of returns.
of two quantities. The MACD employs finanziari. Metodologie, applicazioni,
the difference between two exponential strategie operative, Hoepli, Milano. 1997
moving averages. The Dynamic MACD iv
Murphy J. J., Analisi tecnica intermarket:
embeds two Vidyas and calculates their
strategie di trading sui mercati finanziari
difference. The results, produced by the globali, Il Sole 24 ore libri, Milano.1998
tests over the period of eleven years,
confirm the efficacy of the Dynamic MACD v
Achelis S.B. Analisi tecnica dalla A alla Z,
yielding better returns in most tests and Trading Library, Cinisello Balsamo. 2004
with a considerably smaller number of vi
Coliva E., Galati L. Analisi tecnica
trades. This peculiarity becomes evident finanziaria, UTET, Torino. 1992
if transaction fees and costs are taken
into account. A limit of the Dynamic vii
Edward R., Maggee J. Technical analysis of
MACD is a slight delay making it lag the stock trends, Amacon, New York. 1996
MACD indicator during trending periods.
DeMark T. R., Market timing: tecniche
viii

Future research will involve studies and innovative per operare in sincronia con il
comparisons between Variable MACD and mercato, Trading Library, Cinisello Balsamo.
Dynamic MACD aiming to determine which 2003
of the two novel indicators would be a
better performer in both trading range and
ix
Fornasini A., Analisi tecnica dei mercati
finanziari e trading on line: metodi e sistemi
trending markets. Transaction fees and
a supporto delle decisioni operative, Etas,
costs need to be evaluated to better assess Milano. 2003
the practical implications for the money
management industry. IFTA x
Fuller R. J., Farrell J. L., Analisi degli
investimenti finanziari, McGraw-Hill, Milano.
1993

xi
Kaufman P., Smarter trading: improving
performance in changing markets,
McGraw-Hill, New York. 1995

Chande T. S., The new technical trader:

xii

boost your profit by plugging into the latest

indicators, John Wiley & Sons, New York.
1994

Gandolfi G., Rossolini M., Sabatini A.

xiii

(2006), “Variable MACD: adapting to

financial market dynamics”, STA’s Journal.

This concept of “volatility” is different

xiv

from the more commonly used definition in

statistics and financial terms.

xv
Pring M. J., Analisi tecnica dei mercati
finanziari, McGraw-Hill, Milano. 2003

Presented to the 2006 IFTA Conference in

xvi

Lugano (Switzerland) and the Variable MACD

has been programmed and implemented in
Bloomberg tools.

IFTA.org page 21
 IFTA Journal 2008 edition

An Exploration of the Nature of Bull Market Tops

by Paul F. Desmond

Almost every investor harbours the secret identified by evidence of panic selling (one our almost total lack of understanding
wish of being able to sell out on the exact or more 90% Downside Days) in which about the end of bull markets that is
top day of a bull market. The bragging investors dump stocks with abandon. responsible for investors’ almost universal
rights would last a lifetime. But, exactly Then, with the desire to sell having been inability to avoid bear markets. A greater
how does an investor identify the top exhausted, buyers suddenly rush in to snap understanding of investor psychology near
day? An easy answer might be that it up the bargains (and cover short positions), bull market tops might emit warning signs
is the highest level reached by the Dow resulting in a 90% Upside Day. The in the making, and allow at least some
Jones Industrial Average (DJIA) before a combination of panic selling across a broad alert investors to be able to take defensive
major market decline. This is probably spectrum of stocks, followed quickly by actions in advance of the devastating losses
a reasonably good answer for historians broad, enthusiastic buying, produces what that typically occur in the subsequent bear
studying the long-term trends of the might be described as a classic “V” pattern market.
stock market, but it is not a practical, of prices at major bear market bottoms. There are several helpful tools that
working answer for investors, as it can Bull market tops, on the other hand, Technical Analysts have used for many
only be known long after the top occurred. tend to develop gradually over a long decades to warn of impending stock
Another answer might be that the exact period of time. The reasons for this gradual market tops, such as the Advance-Decline
top of a bull market is the point at which process are easy to understand: It is the Line and the number of stocks recording
the vast majority of stocks reach their Law of Supply and Demand at work. Just as new 52-week Highs. History shows that
highest price levels for many years to bull markets result from strong, persistent these indicators often top out and begin
come. More than a few investors would investor demand for stocks, bull market to contract, as individual stocks fall by the
say that the first answer and the second tops evolve when investors gradually stop wayside, months in advance of the final top
answer are synonymous; that the majority buying. Some investors simply run out in the DJIA. Therefore, it would not be a
of stocks reach their peaks at the same of new money to invest. Others begin to surprise to find that all stocks do not reach
time as the peak of the DJIA. But, is that see individual stocks as being overvalued, their peaks simultaneously or in unison
actually the case? Do most stocks reach and begin to hold back on new purchases. with the DJIA. But, it is the degree and the
their price peaks in unison, and do they Whatever the reasons, the stock market intensity of the divergences of individual
do so simultaneously with the major price cannot continue to advance without stocks from the DJIA that had never been
indices? Does what seems so logical match Demand exceeding Supply. The evolution measured before —until now.
actual experience? of investor psychology from strong buying Discoveries in science are frequently the
There is a dearth of information about enthusiasm for stocks to passivity or result of happenstance rather than great
the nature of major stock market tops, complacency does not occur suddenly. scientific detective work. The discoveries
and the sparse information that does Thus, bull market tops are commonly to be related in this paper regarding bull
exist is more theoretical than statistical. diffuse, possibly lulling most investors into market tops began in exactly that fashion.
Stock market guru, Joseph Granville, once inaction. Perhaps it is the slowness of the My firm, Lowry Research Corporation, had
surmised that one-third of stocks reach entire process that makes it difficult to purchased rolls of microfilm of the Wall
their final bull market price peaks in recognize a bull market top. Street Journal covering the period from
advance of the DJIA’s peak, one-third reach However, beyond this vague and 1920 through 1930. Being able to step
their highs in unison with the DJIA’s peak, somewhat hypothetical supposition, back in time, if only in recorded history,
and one-third reach their peaks after the little or nothing more is known about the is a special experience. The first frame to
DJIA’s peak. However, the sheer simplicity nature of bull market tops. Despite our be viewed in the microfilm reader, purely
of Granville’s theory suggests that it was almost total lack of understanding of the out of curiosity, was the page containing
based more on guesswork than on hard subject, the end of a bull market and the the New York Stock Exchange trading of
statistical analysis. simultaneous start of a new bear market September 3, 1929—the absolute top day
One thing that investors have known, is undoubtedly one of the most important for the DJIA prior to the 1929 Crash. It
if only in a very vague sense, is that major moments in time for any investor. Many is ironic that 1929 is undoubtedly one of
market tops are not the same as major investors have experienced the frustration the most important dates in stock market
market bottoms. Much more work has been and anguish of making big stock market history, and so little is known about the
done in defining the nature of major stock gains in a bull market, only to watch forces of supply and demand at work in
market bottoms than in understanding the gains turn into big losses during the the market during that period.
the nature of bull market tops. A 2002 subsequent bear market. Thus, the ability In simply looking around at the trading
Lowry study titled Identifying Bear Market to avoid capital losses is, in many ways, data from that day—at the many unfamiliar
Bottoms and New Bull Markets, showed a more important objective for investors names of the companies traded, at the
that major market bottoms can often be than making big gains. Perhaps it is volume of trading, at the highest prices

page 22 IFTA.org
 IFTA Journal 2008 edition

for each stock—it became apparent that On the day on which the Dow Jones more from their 1929 highs. Thus it became
some stocks had traded that day at prices Industrial Average reached its absolute high apparent that the absolute top for the vast
below their 1929 highs. Some stocks were for the 1920s bull market, the percentage majority of stocks had probably occurred
considerably below their yearly high. That of stocks making new 1929 highs that day months—perhaps many months—before
seemed strange for a day on which the was not 80% or 75% or even 70%. It was September 3, 1929. And yet, there had
DJIA was at the absolute highest point in 2.30%. Out of 826 stocks that were traded been no single, outstanding day of rally
history and at a level that would not be seen on the New York Stock Exchange that day, prior to September 3 that investors could
again for the next 20 years. Upon closer only 19 stocks made their highs. Equally identify as the ideal point at which to shift
examination, it was difficult to find stocks surprising, only 15.62% of all issues traded portfolios to a more defensive composition.
that were at their highs on that fateful day. on the NYSE were either at, or within 2% of The pressing question was whether the
Intuitively, something seemed to be their 1929 highs. In other words, about 84% 1929 case was a total anomaly, or whether
very wrong. On a day when common sense of all stocks had topped out and had begun somewhat similar conditions would be
would dictate that most stocks should to decline at some time prior to September found at other important bull market
have closed at their all-time highs, it was 3. In fact, it was determined that, on the tops throughout history. Therefore, we
determined that very few stocks had closed same day that the DJIA reached its all-time expanded our study to include each of the
at, or even near, their 1929 highs. Many high, 31.84% of the stocks traded on the fourteen major bull market tops, based on
stocks were down from their highs by NYSE had already declined by 20% from the Dow Jones Industrial Average, from
20% or more (Last price was lower than their 1929 highs. 18.77% of stocks had 1929 through 2000 (table 2). Our basic
1929 High price). Thus began a detailed declined by more than 30%. Stocks at, or assumption was that most stocks reached
examination of the trading of September 3, within 2% of their highs were dwarfed by their highest prices in unison with the
1929. The results were most surprising. the number that had already lost 20% or Dow Jones Industrial Average. But, our

Figure 1.

Source: Wall Street Journal

Table 1. Examination of Trading on September 3, 1929.

BULL MKT TOP DAY % STOCKS AT NEW HIGHS % AT OR < 2% OF NEW HIGHS % OFF 20% OR MORE % OFF 30% OR MORE

09/03/1929 2.30% 15.62% 31.84% 18.77%

IFTA.org page 23
 IFTA Journal 2008 edition

examination of each stock traded on the Average. If a portfolio manager had Our study appears to show that the
New York Stock Exchange, comparing their somehow been able to sell out on the absolute Dow Jones Industrial Average is a less than
bull market highs to their closing prices on top day of the DJIA in each of the fourteen ideal proxy for the broad list of stocks. For
the peak days of the Dow Jones Industrial cases studied here, in most instances example, as shown in Table 3, in the 1929
Average, showed an unexpected picture. the portfolios would have already lost a case, none of the thirty component stocks
These findings defy the conventional considerable amount of value by that time. were making new highs along with the
wisdom about the nature of stock market Investors who may have thought themselves Industrial Average on September 3, 1929.
tops. In each case, 11% or less of stocks lucky enough to sell all of their stocks on the This is due to a large extent to the reporting
(average 5.98%) were making new highs exact top day of the DJIA could have actually of closing numbers for the Average on a
along with the new high in the DJIA—a suffered significant losses. The amazing theoretical basis.
generally accepted proxy for the broad list similarity of the statistics in these fourteen The study also suggests that, even
of stocks. Further, in nine of the 14 cases cases suggests a pattern of deterioration at that early time in the history of the
covered in this study, a significant number at major market tops that investors cannot 30-stock Average, the price weighting of
of NYSE-listed stocks (average 21.97%) afford to ignore. In searching for a way the components was producing an undue
had already dropped in price by 20% or to describe this phenomenon of market influence on the movements of the DJIA.
more before the DJIA had reached its bull deterioration—the gradual process of However, the bigger issue is that the
market peak. hundreds of individual stocks rolling over evidence drawn from all fourteen cases
The primary conclusion to be drawn from into their own bear markets, one by one, over suggests that the highest price levels
these fourteen cases is that the vast majority a period of many months—the picture of a for the vast majority of New York Stock
of stocks reached their bull market highs well feather emerged. We think that image is just Exchange listed stocks have tended to occur
before the peak of the Dow Jones Industrial about right (figure 2). well before the final peak in the DJIA.

Table 2. Examination of Trading at Fourteen Peaks in the Dow Jones Industrial Average.

BULL MKT TOP DAY % STOCKS AT NEW HIGHS % AT OR < 2% OF NEW HIGHS % OFF 20% OR MORE % OFF 30% OR MORE

09/03/1929 2.30% 15.62% 31.84% 18.77%

03/10/1937 6.05% 21.34% 5.94% 1.06%

05/29/1946 8.59% 30.44% 6.30% 0.86%

04/06/1956 5.32% 23.36% 1.92% 0.42%

01/05/1960 1.60% 5.83% 23.25% 7.67%

12/13/1961 3.56% 11.83% 25.29% 11.60%

02/09/1966 9.66% 19.04% 9.52% 2.68%

12/03/1968 9.43% 20.12% 9.51% 2.36%

01/11/1973 5.30% 11.82% 34.22% 20.51%

09/21/1976 10.97% 22.88% 21.65% 10.09%

04/27/1981 7.09% 15.18% 28.01% 9.39%

08/25/1987 6.23% 15.23% 17.37% 7.44%

07/16/1990 5.35% 18.11% 37.31% 22.74%

01/14/2000 3.54% 6.31% 55.33% 32.45%

AVERAGE 5.98% 16.88% 21.97% 10.54%

page 24 IFTA.org
 IFTA Journal 2008 edition

The final days of a bull market are Figure 2.

substantially different than the final days
of a bear market. At most bear market
lows, because fear and panic are the
dominant emotional drivers, the vast
majority of stocks tend to bottom in unison.
At most bull market tops, where investors
have been lulled into complacency, the vast
majority of stocks seem to top out on an
individual basis. This is not much different
than observing that a farmer usually plants
all of his seeds at the same time in the Table 3. The Dow Jones Industrial Average components as of
spring. However, not all of the fruit reaches September 3, 1929.
the point of peak ripeness at the same time.
The ripe fruit must be picked individually,
DJIA Components 1929 High Sept 3, 1929 Close
rather than all at once. In the same way,
investors must commit to buying stocks Allied Chemical 354 ¾ 354
quickly after a major market bottom, but American Can 184 ½ 181
must sell stocks one by one, as they reach
their individual peaks. American Smelting 129 1/8 128 1/8
This simple study of bull market tops American Sugar 94 ¾ 81 ¾
should have far-reaching implications for American Tobacco 205 200
all investors. The conventional wisdom of
what a major market top looks like must Atlantic Refining 77 /8
7
65 5/8
be completely revised. Every portfolio Bethlehem Steel 140 ¾ 136 ¾
manager must create a new strategic plan
Chrysler 135 71 7/8
as to how and when to take defensive
action. And, new indicators must be Curtis Wright 30 1/8 29
devised to eliminate the current guesswork General Electric 403 391
of where individual stocks are within
the primary trend. Investors must be General Foods 81 ¾ 71 7/8
able to see, and have time to react to, the General Motors 91 ¾ 71 ¾
gradual deterioration of market breadth General Railway Signal 126 ½ 123 ½
that precedes periods of substantial stock
market losses. Goodrich 105 ¾ 73
We will leave it to other researchers and International Harvester 142 140
analysts to determine all of the various
International Nickel 72 ¾ 54 ½
reasons why so few stocks have reached
their bull market highs in unison with Mack Truck 114 ¾ 97
the Dow Jones Industrial Average. Our Nash Motors 118 /8
7
84 5/8
principal concern, at this point, is to alert
investors to the conditions that have National Cash Register 148 ¾ 125 ¾
consistently occurred at important stock North American 186 ¾ 184 1/8
market tops. Future studies will address Paramount 74 72
the need to develop new indicators and a
new portfolio management strategy to deal Radio Corporation 114 98 1/8
with the challenging conditions revealed in Sears Roebuck 181 171
this study. IFTA
Standard Oil N. J 73 /8
7
70 ¾
Texas Corporation 71 ¾ 68 ½
Texas Gulf Sulphur 85 ¼ 72
Union Carbide 137 7/8 135 ¾
U. S. Steel 261 ¾ 257 5/8
Westinghouse 295 5/8 285 7/8
Woolworth 100 7/8 99

IFTA.org page 25
 IFTA Journal 2008 edition

Sentix: Behavioral Indices

A Behaviourally Oriented Development
of the TA Tool-Kit
by Manfred Hübner

Abstract and its actors. The homoeconomicus know this, but we have never been able to
Behavioral Finance is the theoretical prevailed and the efficient market state it in such a skilfully scientific way.
foundation of our discipline. The article hypothesis put technical analysis into Now, three main corner stones of a new,
shows, how Behavioral Finance and question. behaviorally oriented financial market
Technical Analysis are connected and This changed completely twenty theory have emerged:
why we demand better data for investor’s years ago. Since the crash of 1987, the • There are useable momentum effects on
sentiment. It introduces a new set of picture of the rational investor faltered the markets (prices move in trends)
sentiment indicators, the Sentix indices, considerably, and again more recently, • In specific situations, investors tend
which enables anyone, without costs, for example, after the meltdown of LTCM to over-react (over bought—over sold
to participate in a global project with hedge fund and the tech bubble burst conditions) or under-react (strengthening
sentiment data to various markets. Some of 2000. The idea of efficient markets of trends)
examples are shown how to use this new with rational investors could no longer • The perception of investors is a function
indicator set. be supported. This paradigm change is of the price, the news flow as well as the
accompanied by a rise and development of prevailing positioning (sentiment can
Behavioral Finance a new scientific area, behavioral finance. dominate fundamental factors)
as the Foundation of In contrast to the classical capital market
Technical Analysis theory, behavioral finance places the From Laboratory Experiments
For most market participants, technical individual and his (sometimes hardly to a Real Time Application
analysis is more of an art than a science. rational) decision making procedure in Most of these realizations have been gained
This is connected to many factors, the centre of attention. For technical in laboratory experiments or through
especially the mostly visual concepts which analysts, the realizations that can be taken tedious examinations and extensive data
are the basis of any technical analysis. The from here often read like a repetition material. An analysis in real-time seemed
nature of markets is also an important of basic convictions that have long impossible. Richard Thaler, one of the
factor. From the beginning, technical been internalised. And, nevertheless, pioneers of behavioral finance, pinpointed
analysis was the attempt by practitioners this scientific work is of fundamental the dilemma: “There are better data on
to get to grips with the nature of financial importance for our work. For the first time, prices than people!”1
markets. Be it the trend analysis (herd basic assumptions of technical analysis At this point, the strength of the
behaviour of market participants), the are reviewed scientifically and (partially) technical analysis and its focus on
Elliott Wave Theory (“Nature’s Law”) or corroborated. practical requirements becomes apparent.
Japanese Candlestick-Charts—all these A central concept of behavioral finance Especially for the analysis of prices,
concepts have one thing in common: they is for example the “prospect theory”, for technical analysis possesses a large pool
base their statements on the psychology which Daniel Kahneman received the of (more or less suitable) tools. Be it
of individuals and of the masses and Nobel-Prize for Economics in 2002. His trend following systems, oscillators or
derive their forecasts from the realization theory states that people experience losses the formation theory—we are armed.
that human behaviour repeats itself in much more than gains, and that we are Specifically, however, in the field of
certain situations and can, therefore, be more likely to behave loss-aversive than sentiment analysis with the consideration
prognosticated to a certain degree. risk-averse of psychological factors, technical
For a long time, this approach was In this context, reference points and analysis has a gaping void. While asset
considered unscientific, as Technical mostly the entry price, play a central role. prices and turnover can be considered to
Analysts, being practitioners themselves, Translated into the language of technical be standardized and comparable, there
having little interest in writing scientific analysts, this states simply that bear is nothing comparable concerning the
texts. It is also based heavily on the fact markets have a different dynamic than bull sentiment, the expectations and the
that at the very latest since the 1960s, the markets and that resistance and support (as positioning of investors. “There are better
prevalent capital market theory painted a significant acting points in the market) play data on prices than on people” is, therefore,
completely different picture of the markets a central role. As technicians, we already also a valid maxim for Technical Analysts.

page 26 IFTA.org
 IFTA Journal 2008 edition

Emotions Cannot be Measured, better, standardized information on successful as a conduit to collect sentiment
Effects Can sentiment, expectations and activities input. At present, over 2,500 investors
We can say that dispositions in the of actors on the finance markets while have made use of this opportunity. Among
proper sense of the word cannot be satisfying academic standards. An extensive the participants are more than 550
measured, as they arise from inside the survey panel was to be established which institutional investors, portfolio managers
individual. However, the effects, i.e. the would allow for comparisons between as well as analysts and economists. We
manifestation of emotions in activity, can investment groups, markets and different receive sentiment data from more than
be measured. Now, an effect can also be a time-periods. The information should be 20 countries, especially from Europe.
neglect or omission, and this is the reason available promptly to allow for a portfolio Addittionaly, Sentix is known and used in
why survey-based sentiment indicators regulation that oriented itself according to Asia, including Japan. Currently, more than
have an advantage over turnover-based the published data. 700 investors take part in Sentix surveys
indicators such as the put-call ratio for the every week.
visualization of emotions. Survey-based Introduction to the
indicators also mirror the sentiment of Sentix Indices Methods: How Sentix Works
investors on the sidelines, who could This idea gave birth to Sentix—behavioral Every Friday, participants receive a mail
create future supply and demand just like indices which, for the first time, presents with a link to our website, on which the
invested investors. On the other hand, a comprehensive sentiment picture of a survey is then conducted. Standardized
most price- and turnover- based concepts large number of investors and allows for questions include questions concerning the
are easier to standardize. With most a comparison of sentiments of different short- and medium- term expectations for
capital markets surveys, one cannot be investor groups and nationalities, as twelve markets; for equity markets as well
sure who is participating and what their well as different markets. Without the as bond-, FX-, and commodities markets.
underlying motivation is. Also, the results, internet, such a project would have been In this context, short-term defines a time
which are usually directed towards very impossible to realize. The combination of horizon of one month, while a six month
specific questions or markets, are difficult e-mail and internet based surveys gave horizon is set for the medium-term view.
to compare due to a lack of transparency us the opportunity to collect the answers The answers are calculated anonymously
in the survey modalities. For this reason, from thousands of people in a very short and turned into indices. In addition to the
inter-market statements tend to be nigh time-span, to process this information and standard questions, participants are given
impossible. to make the results available. The central topics that change on a weekly basis, such
These preliminary thoughts were the node of the project is the Sentix website as preferred investment styles, sector
basis of a new, behaviorally oriented (http://www.sentix.de), where anyone assessments or positioning. Figure 1 gives
concept in the year 2001. This concept can register free of charge to participate a schematic overview of the more than 400
came from commercial practice and was in the surveys and the data. Since its individual indicators in the Sentix family.
conceived for the practitioner to receive introduction, it has been extremely

Figure 1. Overview of the

sentix indicator family

IFTA.org page 27
 IFTA Journal 2008 edition

The Indicators Figure 2. Sentix-Sentiment for the DAX Index (short term)
In the following, we would like to introduce
some of the Sentix indicators and to point
out appliance possibilities. We would like
to start with the mother of all sentiment
indices, the Sentix Sentiment Index. This
displays investor sentiment in the form
of a classic bull-bear index. Participants
can give one of four possible answers:
bullish, neutral, bearish and no opinion.
The index is then computed from the
balance of bullish minus bearish answers
in proportion to the number of answers
(excluding no opinion).

sentix = B
ulls - Bears
All Votes

Should there be, for example, for a short-

term assessment of the DAX-index, 120
bullish, 80 bearish and 45 neutral answers,
the sentiment index would result in +16.32%.
The Sentix Sentiment Indices are, therefore,
valued between -100% and +100%.
Figure 2 shows an example of the short- Figure 3. Short-term DAX-sentiment as Z-Score-Index
term sentix sentiment for the DAX index.
The index behaves in the same way as
other classical sentiment indices. It
can be seen that investor sentiment is
strongly influenced by the development
of market prices. However, in comparison
to “normal” technical indicators, these
oscillators possess one very significant
advantage: they adapt themselves
automatically to the respective market
dynamic. A variation of the parameters to
adjust to the specific market surroundings
is not necessary. A further result of
our research is that an upper trend
turnaround differs significantly from a
lower one. Market bottoms usually come
about against the background of a very
bearish sentiment, while bull markets
hardly ever die in euphoria. Mostly, upper
turnarounds follow precursory sentiment
divergences which—just as with technical
indicators—don’t support new price highs this case, the sentiment divergences as soon. Often, activity with a trend which has
with accordingly bullish sentiment. The well as the lower extremes are even more developed from a gradual transformation
bull camp, therefore, must fall apart easily recognizable. form bear to bull, is more promising.
“sentimentally” before a trend turnaround As we stated earlier, we survey market
is probable. This is also confirmed by Different Behaviour of expectations on a short and medium-
Sentix Sentiment data, namely that bull Sentiment in Bull Markets and term basis, and we also divide according
and bear markets are structurally different. Bear Markets to individual and institutional investors.
It is also interesting to smooth out This assessment of the market sentiment Interestingly, medium-term investor
indices, to statistically standardize them as shows us that activity against the assessment shows completely different
so-called Z-scores. Figure 3 shows a short- predominant sentiment must be well patterns than short-term assessments.
term Sentix sentiment for the DAX which considered. In a bull market movement, A study conducted by the University
has been standardized in this fashion. In one should not go against the herd too Maastricht2 showed that medium-term

page 28 IFTA.org
 IFTA Journal 2008 edition

expectations of those investors surveyed Neutrality Index. This index represents tends to go down. The opposite is the
by Sentix are suitable as a basis for the quota of neutrally positioned case when irritation (high neutrality) is
forecasts. While short-term sentiment investors over time. A high neutrality predominant amongst investors. In the
behaves like a price-based oscillator, index means that many investors have no aftermath, high absolute value changes are
medium-term expectations better mirror clear view on the market. One could also probable, as well as rising volatility.
the assessment of the market. Institutions call this situation an “irritation”. A low Table 1 shows the statistical processing
especially show clear anti-cyclical quota of neutrally positioned investors of future absolute oscillations of the Bund
tendencies. Professionals demonstrate means that almost everyone has a clear Future depending on the value of the Sentix
a relatively sound feeling for overpriced opinion. This could point to a situation Neutrality Index. “No. of. occ.” means
or underpriced markets. Therefore, of “overconfidence”. Figure 4 shows the the number of observations in the Sentix
Sentix data does not only deliver insights short-term Neutrality Index for the Euro- database.
into market sentiment, but also into the Bund-Future (Z-score index).
valuation assessment of securities from Turn-around points on this index in the European Sector Sentiment
the viewpoint of investors. vicinity of +/- 1.5 standard deviations are Allows Sentiment Arithmetic
regularly accompanied by turn-around A significant advantage of the Sentix
Neutrality Counts: How to points in the market. It is more interesting indicator family is that data from a wide
Profit from Uncertainty to note, however, that after a phase of low range of markets can be surveyed with a
In the following, we introduce two other neutrality, only small changes in absolute comparable underlying set of principles
indicators. The first is the so-called Sentix price value tend to follow and that volatility and with the same survey clients. Once
a month, investors are asked for their
Figure 4. Medium-Term Neutrality Index for the assessment of a range of equity sectors
Euro-Bund-Future (Z-Score) (18 sectors; STOXX system). As currently
the majority of survey participants are
based in Europe, this is a European sector
sentiment. Every equity sector can be rated
on five levels from strongly over average
(++) to strongly under average (–). For
the calculation of the index, the answers
are average-adjusted and statistically
normalized as Z-scores. Figure 5 shows,
as an example, the relative sentiment for
European telecom values in comparison to
the relative performance of this sector to
the overall market.
Comparable sentiment data for equity
sectors did not exist until now; the utility,
however, is evident. Extreme readings in the
sector sentiment are a precursor for future
relative price movements in the sector.
Sometimes, as a group, investors do
not accept a dominant price trend. This

Table 1. Medium-Term Neutrality Index and future Bund-Future price movements

IFTA.org page 29
 IFTA Journal 2008 edition

Figure 5. Telecom sector sentiment and relative Performance of Telecoms versus STOXX600

Figure 6. Pharma sentiment stays bullish and price declines go beyond “fair values”

page 30 IFTA.org
 IFTA Journal 2008 edition

happens, for example, if the valuation case discussion of the all time high in the REFERENCES
seems to favour the sector. The price and DAX Index. As expected, the medium- 1
Richard H. Thaler, “Advances in Behavioral
sentiment development in the European term sentiment declines near the top of Finance”, Preface, page XV, Russell Sage
Foundation, 1993.
pharmaceuticals sector is such an example. 2000. That shows that the imagination
Figure 6 illustrates the relative weakness of of investors did not go beyond that point. 2
Marco van Daele, “Investors Sentiment,
the sector compared to the bullish sentiment, A higher willingness to take profits was Stock Returns and Volatility in Germany”,
which dominates for a relatively long time. therefore expected .3 And indeed, as official University of Maastricht, 2005.
As all Sentix indices are surveyed on flow statistics showed, investors behaved in 3
Manfred Hübner, “Die nachgelieferte Mai-
a comparable basis, one can also work a loss-averse manner and pulled the money Schwäche”, Börsen-Zeitung vom 30.05.2007
with them and, for example, determine from the table. It’s clear that as soon as new
the relative sentiment between telecom highs are confirmed, regret will dominate Bibliography
and energy values. Figure 7 shows the investor’s feelings and people will have to 1. Dorsey, Woody, “Behavioral Trading”,
difference in sentiment between both buy into their underinvested portfolios. Thomson Texere, 2004
sectors as well as relative performance. 2. Schmehling, Maik, “Institutional and
Here, one can see the completely new Sentix: A Unique Project with individual sentiment: Smart Money and Noise
appliance functions for which this Global Reach Trader risk”, Discussion Paper No. 337, ISSN
data allows. The Sentix indices are a unique project to 0949-9962, 2006
One could also combine the sector survey investor sentiments and investor
3. Thaler, Richard H., “Advances in
sentiment with the Crude Oil sentiment, activities. They possess the potential to Behavioral Finance”, Russell Sage
The bullishness for Oil in the summer of give “better data on people”. Without Foundation, 1993
2006 compares nicely with the strong appropriate data, which also needs to
bullishness for Energy stocks and gives be available in real time, a systematic 4. Van Daele, Marco, “Investors Sentiment,
additional hints on a coming weakness in implementation of the insights of Stock Returns and Volatility in Germany”,
Thesis paper University of Maastricht, 2005.
the relative performance of BP & Co. behavioral finance would not have been
possible. For technical analysts, the Sentix 5. http://www.sentix.de/index.php?
Practical Considerations indices are an important supplement to pagename=knowhow.htm
What should one take into consideration our tool box. Every investor can actively
when using the Sentix data in practice? participate in the surveys and get free access
First, and especially, short-term indicators to the results. If many investors from
may be used as classical sentiment around the world take part in this project,
indicators. However, the analyst should we can gain a completely new global
use caution with the assumption that a perspective. IFTA
bull market dies in a phase of optimism.
We have regularly observed the break-
up of bull camps before the actual Figure 7. Relative sentiment—Energy sector vs. Telecom sector
course summit. Medium-term investor and the relative performance between the two sectors
expectations also fall backward long before
a top is achieved. In a bear market, it is
often worth grabbing the falling knife with
conviction and confidence in the face of
highly pessimistic values. Bear markets are
simply different.
It is always important for us to inspect
sentiment data according to which
psychological state they represent. If
behavioral finance theory defines cognitive
dissonance to be events going other than
what was expected, then this corresponds
with a falling market with a high level
of investment positions. In this case,
the theory allows the expectation that
investors will use suppression methods.
This, in turn, can be reviewed with the help
of Sentix indicators in the fundamental
group, which measures those topics which
are favoured by investors.
Another example of how to spot
behavioral anomalies in real time is the

IFTA.org page 31
 IFTA Journal 2008 edition

Using A Color Spectrum To Represent Changes In

The WAKO Volume Ratio On Candlestick Charts
by Stewart Gault, MFTA

Introduction and Abstract other hand, the daily or weekly volume is

Recently an increasing number of articles defined as accumulative when the closing
have claimed volume and indicators price at the end of the period is lower than
derived from it are underutilized. The fact the opening price at the beginning of the
that volume indicators are independent period. Variations in the magnitude of the
and often leading evidence of a potential energy and the actual level of the WVR
price trend reversal is well documented. estimate the market phase. Rising prices
Disregarding volume data seems illogical represent energy being distributed into the
when integrating it into a strategy is not a market, whereas falling prices represent
difficult task. Like price indicators, volume energy being stored in the market.
indicators come in many different forms
from indicators which only input volume, Calculation
such as Joe Granville’s On Balance Volume The general equation to calculate the WVR
and Volume Moving Average, to indicators is as follows:
that combine both price and volume, for
example the Money Flow Index and the Total volume of the periods in Total volume of the periods in
Volume Weighted MACD. which prices advanced - which prices fell or levelled off
This paper will introduce a still- Wako (Distributive Envergy) (Accumulative Envergy)
unfamiliar volume indicator, originating Volume =
in Japan, the WAKO Volume Ratio. After Ratio(%) x 100
focusing on three factors that affect this Total volume of all periods
indicator, an original charting method that
combines price and indicator data on a
single chart will be introduced and applied The method of calculating the following energy to a level where the market can no
to the WAKO Volume Ratio. If you are not weeks is the same as that used for longer sustain more distributive energy.
using a volume indicator, are searching for calculating the value of a moving average: The WVR then enters a phase of decreasing
another indicator to test and possibly add delete the oldest week’s total volume from distributive energy, and prices decrease.
to your strategy, or are looking for a unique the denominator and add the newest week’s Prices usually hit a top towards the end
new charting method, you may find this total volume. For the numerator, add or of the phase of increasing distributive
article informative. subtract the oldest week’s period volume if energy, or early in the phase of decreasing
the oldest week’s energy was accumulative distributive energy. As the WVR passes
Background: or distributive respectively, then add or over the zero reference line, a phase of
WAKO Volume Ratio subtract the newest week’s volume if it is increasing accumulative energy begins.
The WAKO Volume Ratio (WVR) represents distributive or accumulative respectively. When the market can no longer store any
the difference between distributive volume For easier comparison with price more energy, prices start to rise and a
and accumulative volume as a percentage movements, WAKO suggested plotting the phase of decreasing accumulative energy
of the total volume for a defined period. In WVR, with a range from –100% to +100% begins. It is around this time that prices
1974, Yukiharu Abe who worked at WAKO and centered on a zero reference line, on normally hit a bottom.
Securities Co. (merged with Shin Nihon top of the price chart (see figure 1). Abe proposed the following six rules
Securities in April 2000 to form Shinko when using his WVR with stocks:
Securities) developed the ratio as a timing Interpretation and Rules
mechanism for the medium and long-term As the WVR generally oscillates around the 1. When prices are in the low or high price
trading of stocks. It is based on the theory zero reference line, identifying the WVR zone, as seen from a medium or long-term
that price movements are formed by a cycle level and phase is essential. The WVR cycle, perspective, the WVR hits its significant
of distributive and accumulative energy. and hence the market, can be broken down bottom or top.
The total daily or weekly volume is defined into phases as shown in figure 2.
as distributive when the closing price at the When the distributive and accumulative 2. Although the depth of the bottom will
end of the period is higher than the opening energy are equal the WVR will be zero. As differ depending on the stock, generally
price at the beginning of the period. On the prices increase, so too will the distributive when the WVR reaches –40% to –60%,

page 32 IFTA.org
 IFTA Journal 2008 edition

prices stop falling indicating a low price Table 1: Calculation of a 14-day WVR
zone. More popular stocks however tend
to have a shallower bottom, around –20%
to –30%. Based on previous bottom levels,
one can buy if the WVR has increased
from below –20% for two consecutive
weeks (13-week WVR) or when the WVR
has stayed in negative territory for several
Col. E:
periods and shows signs of a reversal. if Open , Close
then Energy Type = + (Distrib.)
else Energy Type = - (Accum.)
3. When the WVR increases to around
Col. F:
+60%, the rising price movement slows sum of + energy volume during
down, indicating a high price zone. WVR 14-day period

levels exceeding +80% signify a high Col. G:

sum of – energy volume during
probability of prices falling. Before selling, 14-day period
one must make a distinction between
col. H:
this rule and Rule 4, otherwise, the stock sum of all volume during 14-day period
could be sold too early. High WVR levels, Col. I:
above +60%, should be taken as a warning F-G) / H x 100

sign that a reversal may be imminent.

One may consider selling if the WVR
has decreased from above +60% for two
consecutive weeks (13-week WVR) or
prices have crossed under a short-term
moving average. For short-term buying
opportunities, look for stocks whose WVR
exceed +40% and continue increasing.
The risk of buying a top is significant, so
sound money management strategies are
essential.

4. In the case where the WVR increases

rapidly in a single period due to a sudden
increase in volume, Rule 3 does not apply.
In fact, this indicates the start of an
advancing market, and a possible buying
opportunity.

5. A bullish signal forms when the WVR

follows a declining trend and prices are
either steady or increasing. Subsequently, Figure 1: NAB (daily) with 14-day WVR overlaid
prices start rising or continue rising
for a while indicating a possible buying
opportunity.

6. When the WVR moves from minus(-)

to plus(+) and prices are trending higher
from an earlier bottom, prices continue to
rise further. Contrary to this, stocks where
the low price can be determined from past
WVR levels, scale-down buying can be
performed. Buy the first lot when the WVR
crosses from positive to minus, a second lot
when the WVR falls below –20% and a final
when below –40%.

Usage Today
Other than in the English version of the

IFTA.org page 33
 IFTA Journal 2008 edition

Figure 2. WVR Cycle and Phases Nippon Technical Analysis Association’s

(NTAA) book titled Analysis of Stock Prices
in Japan, finding English literature on the
WVR is very difficult. Determining the
usage of WVR amongst analysts in Japan is
not easy but likely to be limited. In addition,
research does not reveal widespread use of
the WVR outside Japan. A small number of
books, in Japanese, cover the indicator and
a few Japanese technical analysis software
programs include the indicator. In the past,
methods like Candlesticks and Ichimoku
Equilibrium, took some time to reach, and
be adopted by, analysts outside Japan, and
the WVR will hopefully add to this list of
useful indicators reaching the West.

Figure 3. Telstra (daily) with 10-day WVR highlighting rules WAKO Volume Ratio:
Factors to Consider
Time Frame: WVR was originally
developed with medium to long-term
investment in mind, however as shown
here, the WVR can successfully be applied
to shorter time frames as well. Figure
4 compares six different WVRs from 10
days to 20 days. All WVRs respond to
significant price moves, late November
2002 (+60%) and mid Dec 2002 (-80%),
and do not deviate too much from each
other. The short WVRs (10 and 12-day)
respond faster to changes in energy
and reach more extreme levels in both
distributive and accumulative energy,
however with additional noise.
The longer 18 and 20-day WVRs
generally respond more slowly, but in some
cases (see Figure 5) will reverse on the same
day as the short WVRs. The WVR curve
Figure 4: Multiple WVR periods for Nikkei 225 Index appears smoother albeit more attenuated.
From mid-January to June 2003, the WVR
oscillated in a narrower range, between
–40% and +40%, which suggests a possible
change in market direction.
A closer examination of the short 10-day
WVR and the longer 20-day WVR shows the
attenuated longer WVR occasionally (late
November and mid-January) picking up
reversals quickly, but at other times (mid-
October and mid-December) not indicating
the change from increasing accumulative to
decreasing accumulative energy until seven
days later.
During the 90’s bull market, the short
10-week WVR crosses the +60% line eight
times, but only crosses the – 40% line
twice. The September 1998 correction
generated the first significant crossing of
the –40% line in more than six years. For

page 34 IFTA.org
 IFTA Journal 2008 edition

the following 34 months, the short WVR Figure 5. Nikkei 225 Index comparing 10 and 20-day WVRs
crossed the +60% line less frequently,
and accumulative energy increased.
Interestingly, during this period the bull
market slowed and later reversed.
As with any other indicator, back testing
and optimization of the WVR are necessary
to determine the appropriate period for a
balance between too many buy/sell signals
and not enough. Periods of 10, 14 and 20
days or weeks all seem to produce useful
signals, which should allow easy integration
with other indicators like RSI and Bollinger
Bands.
Market Direction: While studying
many charts, the influence of the market’s
trend on the WVR became apparent.
During a sideways market the energy
equilibrium point is at or very close to the
zero reference line. As the market trends
up or down, the equilibrium line moves
up or down producing higher highs and
lower lows respectively in the WVR. These Figure 6. Dow Jones Industrial Index comparing 10 and 26-week WVRs
observations allow technicians to adjust
their rules accordingly and take advantage
of important moves.
During the S&P 500’s short downtrend
(Figure 7), the 10-day WVR spent 60
days above and 68 days below the zero
reference line. Also, the 10-day WVR spent
one day above the +60% line, but 25 days
below the – 40% line.
Throughout the S&P 500’s short
sideways trend (Figure 8), the 10-day WVR
spent 61 days and 62 days above and below
the zero reference line respectively. Also,
the 10-day WVR spent 2 days above the
+60% line, and 3 days below the –40% line.
During the S&P 500’s short-term
uptrend (Figure 9), the 10-day WVR spent
73 days and 55 days above and below
the zero reference line respectively.
Surprisingly, even though the WVR crossed
above the +60% line three times and below of 17,609 and a market capitalization
the –40% line once, the 10-day WVR of US$95.62 million. Although a strong
remained a greater time (10 days) below downtrend is not evident, the accumulative
the –40% line than above the +60% line energy considerably exceeds the
(9 days), which could suggest the current distributive energy. In late March, ATX
uptrend was slowing. closed lower for twelve consecutive days,
Low Volume: Even though the WVR which resulted in the 10-day WVR hitting
calculation normalizes volume, very low the –100% level for three days. Such a low
or widely varying volume may reduce the level indicates a very high probability that
correlation between price and WVR. Two price will increase, which it did. The same
stocks each with relatively low daily volume situation occurred again in late July, and
and small market capitalization were again the price increased.
chosen to study this phenomenon. Skyline Corp (NYSE:SKY, figure 11) had
AT Cross (AMEX:ATX, figure 10) had an an average daily volume of 8,671 for the
average daily volume for the period shown period shown and a market capitalization

IFTA.org page 35
 IFTA Journal 2008 edition

Figure 7. S&P 500 (GSPC) Downtrend

Figure 8: S&P 500 (GSPC) Sideways Trend

Figure 9. S&P 500 (GSPC) Uptrend

page 36 IFTA.org
 IFTA Journal 2008 edition

Figure 10. AT Cross (AMEX:ATX) bar (or candlestick). Charts have changed
significantly from the early 1800’s when
Hoshi-feet (Asterisk) were plotted in a
time series to represent rice prices in
Japan’s Dojima rice market. Today, analysts
employ various indicators based on price
and volume in an attempt to achieve the
most out of their trading strategies. The
increasing number of sub-charts, and
their decreasing size, tends to reduce their
significance, if only physically. Can these
indicators, correlated or uncorrelated, be
combined with price data and displayed
at the same level on the same chart?
Many fields, like finance, electronics
and aeronautics, use numerical analysis
techniques such as FDM, FEM and BEM to
display results in 3D using a range of colors,
which highlight areas of significance.
Representing standard OHLCV data in this
Figure 11. Skyline Corp (NYSE:SKY) 3D form may be unnecessarily computer
intensive; still the concept of using color
variations to represent changes in indicator
values seems practical. Color by itself
can add another dimension to the price
chart without a significant increase in
computation time. Candlesticks, which
have a larger physical (body) area than bar
and line charts, will be used to clarify the
colors.
It should be noted that even though the
concepts presented here apply to other
volume and price indicators, this report
focuses on WVR indicator applications.

Similar Concepts
The use of color on charts to represent
rising or falling prices is hardly new. In
fact, Japanese colored upticks on Ikari-ashi
(Anchor Foot) charts were used around
1900 to distinguish from downticks, and
of US$277.74 million. Interestingly, in this between a fast and slow average was not long after candlesticks with colored
chart the short 10-day WVR generates some applied to the WVR, subtracting a 20-day bodies became popular because of their
useful signals (mid-February, mid-April and WVR from a 10-day WVR (Figure 12). increased clarity and ease of use. Today’s
late July), but the smoothed 26-day WVR Unfortunately, this idea did not yield usage of green and red candlesticks,
also reaches the +60% and – 40% lines at results any better than normal WVR usage a common combination, to represent
the same time without as much noise. and was therefore not pursued. The noise rising and falling prices is useful for
WVR can generate some reliable signals on both WVRs, which could be reduced quickly identifying a large number of
when used with low-volume stocks; by additional smoothing, contributes to consecutive up or down periods in addition
however increased volatility must be taken the volatility and false signalling of the to identifying other market conditions.
into account. A longer WVR can avoid difference curve. Yet the limited number of colors does not
some of the false signals, but may not show subtle changes in average price,
reach the trigger lines, so adjust the rules Spectral Candlesticks momentum or sentiment. Most popular
accordingly. In basic chart construction, time and price charting programs allow users to define
are represented on the x-axis and y-axis basic color settings, however research has
Modification of the WAKO of a 2D graph respectively, and volume is not uncovered any usage of color spectra
Volume Ratio recorded as a histogram at the bottom of like that presented here.
The MACD concept of finding the difference the chart under its corresponding price

IFTA.org page 37
 IFTA Journal 2008 edition

Figure 12. BHP Billiton (ASX:BHP) with a 10 and 20-day represents the red, green and blue values,
WVR difference and each row, the 201 colors (see figure 15).
Due to the length of the matrix, only the
first and last few values are shown.
The charting program first calculates
the WVR value for the period then finds
a color in the spectrum matrix closest to
the indicator value (Figure 16). To simplify
programming, the WVR values were
calculated in Excel and imported into
MATLAB. Assuming the indicator value is
+50.2%, since we selected 201 points the
program rounds this to the nearest integer
(+50%), which equates to a RGB value from
the above table of [1.000, 0.5000, 0] or
orange.
The MATLAB code in Figure 17 plots the
candlestick in two stages; firstly the wick,
and secondly the body. The conditional
statement decides whether the candle has
a clear or colored body before drawing the
candlestick on the screen.
Calculation to other sensory thresholds, such as line
For stocks daily volume can range from zero length and mass. Coloring Candlestick Charts
to the number of stock issued. However, The color on the left (135) and the one with WAKO Volume Ratio-
depending on several factors, such as the to its right (134) may look very similar or Based Spectra.
size of the company, industry and number even the same. They are, however, slightly Figure 18 is the same as Figure 3, however
of stock issued, volume varies significantly. different. Distinguishing the colors on the the WVR has been converted into a color
Using a normalized volume indicator such right (130 and 122), which have a greater spectrum and the candlestick colors
as the WVR, simplifies the calculation. difference, is much easier (Figure 13). Using changed accordingly. On January 28, 1999
Plotting this was performed in MATLAB, a thousands of colors with unrecognizable (WVR = +78.42%) and November 22, 1999
technical computing program. differences would provide little extra benefit (WVR = +100%) dark red indicates a very
After determining the indicator’s range, to the analyst and would only increase high level of distributive energy, and
the number of colors in the spectrum used computation time. For the WVR, which according to Rule 3, suggests a possible
to represent the indicator can be chosen. ranges from - 100% to +100%, 201 points trend reversal that actually occurred
The number of colors the human eye can (or one for each integer between –100 and several days later. The extreme low on
distinguish is finite and ranges from one +100) seems to be a logical choice. February 15, 1999 (WVR= -100%) indicated
to several million depending on light Many computer programs use a RGB by dark blue did not see the prices increase
conditions. Due to technical limitations, (Red Green Blue) color code to specify significantly, however the May 11, 1999 and
computer and television screens cannot the intensities of the red, green and blue October 19, 1999 lows saw prices increase
produce all visible colors, especially most components that make up color. Depending to previous highs.
saturated colors. In addition, the amount of on the application, the value used to specify A possible price decrease in the Dow
light falling on the screen can also reduce the intensity varies with common ranges Jones Industrial Index (Figure 19) was
the number of colors we can discriminate. between 0 and 1, and 0 and 255 (an 8 bit indicated by the 10-week WVR crossing
The minimum difference in magnitude number). The highest value (1 or 255) the +60% line in the week of March 16,
between two colors a person can perceive specifies the highest intensity of red, green
is defined as the difference threshold, or or blue, and the lowest value (0) specifies
‘just noticeable difference’. For example, the lowest intensity of the corresponding Figure 13. Weber’s Law Applied
on a screen with uniform luminance L, if color. For example, [0 0 0] represents to Colors
the luminance in an area is increased to black, the minimum intensity of each color,
L+D, then the smallest detectable increase while [255 255 255] represents white, the
D is the ‘just noticeable difference’ at L. maximum intensity of each color.
A 19th century German experimental The analyst can define the color range and
physiologist, Ernst Weber, found that even highlight levels of interest, such as +60%
the size of the difference threshold is and –40% for the WVR. Figure 14 shows
a constant proportion of the original some examples of user defined spectra.
stimulus magnitude. This relationship, In MATLAB the spectrum is defined by
known as Weber’s Law, can be applied a 3-by-201 matrix, in which each column

page 38 IFTA.org
 IFTA Journal 2008 edition

1998 and again two weeks later. Although Figure 14. Possible Spectra
not immediately, prices fell to a low, as
indicated by the solid dark blue candlestick
on September 14, 1998, and then quickly
rose over 1000 points. This increase
slowed after distributive energy peaked at
+60.30% on November 23, 1998.
The 11-month daily chart of Boeing
(Figure 20) shows distributive energy
reaching high levels in early June (WVR
= +69.04%) and early September (WVR
= +84.46%), which are soon followed by
downturns. The January 29 low of –61.13%
did not signal the expected price rally,
however the ten days from February 28
remaining below – 40% was an early
indication prices would rise.
This daily chart (Figure 21) uses a
different spectrum from the earlier charts.
Three white bands in the spectrum, +57
to +63%, -3 to +3% and –43 to -37%, were Figure 15. Defining the spectrum in MATLAB
used to highlight the important +60%,
0 and –40% levels. When the WVR fell
quickly after July 8 it skipped the –40%
band, because -44.06% on July 15 was just
outside the band’s width. The subsequent
rise showed a white candle on August 11 at
the zero level and two at +60% on August
18 and 19. For shorter, more volatile WVRs
consider using wider bands (~10%) or else a
smoother, longer WVR.

Conclusions and Further Study

Although the WAKO Volume Ratio provides
some practical signals in all time frames for
large and small stocks, using it to confirm
signals with other indicators, especially
price indicators, is highly recommended.
The WVR is relatively simple to program Figure 16. Finding the closest color
and interpret, so integrating it into a
trading system for back testing should be
easy. Another modification of the WVR
that will be explored is a weighted WVR for
longer periods in an attempt to increase
response time but not noise.
The spectral candlestick method, though
visually appealing, does not instantly
provide quantitative information usually
required for trading decisions. The full
range of applications has not yet been
explored. At this point the method seems
most useful when quickly scanning large
numbers of stocks. A major drawback
is programming this method. Technical
software like MATLAB can perform the task
but not as effortlessly as most analysts may
require. A future task will be to develop a
user-friendly plug-in to Excel or a stand-

IFTA.org page 39
 IFTA Journal 2008 edition

alone C program. Another area of interest Figure 17. Plotting spectral candlesticks
is combining a price indicator and the WVR
indicator, into a single indicator then using
that as the base for spectral candlesticks to
quickly identify potential investments.

Acknowledgements. There are two people

who deserve mention for providing
guidance and support while completing this
article; Hiroshi Okamoto, who introduced
the WAKO Volume Ratio and kindly lent me
several books, and Yukiharu Abe, the WVR
developer who answered more specific
questions about his indicator. IFTA

Figure 18. Telstra (in AUD) with a 10-day spectral WVR

Figure 19. Dow Jones Industrial Index (DJI) with a 10-week

spectral WVR

page 40 IFTA.org
 IFTA Journal 2008 edition

BIBLOGRAPHY
Figure 20. Boeing (NYSE:BA) with a 10-day spectral WVR 1. Achelis, S.B., Technical Analysis from A to Z,
2nd Edition, McGraw-Hill, 2000.

2. Bollinger, J., (2002), A Survey of Volume

Indicators, IFTA Journal, 2002 Edition.

3. Hull, J.C., Options, Futures and Other

Derivates, 5th Edition, Pearson Education,
Saddle River, New Jersey, 2003.

4. Murphy, J.J., The Technical Analysis of

Financial Markets, New York Institute of
Finance, Paramus, New Jersey, 1999.

5. Okamoto, H., Practical Utilization of

Japanese Technical Charts—Overview of
Current Major Stock Market by Candlesticks
and Triangular Charts, Paris Technical
Analysis Forum, 2003.

6. Pring, M.J., Technical Analysis Explained,

3rd Edition, McGraw-Hill, 1991.

7. Color Discrimination, (n.d.), ERGO/

GERO Human Factors Science Homepage,
Retrieved September 3, 2003, from http://
www.ergogero.com/FAQ/Part2/cfaqPart2.
html

8. Weber’s Law of Just Noticeable Differences,

(n.d.), University of South Dakota, Internet
Figure 21. AT Cross (AMEX:ATX) with a 10-day spectral WVR Sensation & Perception Laboratory,
Retrieved September3, 2003, from http://
www.usd.edu/psyc301/WebersLaw.htm

9. History, (2002), Shinko Securities Home

Page, Retrieved August 6, 2003, from http://
www.shinkoseco.jp/english/hist.html

10. The MathWorks Inc., (2002), Using

MATLAB—Version 6.

IFTA.org page 41
 IFTA Journal 2008 edition

How Well do Traditional Momentum Indicators Work?

by Cynthia A. Kase, MFTA

Introduction took place. This was done by finding all the average stop that was hit was exceeded or
Most market technicians believe traditional instances of the average stop being hit. To moving averages (the 10 and 21 referred to
momentum indicators, such as the translate, the “average stop being hit” is above) crossed; and three, the average stop in
Stochastic, RSI and MACD “work.” But determined as follows: (1) Moving averages the opposite direction was hit.
hard quantitative evidence is rare. Most of 10 and 21 were calculated. (2) If the fast Figure 1 shows an example of the count
support for the efficacy of these indicators moving average was above the slow moving being stopped based on a close beyond a
is anecdotal, based on traders’ experience average, the market is assumed to be rising, stop having a reversal value of 3.6 standard
or empirical, based on an indicator’s and if below falling. (3) If the market was deviations of True Range, Stop 3. The chart
performance when embedded in a trading considered to be rising, and then declined shows a valid bearish divergence as marked
program. In this research paper, hard by an amount equivalent to the average by the cyan dotted lines. Two bars later a
evidence of how these three well-known of a two bar True Range, as defined in the new high was made and so the program
indicators perform is presented. For sidebar below, plus one tick, the average stopped looking at stops being hit for that
purposes of this study, stops based on True stop was considered to have been hit, and divergence at that point. At the new high,
Range, as detailed in the Appendix were vice versa for a declining market. Once the another valid bearish divergence took
used to measure reversals. Two aspects average stop was hit, the stops based on that place. Then the market turned, prices
of indicators were studied. These were (1) bar were frozen so that the remaining stops dropped and the program counted the stops
whether a divergence took place preceding based on those in place at the same time of until the market closed below Stop 3, as
stops being hit or not, and (2) whether or the average stop hit could be evaluated. The shown by the blue arrow.
not following a divergence, the market key is that because the stop is based on True Figure 2 shows a case where the program
turned sufficiently to hit the stops. Range, which varies over time and is based stopped counting the stops. This occurred
The data used for this study included the on volatility, a trailing stop can change even because the stops where “flipping” based
most actively traded futures contracts per if a new high or new low was not made. This on the underlying moving averages crossing
the July 2006 issue of Technical Analysis of can happen because the value of the amount and the average stop being hit in the
Stocks and Commodities magazine. In all, a added to a low or subtracted from a high opposite direction. The chart shows a valid
total of 43 commodities and six FOREX pairs itself changes due to the change in volatility, bullish divergence, marked by the dark red
were used: the Australian Dollar, Canadian even though the high or low didn’t change. lines. One bar after Stop1 was hit the moving
Dollar, Swiss Franc, British Pound, and This is why the stop levels are frozen at the averages underlying the stop placement
Japanese Yen, against the US Dollar. This data time the average stop is hit. Having stops crossed at which point the stops flipped
extended back fifteen years, where available, that don’t change is important for purposes from short to long, as shown by the red
otherwise the maximum data available was of the study because the stops in place at arrow. After the stops flipped from short to
used. The data was provided courtesy of www. that moment in time that the average stop long, the average stop was hit, which was by
GenesisFT.com, and the format employed is hit are all the analyst has to work with, definition in the opposite direction of the
was back-adjusted normalized data. Back as opposed to future stop points that may initial count. At this point, shown by the dark
adjusting takes the difference between the change due to changes in volatility. The red arrow, the count was stopped.
first and second nearby contract prices process ended under these conditions: one, if It was found that of the 157,206 bars
upon expiration and adjusts all previous the market closed beyond a stop based on 3.6 of data, the average stop was hit a total
price points by that difference to remove standard deviations of a double True Range of 14,582 times, or about 9% of the time,
any rollover gaps. The raw data stream, excursion (referred to as Stop 3 hereinafter); approximately once every eleven bars. Once
when normalized in this manner, may have two, if the peak or low dip just prior to the the average stop was hit, the follow through
negative numbers so Kase wrote a program
to identify which streams had negative
values and adjusted them upwards to ensure
that this was no longer the case. As it turned Calculating Two Bar True Range
out there was an average of about thirteen
years of data per instrument, or about 630 H = highest high of two consecutive bars
years of data studied. L = lowest low of two consecutive bars
C = close of the most recent bar
Stops Hit and True Range Two Bar True Range = TBTR = maximum of absolute value H – L, H – C, C – L.
Excursions Average TBTR = average over n bars (default 30) of TBTR
In performing this study the first step was Standard deviation TBTR = standard deviation over n bars (default 30) of TBTR
to determine the behavior of the stops
regardless as to whether a divergence

page 42 IFTA.org
 IFTA Journal 2008 edition

to the remaining stops was evaluated. Figure 1. Methods of Stopping Stop Hit Count,
Specifically reversals were defined as Corn Continuation
magnitudes equal to Stop1, Stop 2 or Stop 3,
where the Stops are defined as reversals of
1, 2.2 and 3.6 standard deviations above the
mean of a two bar True Range, as defined
in the sidebar. Once the average stop is hit,
the probability of hitting the other stops
can be estimated as well, as shown in Table
1. For example, if the average stop is hit,
there is a 63% chance that Stop 2 will be
hit. Follow through based on other stops
can be calculated also. In such an event if
Stop 3 is hit, there is an 84% chance that
there will be a close beyond that stop.
The probabilities can be used both for
forecasting purposes and for risk control.

Indicator Functionality
The next aspect of the study involved
identifying when divergence took place on
each of the three indicators noted above,
and to measure if there was follow through Figure 2. Methods of Stopping Stop Hit Count,
in the form of a statistically significant Soybeans Continuation
move in the direction of the divergence
(down for bearish divergence following an
up market and up for bullish divergence
following a down market).
For purposes of this research,
statistically significant moves were defined
by evaluating divergences in two directions.
This required coding a divergence
identification algorithm; given that none of
the canned “divergence” programs that are
available in the public domain on charting
packages meet an appropriately strict
definition of technical analysis.
To find divergence signals, first peaks in
price and in momentum were defined and
identified. A peak was defined as a high in
price or momentum such that the day on
which the high took place was preceded
and followed by lower values. The program
allowed for peaks to be formed as plateaus
that consist of up to three equal bars Table 1. Stop Hit Follow Through
preceded and followed by lower values. The
reverse logic was used for dips. Once peaks
and dips were identified, if price peaks took Measure Number Follow Through >>
place on the same bar as momentum peaks, Total Bars 157206 Total
or within a tolerance of plus or minus two
bars of momentum peaks, a matching pair Average Stop 14582 9 Hit
was found. The inverse was true for dips. Stop1 11601 7 80 Hit
The charts below show examples of within Stop2 9174 6 63 79 Hit
tolerance and outside of tolerance, using
the Stochastic as the momentum indicator. Stop3 7075 5 49 61 77 Hit
Figure 3 shows a divergence in which both Close Beyond 5914 4 41 51 64 84
peaks in price matched the peaks on the
Stochastic exactly.

IFTA.org page 43
 IFTA Journal 2008 edition

Figure 3. Divergence on Stochastic, Corn Continuation hit was preceded by a valid divergence on
each of the three indicators studied, as well
as when combinations of the indicators
were used. Table 2 shows that the results
indicate that divergences found on all
three of the indicators studied, preceded
a turn that hit the average stop about the
same percentage of the time: 18% for the
Stochastic and RSI and a slightly lower 16%
for the MACD.
It is very interesting that there was a
significant increase in the turns caught
when the Stochastic and RSI were combined.
An improvement of 11 percentage points
from 18% to 29%, resulting in 60% more
turns being caught as demonstrated in
Table 3 Adding the MACD to either the RSI
or Stochastic did not improve the results as
much, only improving performance by about
four percentage points. This indicates that
there must be a fairly high degree of overlap
relative to the MACD for the Stochastic and
RSI, and much less overlap between the
Figure 4. Divergence on Stochastic, Gasoline Continuation latter. Combining all three indicators versus
just using the Stochastic and RSI only yields
a marginal two-percentage point improve-
ment. Thus, the conclusion is that using the
Stochastic and RSI is warranted, with the ad-
dition of the MACD, in cases where a trader
might be looking at one chart on a position
basis or if using a computerized model where
the work involved in adding the MACD is
insignificant. Table 3 also shows the percent
of each stop that was caught by a particular
indicator. The overall pattern remains the
same, and the values are roughly the same
regardless as to what stop is viewed, with
a slight peak at the Stop 2 (set at a reversal
value of 2.2 standard deviations of a two bar
True Range). While it is outside of the scope
of this study to determine why there is varia-
tion, the most likely explanation has to do
with minor variations in the degree of skew
relative to the log normality of the distribu-
tion of range.

Figure 4 shows examples or peaks defined peaks or higher dips. The two pairs of Stops Hit Following Divergence
“within” and “outside” of tolerance. The matching peaks or dips were then checked The next aspect of the study had to do with
first peak on the left, as shown by the blue for divergence. In this study bearish follow through. This means that once the
arrow is within tolerance because the peak divergence was defined as a higher or equal average stop is hit, how often there is a
in price is followed one bar later by a peak in peak in price matched by a lower or equal continuation against the direction of the
momentum. The peak in momentum does not peak in momentum, and bullish divergence trend such that the more distant Stops
match the peak in price shown by the cyan as a lower or equal dip in price matched by 1, 2 and 3 (at one, 2.2 and 3.6 standard
arrow until three bars later, and therefore, a higher or equal dip in momentum. deviations) are hit, and/or if a close beyond
with a two bar tolerance, was breached. Stop 3 takes place.
Once a matching set of peaks or dips Divergence Preceding Stops The second column in Table 4 indicates
was found, the algorithm looked up to Once all divergences were found, the next the number of times a particular stop
100 bars back for an earlier pair of lower step was to see how often each average stop was hit, and when a close beyond Stop 3

page 44 IFTA.org
 IFTA Journal 2008 edition

took place. The percent column shows the Table 2. Average stop Hits Caught or Missed
corresponding percent of the time, relative
to the average stop being hit that follow Average stop Number Caught Missed
through took place. So for example, 80% of
the time that the average stop was hit, there Total Hits 14582 – –
was follow through with Stop 1 being hit, Stochastic 2601 18 82
and 41% of the time the market continued RSI 2657 18 82
against the original direction to an extent
that a close beyond Stop 3 took place. This MACD 2885 16 84
was then compared to the follow through Stochastic and RSI 4206 29 71
that took place after a divergence. The results Stochastic and MACD 3177 22 78
indicated that follow through in terms of
all three indicators studied was about the RSI and MACD 3280 22 78
same so a representative column is shown All Three Indicators 4502 31 69
as “Indicator.” What is interesting here
is that the follow through of all the stops
was slightly less, with the follow through
on a close beyond Stop 3 slightly higher. Table 3. Percent of Time Stop Caught by Indicator or Combination
Though the differences are not large it could
mean that the indicators have a slight bias
Stop>>> Average Stop1 Stop2 Stop3 Close
to finding a somewhat larger number of
reversals that are of larger magnitude. Stochastic 18 17 16 17 18
Table 5 shows the percent of the time RSI 18 17 17 18 20
each stop was hit based on each individual
MACD 16 16 16 17 19
indicator and their combinations. Each
individual indicator has similar follow Stochastic and RSI 29 26 25 27 31
through, which is consistent not only with Stochastic and MACD 22 20 19 21 24
the average stop, as noted above, but also
RSI and MACD 22 20 20 22 25
with the other stops. If a signal is received on
the MACD and either the RSI or Stochastic Any Indicator 31 28 27 29 34
at the same time, there was no significant
difference. The same is true for all three, but
the lack of differentiation may be due to the
very small number of signals. If a signal took Table 4. Follow Through after Hitting Average stop
place at the same time for the Stochastic
and RSI, the likelihood of hitting the average Stop Number All Indicator
stop and Stop 1 increased, but the results
Stop1 11601 80% 76%
for the lower stops remained the same. The
implications of this may be that if a trader Stop2 9174 63% 58%
has stops set at the average stop or Stop 1, Stop3 7075 49% 48%
the odds of being hit are slightly larger so a
Close Beyond 5914 41% 44%
slightly more aggressive exit strategy may

Table 5. Percent of Time Indicator Hit Reversal

Stop # Signals % Signals Average Stop1 Stop2 Stop3 Close Beyond

Stochastic 2601 25 85 63 47 39 35
RSI 2657 26 82 61 47 39 36
MACD 2354 23 83 65 51 43 38
Stochastic RSI both 1325 13 90 67 49 42 40
Stochastic MACD both 472 5 86 62 48 42 40
RSI MACD both 544 5 85 62 48 41 39
Improvement – – 4 3 2 34
All Above, same time 342 3 85 62 48 40 39

IFTA.org page 45
 IFTA Journal 2008 edition

Figure 5. Divergence on Stochastic, Gasoline Continuation be warranted. However, if the stops are set
at Stop 2 or greater, no change in strategy is
justified. In the writer’s opinion the change
is not large enough to call for modification of
trading strategies.
Another implication of the results is
that the indicators predict average stop
hits about 9.2 times as frequently as
random. The odds of hitting an average
stop are about 83% following an indicator
divergence signal, versus 9% for random
hits. The Stochastic/ RSI combination is ten
times more frequent than random.

Impact of Optimization
In this portion of the study, the periodicity
of the Stochastic and RSI was varied, using
values of 5, 8, 13, 21, 34 and 55 in addition to
the eSignal® defaults to determine if there
was any improvement in functionality as
periodicity varied. As reflected in Figure 5,
there was only a slight degradation in the
Figure 6. Divergence on Stochastic, Gasoline Continuation percent of the average stop was hit and follow
through after a divergence took place. The
major difference was found when evaluating
the performance of the indicator from the
opposite direction. Meaning the rate of
the average stop was hit, how often was it
preceded by a divergence signal, or to put it
another way, how many of the hits did the
indicator catch. Here the performance more
or less increased as periodicity decreased,
making the RSI and Stochastic set at the
smallest reasonable setting, the best choice.
Figure 6 shows the relationship between
the percent of the time the average stop
was preceded by a Divergence, with the
relationship clearly one in which the
accuracy of the indicator declines as the
periodicity increases.
Table 6 shows the detail of the optimal
indicators versus the defaults in table format.
While it makes sense that the shorter the
periodicity, the more accurate the indicator,
one might expect that there would be a
Table 6. Average stop Hits Optimized vs. Default Indicators degradation in the amount of time the
average stop was hit after a divergence. That
Average Stop # Optimal % Optimal % Default Δ Points Δ% is, one might have expected an increase in
false signals as a trade-off for fewer stops
Stochastic 3490 24 18 6 33%
being missed. As shown in Figure 6 and
RSI 4165 29 18 11 61% discussed earlier this was not the case. The
Stochastic and RSI 6118 42 29 13 45% table summarizes the difference between
the indicators when using a periodicity of
Stochastic and MACD 4381 30 22 8 36%
five versus the default settings at 14. Again,
RSI and MACD 4921 34 22 12 55% the best combination of two indicators
All Three Indicators 6450 44 31 13 42% is the Stochastic and RSI, with a 13 point
improvement for the 5 versus 14 period
indicators, a 45% improvement overall.

page 46 IFTA.org
 IFTA Journal 2008 edition

Variation Among Markets Table 7. Percent of Follow Through after Signal

The last part of the study was conducted Difference Between FOREX and Agricultural
to determine if there was any significant Average Stop and Close Beyond Stop3 (Italic)
difference in the performance of the
indicators among commodity or instrument
All FOREX Ags All FOREX Ags
types. Thus, approximately 2,000 data
points for each of six FOREX pairs of US Stochastic 85 88 83 35 33 30
Dollar to Australian Dollar, Canadian RSI 82 81 81 36 32 36
Dollar, Swiss Franc, Euro, British Pound and
MACD 83 82 81 38 38 39
Japanese Yen was compared to the six most
active agricultural commodities, also based Stochastic RSI both 90 91 91 40 33 34
on approximately the 2,000 most recent Stochastic MACD both 86 88 83 40 37 33
data points. These included Corn, Cotton #2,
RSI MACD both 85 84 79 39 32 38
Soybeans 5000 bushels, Sugar - World #11,
Soybean Meal, and Wheat - Soft Red. All Above, same time 85 86 86 39 33 38
As shown in Table 7 there was little
difference found. When comparing FOREX
to the agricultural commodities, the average
stop hit percentages were about two percent Table 8. Percent Stops Preceded by Signal
better, and the close beyond Stop 3 one Average Stop and Close Beyond Stop3 (Italic)
percent worse. Comparing the FOREX to all
the data contained in the entire study the Average Stop All FOREX Ags All FOREX Ags
results were one percent better and four
percent worse, and for agricultural products Stochastic 18 19 15 24 23 22
two and one percent worse, respectively. RSI 18 17 18 29 29 27
These minor differences can be attributed MACD 16 15 15 16 15 15
to small variations in market activity, such
as fewer trend reversals, than any factors Stochastic and RSI 29 28 26 42 42 39
inherent in either market segment. Stochastic and MACD 22 22 20 30 29 29
In evaluating performance in the reverse RSI and MACD 22 22 23 34 33 32
direction, that is, relative to how often a
turn was preceded by a particular signal, no All Three Indicators 31 30 29 44 43 42
differences were found on average between
all the data and FOREX. Agricultural
products scored two percent worse, which
Bibliography
is again considered a minor variation. 1. Easy Language User’s Manual, Version 4, 9. “The Kase Dev-Stop: Accounting for
Omega Research Inc. 1997. Volatility, Variance and Skew in Managing
Conclusions Trading Risk.” Journal of Technical Analysis,
This study has shown that momentum 2. “Futures Liquidity” Technical Analysis of Summer 1993.
Stocks and Commodities July 2006, p. 97.
indicators can predict market turns that
10. “The Two Faces of Momentum.” Stocks,
are of sufficient magnitude to generate an 3. Kase, Cynthia A., “Knowing When to Step Futures, and Commodities October 2003.
average double-bar True Range reversal in Back From the Market.” Futures Magazine,
increasing rates of accuracy as indicator June 1991. 11. “Tools for Technical Analysis.”
periodicity decreases, from about 18% for Energy Markets, April 2005.
4. “Managing Trade Risk.” Trader’s Catalog &
a single indicator with eSignal® default
Resource Guide July 1999. 12. Trading with the Odds McGraw-Hill, 1996.
settings to 44% using all three indicators
with optimized settings. Combinations of 5. “Momentum Divergence.” NYMEX Energy 13. Wilder, J. Welles, New Concepts in
the Stochastic and RSI are far better than in the News Fall/Winter 1993. Technical Trading Systems. Hunter
either indicator combined with the MACD. Publishing Company, 1978.
6. “Proof That Technical Analysis Really
Once the average stop has been hit, there is
Works.” Commodities Now, March 2005. Software and Data
no significant variation in follow through All code used in this study was programmed
between instances in which a stop was hit 7. “Redefining Volatility and Position Risk.” in-house by Kase and Company, Inc. Data
and not preceded by a divergence or in cases Technical Analysis of Stocks and Commodities courtesy of www.GenesisFT.com, eSignal
in which a divergence did take place. Finally, October 1993. 8.0, eSignal, Hayward, CA MS Office 2000
Standard Edition, Microsoft Corporation,
no significant variations between markets,
8. “Setting Stop-Losses Using Price Redmond, WA TradeStation 2000i,
specifically in the cases of FOREX and Volatility.” The Technical Analyst, TradeStation Securities, Inc., Plantation, FL
agricultural products, were found. IFTA July/August 2005.

IFTA.org page 47
 IFTA Journal 2008 edition

Harmonic Ratios as Applied To Commodity Market

Technical Analysis
by George Alexander MacLean, MFTA

Abstract are excellent predictors of target objectives modern tuning in addition to these
This research paper tests if further and of stop-loss levels in short term three ratios.
retracement levels can be found and trading, I believe the importance of choice First, a description of the three Greek
usefully applied in technical analysis of origin of a move is often overlooked. terms in the previous paragraph. Taking
in addition to the existing retracement Gann insists that the lifetime high/low a string of any length and bisecting it and
systems such as Gann and Fibonacci which is critical to his analysis and intra-day then plucking one of the lengths will give
are already successfully used. In addition, if traders often forget this. It would be a note exactly one octave higher than that
early signals from penetration of a nearby more beneficial if support and resistance heard from plucking the original length.
retracement could give warning of reversal levels much closer to the current price This is normally accredited to Pythagoras,
moves developing. action could be applied. Whilst important listening to the different sounds of
The author believes that financial retracements are undoubtedly key target hammers hitting anvils. As his starting
markets act as a natural system, in that objectives, it could be many days if not point, the halving of a string length gave a
not only do prices have a relationship with weeks before the price comes anywhere note exactly one octave higher and this he
time (Gann theory) and with each other near these levels, e.g. using the 1974 low called the Diapason. Other divisions of the
(broad technical analysis) but also display a for a Gann retracement may see great original length such as dividing in the ratio
mathematical relationship to recent highs distances between levels. I found this to be 4:3 gives Diatesseron and in the ratio 3:2,
and lows (harmony). a significant drawback in intraday or very Diapente. These will be seen in tables of
Constraints set by this examination short term technical analysis. My research Harmonic ratios. It is my hypothesis that
allow for research in this paper to be into finding retracements which have the relationships on a musical scale i.e. the
restricted to one broad area (agricultural a “natural” basis, e.g. Fibonacci ratios, ratio of one to another, can also be applied
commodities). The author believes, where the 1, 1, 2, 3, 5…sequence and the to charts as an additional technical tool.
however, that research done in other ratio 61.8% occur so frequently in nature, This paper will investigate the success of
markets (MacLean, 2005) would give has led me to consider musical Harmonics. this hypothesis.
similar results. This paper will use the The search for derived or “synthetic” There is a branch of Technical Analysis
inverse decimal value of harmonic ratios ratios (MacLean 2005) is often a laborious which looks at Harmonics as derived from
(the size of the musical interval from one and thankless task. Le Corbusier, in his the motion of heavenly bodies such as
note to another) as a test of whether this search for architectural proportion, for the planets, moons and comets. This is
hypothesis is valid. example, fudged the issue when creating on a very grand scale. I am instead going
the Modulor (Le Corbusier, 1954) by to focus on Harmonics on a much smaller
Introduction ignoring the average height of a Frenchman scale, while retaining some of the natural
Prices in financial markets reflect pressures and instead took the total height of an characteristics of my starting point: that
on trader’s activity in that the price of a Englishman wearing a hat in order to arrive of music and harmony. In this paper,
commodity at a point in time is the result of at the proportions on his Red and Blue I shall concentrate on Western music
the action of buyers and sellers. This paper Modulor scale. especially that of notation and harmony
tests if financial markets display some It is my intention to look for such a described since 1290. This Western
Harmonics, in that Harmonic support or universal measure of proportion, one tuning, as seen in Table 1 through Table
resistance lines will be evident where price coming from a natural source: that of 3 is called Lydian Mode, from Mode V of
moves are seen. Gann, Fibonacci and Elliott music. I have developed in this paper work the Gregorian Chant of the early Classical
Wave studies look at price moves over very I started in my book Fibonacci and Gann Period which is now the most common
large vertical (price) scales and whilst this Applications in Financial Markets (MacLean Pythagorean tuning mode. Other modes,
has a place in traditional technical analysis, 2005) which examined Pythagorean such as Equal Temperament (where
my experience of intraday market analysis musical Harmonic periods. As in the work the difference in notes is measured by
sees these as longer term objectives, often of Gann and Fibonacci analysis, I followed frequency (Hertz, Hz) and measured in
too far away from current market action the basic rule that a significant high or low cents, are not discussed as these are
and thus not of great utility for short term followed by a reversal move as my starting derivatives of the initial 440Hz measure
benefit to intraday and other short term point. However, instead of looking only at for Middle C and have no “natural” basis.
traders. the Diatesseron, Diapente and Diapason as The other Gregorian Modes are detailed in
Although Gann and Fibonacci studies I did in my published work, I now turn to the Appendix.

page 48 IFTA.org
 IFTA Journal 2008 edition

Table 1. Harmonic Ratios The Harmonic Ratios and

an Introduction to Musical
Harmony
This paper uses the term “harmony” to
Ratio Interval Decimal Inverse Decimal Traditional Name
describe proportions which are linked and
10:01 Unison 1.00 1.00 derived from a much larger measure. The
measure is an Octave (a musical notation
2187:2048 Major chorma 1.07 0.94 having eight parts) and used as the distance
from a low to a high (or a high to a low)
09:08 Major second 1.13 0.89 as is the case in measuring Fibonacci and
Gann retracements. However, by avoiding
32:27 Minor third 1.19 0.84
the pitfalls of using significant highs/lows
81:64 Major third 1.27 0.79 Ditone as is necessary on Gann (lifetime high
or low in that case), and as in the case of
04:03 Perfect forth 1.33 0.75 Diatesseron Fibonacci where looking at a move that
may have taken some time to develop, the
729:512 Diatonic tritone 1.42 0.70 use of “Harmonics” will take only localised
significant highs or lows and a nearby
03:02 Perfect fifth 1.50 0.67 Diapente
extreme. This makes it useful in very short
6561:4096 Minor sixth 1.60 0.62 timeframes.
Table 1 is constructed as follows: Taking
27:16 Pythagoreran 1.69 0.59 the original measure (1:1) as a starting point
major sixth and then adding a perfect fifth will give
3:2, the second harmonic. Applying a fifth
16:09 Minor seventh 1.78 0.56 onto this measure, we get 9:4 (multiplying
the ratio 3:2 by 3:2) but this will take us out
243:128 Major seventh 1.90 0.53
with the scale being greater than two (an
02:01 Octoave 2.00 0.50 Diapason octave) so we move down an octave from
9:4 which results in 9:8 (9:4 multiplied by
1:2). This results in the third harmonic.
We now have the series 1:1, 3:5, 9:8. The
fifth column (Table 2) is the standard
Table 2. Creation of the Octave from F below Middle C (C’) musical notation, starting at the F below
Middle C (C’).
The intervals from this table result
Measure Note Action Result Note from the relationships within the ratios.
For example, taking the Fourth B-E’ is
1:1 F Add a Fifth 3:2 C` the difference between F-E’ and F-B. As
we divide to get the interval, this means
3:2 C` Add a Fifth 9:4 (243:128)/ (729:512) = 4:3 and this is

9:4 Subtract an Octave 9:8 G

Table 3. Ordering of Table 3
9:8 G Add a Fifth 27:16 D` Putting these in order

27:16 D` Add a Fifth 81:32 Note Measure

81:32 Subtract an Octave 81:64 A F 1:1
G 9:8
81:64 A Add a Fifth 243:128 E` A 81:64
B 729:512
243:128 E` Add a Fifth 729:256
C`(middle) 3:2
729:256 Subtract an Octave 729:512 B D` 27:16
E` 234:128
1:1 F Add an Octave 2:1 F` F` 2:1

IFTA.org page 49
 IFTA Journal 2008 edition

Figure 1.
Corn: Daily Continuation chart
with comparison Fibonacci
retracement

Figure 2.
Corn: Continuous, Longer-
term Fibonacci retracement
and shorter term Harmonic
retracement

Figure 3.
Corn: Continuous Figure 1
Fibonacci retracement and
Figure 2 Harmonic retracement
combined

page 50 IFTA.org
 IFTA Journal 2008 edition

repeated throughout the scale (A-D’ – tools more sympathetic and in line with the key, as a break there would trigger a move
27:16/81:64 = 4:3). Other differences will development of Gann analysis as applied to to the 5-6th 241.56/239.75 congestive
give the missing ratios from Table 1. Note commodity markets in the early part of the area and lower. This is the ideal test for
that this calculation method does not last century. The initial area to study will Harmonic supports as the contract has
give rise to values less than 50% (from be the Grains Market, where chart analysis moved from traditional narrow range
Table 1, Inverse Decimal), but does for the will be given and the Harmonic result of trading to a more violent period where
Decimal value. This paper will look at the success or failure discussed. volatility has substantially increased. This
inverse decimal values only as testing is change of contract behaviour should draw
beyond the scope of this paper. As with Corn attention from analyst and trader alike as
Fibonacci and Gann analyses, I look for In Figure 1, the sharp move in late October there can be opportunities for profit in this
a nearby significant low (high) and start saw key resistance levels taken out. environment. Using Harmonics here gives
my measurement. This would then need Applying the Harmonic retracements from key early triggers and should be useful
a high (low) to act as the next “octave”, the July low at 204.50 to the October 251.00 for nearby entry trigger levels and price
being the extreme of the rally (fall). From high shows that 2nd and 3rd retracements objectives.
there the retracements would correspond (245.83/243.73) acted as good support in a
to the reciprocal levels in Table 1 (0.94, retracement. A break of the 4th at 241.24 Rough Rice
0.89, 0.84…) counting from the high or the will be significant, as this will also see the The rally from the October low saw
low. For deeper retracements the original break of congestion from the September traditional congestive resistance levels
decimals from Table 2 would need to be high and the top of minor congestion from successfully taken out. This is an ideal
tested. October. The oscillation about the 1-3rd time to apply Harmonics. as a minor
retracements is significant here as it is pullback has occurred from the previous
Application of Harmonic forming a potential bull flag although in session high. To confirm that the rally has
Retracements to traditional pattern recognition technical not run out of steam the 1st Harmonic
Commodity Markets analysis it is still too early to confirm. has to be taken out at 7.54. Although this
Agricultural markets by their very nature Comparing this with Fibonacci is under attack, a convincing intraday
are influenced by various natural forces, retracements confirms that the current break is needed; otherwise the threat is
from the effect of seasons, weather, action is consolidative about the 50% slippage back through the 3-4th Harmonic
rainfall, disease, yield, consumer demand retracement of the 285.50/205.00 down at 7.47/43. However, as long as the bull
and so on. Despite the spread of countries move at 245.00 from which a price recovery channel support is not penetrated and
growing various commodities these various appears difficult. Harmonic support levels do not come under
factors combine to make supply difficult Applying Harmonic retracements to the pressure, the recovery should continue. A
to calculate from one growing season to low of October gives a better picture. The break of the 1st Harmonic on a significant
another. This effect of natural pressure current local price action finds support at basis will be required before calling for the
should make analysis using other natural the 4th Harmonic at 243.08. This will be extension to develop.

Figure 4.
Rough Rice: Continuation
chart and Harmonic
retracement

IFTA.org page 51
 IFTA Journal 2008 edition

Figure 5.
2nd line and harmonic
retracement

Figure 6.
Cotton: Continuation chart and
two harmonics

Cotton clear that the move from the late August 6 from the Jun 48.15 low and the August
In figure 5 recent sharp rally in Cotton looks Gap has been overdone. The traditional 54.30 low confirms current price action
to be over with a bearish outside week bull channel (top of channel shown as a is at a key level. This is the conjunction
followed by a gap down. The gap through trendline) which existed for the previous of the 4th Harmonic from the June low at
the 2nd Harmonic at 78.84 and pressure year was penetrated on a spurt of price 77.11 and 5th from August at 77.18 with
building on 3rd at 76.43 should convince action but looks to have over extended. This the joint Harmonic of 5th (June at 75.64)
that the steep rally from late August was slide should bring the contract back within and 6th (Aug at 75.72) looking threatened.
overdone. The target from figure 5 looks to a more “normal” bull channel and at that Again with Harmonics it is clear that the
be at the 4th Harmonic at 73.55, which is point further analysis should be applied. downside is favourable from here and key
at the base of the small congestion before The top of the old channel comes in at support levels and targets are seen at 6th
the gap in September. The current move is 72.42, which is close to the 5th Harmonic at (Jun at 73.89) and 7th (Aug at 74.63).
seen as more than a pullback on modest 71.40. In this chart the contract has broken
profit taking as with the bearish outside out of the 1-3rd harmonic retracement Cocoa (London)
day and week occurring in the previous levels and with the bullish outside day, Again combining two Harmonics on the
session and the gap lower, a major change followed by a gap shows in both Harmonic daily chart confirms the importance of
in market sentiment is seen. This should retracements that a major reversal move the current congestion. The last week has
see the contract continue through the 3rd looks convincing and is in an advanced seen consolidation about the 1st and 2nd
Harmonic and lower. state. from the 1135 high at 854.07 and 868.33
Looking at the long-term chart, it is Combining two Harmonics on figure and the 1st and 3rd from the 1075 high at

page 52 IFTA.org
 IFTA Journal 2008 edition

Figure 7.
Cocoa: Continuous chart and
two harmonic systems

Figure 8.
Rapeseed: Continuation chart
and harmonic

850.25 and 872.50. In traditional technical in a recovery move but as the contract is with a longer-term Fibonacci retracement
analysis the current congestion is forming oscillating about the 1-3rd Harmonics, pattern. Again the sharp slide has seen key
a bear flag, yet with a suggestion that it would be appropriate to wait for a Harmonics taken out and with the pressure
recovery moves from the fairly strong break of the base or of the 3rd Harmonic building the 7th Harmonic at 374.87, which
base, formed above the 835.00 low. This retracement in the 2nd Harmonic system is close to the 23.6% retracement at 374.64,
will be confirmed with the current bullish (just above the 2nd Harmonic in the 1st this area is going to be key. A break of this
outside day succeeding in breaking the 2nd Harmonic system) before taking a position. support zone will then trigger a move to the
Harmonic of the 1075 measure at 861.67, Figure 8 shows a rather stringy Rapeseed 9th Harmonic congestion at 371.75.
and from there to attack 3rd at 872.50 and contract with a very sharp turn down on These close levels to the significant
move higher. Until then, however, there is the last bar. This has seen key support high are useful as confirmation of the
little to suggest that this consolidation is levels taken out as far as Harmonics are downtrend developing. In traditional
going to end. This is a good example of the concerned, with conformation of the end technical analysis a break of the 23.6%
conjunction of Harmonic retracements. of the rally on a break of the 2nd Harmonic support is needed before confirmation
The close proximity of the Harmonic lines at 382.58. The sudden move though the 3rd of a downturn is seen (although other
to each other increases the value and and 4th Harmonic at 380.01/376.96 has techniques could have supplied the
importance of each level and in addition seen acceleration in the decline and there necessary negative evidence before then if
will give early confirmation of a recovery is now evidence of a potential move to the the analyst is using traditional techniques).
move should one develop. This is the key. An congestion about the 6th at 371.96. The significant break out of the 1-3rd
early signal is needed in order to participate Figure 9 compares a shorter Harmonic congestive Harmonics would have been the

IFTA.org page 53
 IFTA Journal 2008 edition

Figure 9. trigger for taking a reversal position. The

Rapeseed: Continuation chart and Fibonacci retracement purpose of Harmonics is to give early and
confirmable triggers that a reversal move
is developing, long before more traditional
measures are called into play and in this
case the failure to continue the rally
through the recent high and the subsequent
slippage through the 3rd Harmonic (382.32)
should be enough warning that a downturn
is ahead. With the break of congestion at
the 3rd and 4th Harmonic (382.32/380.06)
confirmation is seen and the next bearish
move has confirmed the downturn,
putting key congestion under pressure
ahead of the 8th Harmonic and the 23.6%
Fibonacci retracement at 373.08 and
374.64 respectively. The final area of
study is to sample and analyse the Meats
market.

Live Cattle
Figure 10 shows the contract at the peak of
an extended rally. Current action is trying
to recover but the 1st Harmonic at 101.55
is acting as a very strong resistance level.
Figure 10. Previous moves have seen the contract slide
Live Cattle: Continuation chart and Harmonic retracement to test and bounce from the 5th Harmonic
at 95.54, but there is some doubt that the
subsequent recovery move can be sustained
as the contract has oscillated within the
2nd and 3rd Harmonic (100.02/98.56)
before failing to push significantly higher
on the bounce after filling the gap. Bulls
would need to see a significant attempt
to break through the 1st Harmonic on
a sustained basis. Until his happens at
101.55 the contract is likely to be moving
sideways, at best. Support here comes at
the 2nd Harmonic and if that breaks a move
back lower is likely. The break of the 3rd
Harmonic support did not see the contract
reversing significantly, but as the reversal
was followed by a bounce from the 5th with
a series of gaps, taking a recovery position
should have been suggested. Again there
is a threat of a near double top developing
which needs confirmation. An early
confirmation would occur on another break
of the 3rd Harmonic.

page 54 IFTA.org
 IFTA Journal 2008 edition

Figure 11.
Frozen Pork Bellies Pork Bellies: Continuation chart and Harmonic retracement
Figure 11 illustrates another example where
the recovery move from the 83.05 low is
struggling. Despite the bullish outside
day the contract is straining against the
12th Harmonic (50% retracement). Until
there is a break of the 50% retracement
level, consolidation should continue.
Noted should be the identified support
at 86.50 (the 7th Harmonic). Looking at
the warning signals from the Harmonic
penetration with the bullish outside
day, we see a move out of character—the
contract had moved only within three
Harmonics on a daily basis. Once the
bullish outside day broke through the
4th Harmonic at 85.22, this suggested
a change in character in the contract,
keeping the focus on the upside
throughout the day. Indeed, the move
through the 7th Harmonic should have
confirmed that the contract was in a
recovery move.

Conclusion Harmonic are very close to the origin of

The purpose of this submission paper was the reversal move. Constant, building
to discover if there were additional levels pressure on these levels should keep
that could be used to determine likely the trader’s attention focused and once
price reversals. The initial supposition there is a break of the 3rd Harmonic
was that Fibonacci and Gann vertical this will confirm that a reversal move
retracements were useful but that 1/8th is developing. It is my suggestion that
or major Fibonacci levels (38.2%) were 1st-3rd Harmonics be used as the warning
too far away from the reversal start. This area, similar to the warning lines as seen
meant that some profit had to be foregone in Lane’s Stochastics or the Welles Wilder
before confirmation of the reversal move RSI. The trigger should then come on a
was given. Futures markets, by their break of the 3rd Harmonic. This close
nature change rapidly; when a reversal congestive grouping, not only of price
move is developing, acting on it could be action but also of Harmonics, has been
critical to maximisation of profits. a good indicator of market sentiment in
I have suggested that the use of the above tests. From the results of this
relationships developed from another research, this looks to be a successful
natural system, musical harmonics, could method of applying Harmonic Proportion
be used. My premise was that the basis to agricultural commodity markets. IFTA
of Fibonacci numbers, and by extension
retracements, fans, etc., was on naturally
occurring proportion. Musical Harmony,
another natural proportion, with its
resultant levels may give significant help
in a trading environment. I applied the
Harmonic to recent price moves, moves
from a local high or low. The results have
confirmed my hypothesis: The natural
proportions seen in Musical Harmony can
be used to identify change early at an early
stage of a price trend move.
In some of the tests above, the 1st-3rd

IFTA.org page 55
 IFTA Journal 2008 edition

Appendix Bibliography
Gregorian Plainsong 1. Frost, AJ and Prechter, RR Elliott Wave
Glareanus (1547) standardised the Modes Principle: Key To Stock Market Profits. New
into the following classifications (Modes Classics Library (ISBN 0932750079), 1985.
IXII).
Mode I 1:1 9:8 32:27 4:3 3:2 27:16 16:9 2:1 2. Le Corbusier (Charles Edouard Jeanneret)
(Dorian) D E F G A B C D The Modulor: A Harmonious Measure to
the Human Scale Universally Applicable to
Mode II 1:1 9:8 32:27 4:3 3:2 128:81 16:9 2:1 Architecture and Mechanics. Faber & Faber
(Hypodorian) A B C D E F G A Ltd, Birkhäuser Basel; 2000.

Mode III 1:1 256:243 32:27 4:3 3:2 128:81 16:9 3. Le Corbusier (Charles Edouard Jeanneret)
2:1 Modulor 2 (Let The User Speak Next), reprint
of 1955 edition. Birkhauser Verlag AG, Basel.
(Phrygian) E F G A B C’ D’ E’ 2000.

Mode IV 1:1 256:243 32:27 4:3 1024:729 4. Livio, Mario The Golden Ratio: The Story
128:81 16:9 2:1 of Phi, The Extraordinary Number of Nature,
Art and Beauty. Headline Book Publishing
(Hypophrygian) B C D E F G A B Ltd. 2003.

Mode V 1:1 9:8 81:64 729:512 3:2 27:16 5. MacLean, George A, Fibonacci and Gann
243:128 2:1 Applications in Financial Markets: Practical
Applications of Natural and Synthetic Ratios
(Lydian) F G A B C D E F in Technical Analysis. John Wiley & Sons,
2005.
Mode VI 1:1 9:8 81:64 4:3 3:2 27:16 243:128
2:1 6. Plato , Plato’s Cosmology: The Timaeus
of Plato,Translated With A Running
(Hypolydian) C D E F G A B C Commentary.Kessinger Publishing Company
Translated by Cornford, FM, 2004.
Mode VII 1:1 9:8 81:64 4:3 3:2 27:16 16:9 2:1
Acknowledgements
(Mixolydian) G A B C D E F G Charts from in this paper are created with
Metastock of of Reuters PLC.
Mode VIII 1:1 9:8 32:27 4:3 3:2 27:16 16:9 2:1

(Hypomixolydian) D E F G A B C D

Mode IX 1:1 9:8 32:27 4:3 3:2 128:81 16:9 2:1

(Aeolian) A B C D E F G A

Mode X 1:1 256:243 32:27 4:3 3:2 128:81 16:9

2:1
(Hypoaeolian) E F G A B C D E

Mode XI 1:1 9:8 81:64 4:3 3:2 27:16 243:128

2:1
(Ionian) C D E F G A B C

Mode XII 1:1 9:8 81:64 4:3 3:2 27:16 16:9 2:1
(Hypoionian) G A B C D E F G

page 56 IFTA.org
 IFTA Journal 2008 edition

Alteration of Price Movement Dynamics on a Chart

via Quantization of the Change in Closing Prices.
by Mohammed El Saiid, MFTA

Introduction manipulation had proven very useful in the motion through a single implication. This
Part one introduces a method which visualization of long-term trends. concept implies that it is essential to study
proposes a new technical analysis (TA) tool. Several other methods of data pre- the direction of price motion over time
This TA tool is primarily applied to market manipulation resulted in various useful separately (in isolation) from the amplitude
indices and averages, and is generally used market indicators and oscillators, such of price motion over time.
for market forecasting. This method will be as the Momentum indicator, and the Accordingly, this particular implication
referred to here on as “the application.” Stochastic and RSI oscillators. provides leeway to searching for new
The introduction also presents two methods of data representation. The
fundamental concepts related to TA, 1.2 Left and Right Translations main aim of such methods is to somehow
namely the concept of pre-manipulation in Cycles display the direction of price motion over
of data, and the concept of left and right As commonly known within the subject of time, while isolating or at least reducing
translations in cycles. These two concepts time cycles, all trends of the market are the effect of amplitude changes in such
form the core idea of the application. regarded as a series of interacting cycles motion. And hence, the application
with troughs and peaks. An ideal peak presented in this research is intended to
1.1 Pre-Manipulation of Data should occur exactly halfway through the provide a method for pre-manipulating
Through the past decades, various cycle’s period. In practical life, however, price movement data through a form of a
techniques, methods and applications were ideal peaks seldom occur. What does quantization process. The core idea is to
introduced to the field of technical analysis. occur is that cycle peaks tend to behave take into consideration only the direction
Most of them were based on observations differently with respect to the major trend of the change in closing prices (∆ closing
of characteristics of price (value) action on of the market (or larger cycle). In other prices) over time while totally disregarding
charts. Some of these observations were words, during positive trends, peaks of the amplitude of this change.
later developed into concepts defining cycles tend to shift from the midpoint of the The aim of this pre-manipulation
market movements. Nevertheless, all cycle period to its right side. On the other method is to graphically create an
of these observations led to numerous hand, during negative trends, peaks tend to alteration in price movement dynamics
forms of interpretations of price motion. shift to the left side of that midpoint. When appearing on a price chart. Through
Such interpretations were successfully we put this into perspective, the concept this alteration, better visualization
translated and represented graphically on of left and right translation from a cycle’s will be achieved on the chart providing
charts through various methods of pre- midpoint states that the market spends new information that may better aid in
manipulating price action data. more time in the direction of the ongoing predicting future price trends.
Some of these pre-manipulation major trend and therefore, less time
methods provided different forms of correcting from it. Part Two: Methodology
graphical representation. Such graphical Quoting John J. Murphy’s book, Technical Part two presents the method used for the
representations resulted in various means Analysis of the Financial Markets, he construction and implementation of the
for displaying price data as alternative discussed the topic of time cycles, “Stop application. This method is conducted
types of charting techniques. One common to think about it, all we’re saying is that in through a two-stage process, namely the
type—the point and figure charting— a bull trend, prices will spend more time quantization of data (∆ closing prices), and
utilizes the same price action data used going up than down. In a bear trend, prices the alteration of price movement dynamics
for representing line or bar charts. spend more time going down than up. Isn’t on a chart. Within this method, the second
However, this technique displays such data that the basic definition of a trend?” stage (process) is considered to be the
differently by focusing only on significant The key words used by Murphy here resultant or by-product of the first stage.
price changes, while disregarding the are “up” and “down,” which define the
element of time. direction of motion. In general, it holds 2.1 Quantization of Data
Another data pre-manipulation method true that the basic factors determining (D Closing prices)
sought for alternative means of price price movement dynamics over time Quantization, by definition, is the process
scaling. A common example to that is the are amplitude and direction of motion. of limiting (approximating) the possible
graphical conversion of an arithmetic Nevertheless, the concept of left & right values of a continuously progressive data
scale chart to a semi-log scale chart. translation of cycle peaks provides a key to in nature and varying in magnitude to
This particular example of data pre- better understanding of the nature of price a discrete (separated) set of values (or

page 58 IFTA.org
 IFTA Journal 2008 edition

symbols) of restricted amplitude, rather possible outcomes that define price motion Positive, negative “one” and “zero”
than varying continuously. regarding direction, namely UP, DOWN and values are then assigned to each of their
In other words, a group of progressing FLAT. corresponding symbols as shown in
values are reorganized by being placed into Accordingly, any change between two column five.
new categories or sets. Each of these sets consecutive closing prices resulting in And finally the last step before
is separated from the others in such a way either a positive, negative or neutral change presenting the data graphically, the new
that any specific value from the original (no-change) is categorized as UP, DOWN or values of column five are accumulated as
data would belong exclusively to one FLAT respectively. The new categories are shown in column six.
category or set but none of the rest. then each assigned to a positive, negative The phrase “Price movement dynamics”
A most common example to the one, and zero values also respectively used in the heading of the research refers to
quantization process is the approximation (according to the direction of price change). the nature by which prices tend to progress.
of numbers (the process of rounding-up The following example is a table Hence the issue of visually altering
or down numbers). By approximating any representing the steps used for the such nature will be fulfilled through the
one (or more) digit value to the nearest calculation and construction of the following subject matter (the second stage
decimal place, we are somehow creating application. Note that the percent change process).
two discrete categories or sets. is calculated to the nearest two decimal
Extreme quantization is one form of places.
quantization in which we limit the possible Column two represents a 10-day closing
values of a continuously progressive data values sample of an index.
in nature, and varying in magnitude to the Column three represents the change (in
least discrete set of values (or symbols) percentage) between each two consecutive
possible. closing values of the 10-day sample. The
results (as shown) are either positive,
Extreme quantization applied negative values of different amplitudes, or
to data (D closing prices): neutral (no-change).
As previously mentioned, the continuous Column four represents the process
movements or ∆ closing prices over time of categorizing the data of column three
are expressed by amplitude and direction of (the changes in closing values) into the
motion. Nonetheless, by applying extreme new sets previously referred to as UP,
quantization, these ∆ closing prices DOWN and FLAT. Each set is replaced by
become categorized with respect to the a letter (symbol), for which the letter “U”
direction of motion only. Such categories represents “UP”, “D” represents “DOWN”
(or sets) represent the least and only and finally “F” represents “FLAT”.

Table 1.

Days Closing Percent Extreme Extreme Accumulating

Values Change Quantization (in Symbols) Quantization column 5 The product of

1 1271.47

2 1279.24 0.61% U +1 +1

3 1340.77 4.81% U +1 +2

4 1340.72 0.00% F 0 +2

5 1392.27 3.84% U +1 +3

6 1400.55 0.59% U +1 +4

7 1397.07 -0.25% D -1 +3

8 1379.05 -1.29% D -1 +2

9 1421.84 3.10% U +1 +3

10 1421.80 0.00% F 0 +3

IFTA.org page 59
 IFTA Journal 2008 edition

Figure 1. USD vs. Japanese Yen, from August 2000, 2.2 Alteration of Price
to November 2002 Movement Dynamics on a Chart
As a method of pre-manipulation, the
quantization process applied to ∆ closing
prices provided a different interpretation
of data. When presented graphically, such
interpretation created an alternative means
of visualization. And the result was an
alteration in the price movement dynamics
of the original price chart. This result (or
by-product) represents the second and final
stage process.
The following chart (Figure 1) displays
a sample period from the US dollar vs.
Japanese yen currency chart (upper
window), accompanied by the application
(lower window).
The purpose of this example is to
graphically represent and explain how data
(closing price action) is interpreted after
being subjected to the pre-manipulation
process. Thus visually compare that
graphical representation to the basic price
chart.
As shown in Figure 1, the graphical
Figure 2. The NASDAQ composite, from September 1992 representation of the application (lower
to March 2000 window) displays the same progression
of the price action (upper window) in
synchronicity (with respect to time)
to the price chart. However, the visual
interpretation of the price action through
the application appeared somewhat
altered.

Part Three: Testing the

Application
Part three provides the results of using the
application on market averages and indices
as well as other major currency charts in
the foreign exchange market. This section
also intends to tackle the possibility of
implementing this application on individual
stocks and securities, as well as displaying
the results of such implementation.
The following basic TA tools will be
tested on the application:
• Trend and channel lines for both major
and minor trends.
• Continuation patterns, such as the
ascending, descending and symmetrical
triangles, rectangle formations, and flags.
• Major reversal patterns, such as the head
and shoulders formation, and double tops.
As a separate example, a final case will
be provided demonstrating how well this
application lends itself to the concepts of
support and resistance.
It should be noted that throughout this

page 60 IFTA.org
 IFTA Journal 2008 edition

research, the application will only be tested Figure 3. The XETRA DAX Index., from February 1996
on the daily end-of-day closing values or to September 2001
prices. Both line and bar charts will be used
to represent the daily price charts.
Within the examples presented in
this section, several cases were spotted
displaying divergence between the prices
and the application. Consequently, these
divergences were pin pointed and discussed
along with the basic TA tools tested on the
application.
During the analysis of the cases
presented in part three, the need for
applying similar basic TA tools to the
original price chart was required only for
comparative and supportive means.
The primary aim is to point out how the
nature, type, phase and duration of these TA
tools and patterns tested on the application
had varied from that of the basic price charts.
3.1 Trendlines and Channels
The following charts (figures 2, 3 and 4)
represent examples of long-term trend
lines, and channels.
Figure 2 displays a daily line chart of
the NASDAQ composite (upper window)
representing a period of the bull trend
from 1992 till 2000. This same trend was
visualized graphically on the application
(lower window) as a series of reaction lows
progressing upwards represented by a
major positive trend line.
Four black arrows are plotted on the
application indicating significant trend
line support points. Another four black
arrows are also drawn on the NASDAQ
Composite chart showing corresponding
points to those of the application. These
corresponding points are regarded as
major bottoms and/or critical levels of
the NASDAQ Composite during the major
long-term bull trend. Thus, such points
represent buying opportunities for the
market as suggested by the application.
Finally, the breakout from the major
trend line constructed on the application
also provided a decisive signal for the
termination of the bull market (see red
arrows).
The following example (Figure 3)
represents a comparison between two
major long-term trend lines regarding
time of duration, phase difference and
termination of each.
In this example, two major trend lines
are displayed on both the application
(lower window), and the price chart (upper
window). A phase difference of four

IFTA.org page 61
 IFTA Journal 2008 edition

Figure 4. The Egyptian Co. for mobile serv. (EMOB.CA),

from November 1998 to August 2000

Figure 5. The NASDAQ composite, from January to November 1993

page 62 IFTA.org
 IFTA Journal 2008 edition

months was spotted between the starting 3.2 Continuation Patterns

points of both trend lines constructed
on the application and the price chart (A) Triangle formations:
respectively. Figure 5 represents a case where an
Accordingly, a nine-month phase ascending triangle was recognized on the
difference was observed between the application (lower window). However, this
breakout of prices from one trend line triangle was formed during a testing period
(upper window), and the breakout signal of the NASDAQ composite to a critical
of the corresponding trend line of the resistance level in the market (upper
application. This breakout signal generated window).
by the application provided an earlier The breakout signal generated by the
warning to the termination of this major application and the actual breaking of
bull trend in the market the resistance level on the NASDAQ chart
Note that the bottoms occurring on occurred at the same time.
the trend line of the application (lower The ascending triangle (lower window)
window) were at different points in time provided a target projection equivalent to
from those marked on the line chart (see a value of 750 points on the NASDAQ chart
black arrows). Taken as an advantage, more (upper window).
buying opportunities (entry points) to the After the breakout signal from the
market are accessible. triangle, the subsequent trend progression
The following bar chart (Figure 4) of the application took the form of a
displays a local Egyptian telecom stock channel.
(upper window), accompanied by the In this example, a bullish divergence was
application (lower window). realized when prices fell from the previous
Figure 4 shows a major positive channel market high forming two consecutive
on the application (lower window), as well lower lows. The application formed a
as a major positive trend line on the price low followed by a second higher low. This
chart (upper window). divergence was confirmed later when
It was noticed that the time durations the NASDAQ Composite turned upward,
of both the trend line and the channel are forming new highs beyond the previous
somewhat equivalent (15 months). The market high.
phase difference between both durations,
however, was about three months.
Due to this phase difference, an early
breakout from the channel on late March
2000 was signaled on the application.
Subsequently, the actual breakout of
prices from the trend line (upper window)
occurred three months later (June), and
the termination of the bull trend was
confirmed.
During the bull trend, all peaks reaching
the upper channel line on the application
coincided with several critical peaks on
the price chart (see black arrows). These
peaks (lower window) provided short-term
selling opportunities, or reducing—position
signals to the previously held long
positions.

IFTA.org page 63
 IFTA Journal 2008 edition

The next example is of an issue traded

Figure 6. Arab International Contractors Co. (AICR.CA), in the Egyptian stock exchange against the
from September 1999 to August 2000 application.
In the above example (Figure 6),
a descending triangle formed on the
application (lower window), and a head
& shoulders formation (H & Sh) of similar
time duration formed on the price chart
(upper window).
Regarded as support levels, the flat
lower line of the triangle (lower window)
appeared corresponding to the neckline of
the H & Sh (upper window). This flat lower
line was precise and definite to a large
degree resulting in all the intraday price
violations occurring on the corresponding
neckline of the price chart to be not visible
on the application (see blue arrows).
It was also observed that the breakout
signal of the triangle occurred exactly at
the same time as the actual breakout of
prices from the H & Sh formation (see red
arrows). Nevertheless, the breakout signal
of the triangle was more decisive and clear.
Figure 7 represents a third case with a
triangle formation (symmetrical triangle)
Figure 7. NASDAQ 100 TRUST. (QQQ), from June 1999, observed on the application.
to December 1999 The time duration of the triangle (lower
window) occurred slightly before, and
continuing through a period of congestion
in prices (upper window) from mid August
till October.
The breakout signal generated from the
triangle (lower window) led prices to a new
high (target) almost at $65. This breakout
was followed by a series of higher lows on
the application but prices fell to a new low.
Hence a positive divergence was realized
and confirmed later when the prices turned
upward, breaking above the resistance of
the congestion area (upper window).
The breakout of prices from the
congestion area was confirmed three
days later by the breakout of the minor
resistance appearing on the application.

page 64 IFTA.org
 IFTA Journal 2008 edition

Figure 8. Cairo Housing Co. (ELKA.CA), from December 1999,

to August 2000

(B) Rectangle formation

The following example (Figure 8) shows a
daily bar chart (upper window) of an issue
traded in the Egyptian Stock Exchange,
accompanied by the application chart
(lower window).
Figure 8 shows a rectangle formation
formed on the application (lower window)
during a period of sideways movement on
the price chart (upper window).
The upper and lower boundaries of the
rectangle (lower window) appeared to
be corresponding to the resistance and
support levels observed on the price chart
and were well defined. For that matter,
any violation (or penetration) of prices to
the corresponding resistance (or support)
levels on the price chart were not confirmed
by the application (see blue arrows).
Another result due to the well defining
of the rectangle boundaries is that the
breakout signal of the rectangle appeared
very definite and clear, as well as early.
This signal was confirmed three days later
when a breakout of prices occurred from
the support level on the price chart (see red
arrows).

IFTA.org page 65
 IFTA Journal 2008 edition

Figure 9. EUR vs. GBP, from December 2002, to July 2003. (C) Flags (minor continuation patterns)
Figure 9 shows several examples of minor
continuation patterns (flags) appearing on
both the price chart (upper window) and
the application (lower window).
A similarity was noticed between
the duration time of all three minor
continuation patterns on the price chart to
the corresponding ones on the application.
A breakout of prices from flag formation
(1) (upper window) was followed by a
three-day delayed breakout signal from
the corresponding flag formation (1) of the
application. The delay serves as additional
confirmation to the actual breakout of
prices.
The breakout signaled from flag
formation (2) (lower window) occurred
in synchronicity with the actual breakout
of prices from the corresponding flag
formation (2) on the price chart. On the
other hand, the breakout signal observed
Figure 10. Cairo Housing Co. (ELKA.CA) from from flag formation (3) was regarded as an
August 2000 to July 2001. early, definite and clear one. Both signals
provided from the flag formations (2 & 3) of
the application suggest a confirmation and
an early indication to the price movement
(upper window) respectively.
Another bullish divergence was realized
in this example as prices broke out from the
final minor congestion area forming two
successive lower lows. Alternatively, the
application formed a low (bottom) followed
by another higher low. This divergence was
confirmed later in June when prices moved
upward forming a series of higher highs
and lows.

Figure 11. Sun Microsystems Corp (SUNW) from

June 1999 to April 2001.

page 66 IFTA.org
 IFTA Journal 2008 edition

3.3 Major Reversal Patterns Figure 12. Orascom Construction Company (ORTE.CA),
(A) Head and shoulders formation from February 2001 to April 2002.
This example provides a case at which a
head and shoulders pattern was formed
on both the price chart (upper window)
and the application (lower window). This
pattern appeared remarkably identical
to both (price chart and application) with
respect to time of duration, breakout, and
symmetry of pattern components (i.e. head
and left/right shoulders).

(B) Double top formation

Figure 11 shows a twelve-month major
double top formation (DT) on the
application (lower window). A similar
duration was observed between the first
and second peaks of the price chart (upper
window) and their corresponding double
tops appearing on the application. The
neckline of the DT formation corresponded
to the major support level of the price
chart.
The first top of the DT formation (lower
window) provided resistance to the second
top and as a result both tops of the DT Figure 13. USD -$ vs. Australian-Dollar, From March 1997 to
formation appeared equal in amplitude. December 1998
It should be noted that the breakout
signal from the neckline of the DT
formation occurred two days earlier than
the breakout of prices from the major
support level.

3.4 The Concepts of

Support and Resistance and
the Application
Figure 12 displays a daily price chart of
an issue (upper window) in a bear trend.
During this trend, four significant price
lows were recognized on the price chart.
Similarly four corresponding lows were
also observed on the application (lower
window).
An early breakout from the first low
of the application was signaled. Later, a
breakout in prices occurred, followed by
a minor up-correction on the price chart,
coinciding with the end of the pullback to
the first low of the application.
Notice how the price breakout from
the second low on the price chart was not
confirmed by the application, and prices
drifted away through a small price range.
It was only through the second breakout
signal generated from the application that
the prices actually broke out and fell within
the negative trend.

IFTA.org page 67
 IFTA Journal 2008 edition

Part Four: Discussion and Conclusion And all of which (one way or the other) • At certain times, observed through the
The final part of the research intends suggested useful trading signals. calculation process, price changes could
to provide a general discussion of a few be very minor, and of no true significant
basic characteristics of the application. An overview of the advantages provided by value. During such times, the application
An overall assessment of the potential the application: would interpret these minor (insignificant)
advantages and limitations of the changes as up or down moves, rather than
application will in turn follow this general • The usage of trend line and channel flat moves. In other words the application
discussion. analysis appeared to be of significant value. would regard them as significant changes,
There will be a discussion as to whether In general, the application succeeded in and therefore would appear to be deceptive
the application is considered to be a trend providing better visualization for major (or provide false signals).
leading or trend following device. long-term trends. One case displayed the
Finally, there will be an overview of the ability of the application to visualize a • The application does not respond
basic similarities and differences between significant long-term trend line within a sensitively to highly volatile (sharp), minor
the application and the Meisels indicator specific time period. Given this same time or sudden price movements. Consequently
followed by a conclusion to this research. period, however, the actual price chart through such movements, the application
could not support the construction of such does not usually provide timely (useful)
4.1 Characteristics or Features of the trend line. Another case displayed a major signals when tested upon with basic T.A.
Application (Advantages and Limitations): trend line constructed on the price chart tools.
In this section there will be an corresponding to a major channel on the
explanation of the general characteristics application which provided more useful • At rare times the application can
or features of the application. The features information. appear misleading. This limitation only
basically describe how the application occurs at times when prices rise (or fall)
reacts (its nature of motion on the chart) • The significance of various types in the direction of the trend. However
during the continuous price changes in bull, of pattern formations tested on the these moves (in the trend direction) are
bear and flat trends. application proved to be of crucial accompanied by high volatility and occur
One important feature of the application value. Some cases displayed an ability to over a relatively shorter duration than
is that it recognizes motion in terms of identify specific pattern formations on when prices correct from such moves (on
direction only and not amplitude. In other the application at times when no pattern lower volatility). It is only during these
words, it only reacts to the changes in formations could be recognized on the rare cases that the prices would progress
closing values regardless of the amplitude price chart. Other cases showed similar or to higher highs and lows (or lower highs &
of such changes. different pattern formations (of similar or lows) whilst the application would progress
Another important feature was noticed different durations and phases) appearing elsewhere.
through the process of construction of the on both the price chart and the application.
application. Within the calculation process, Nevertheless, these cases may describe
the application considers or interprets The definitive accuracy in the formation of the basic nature or constitute the main
flat (or no-change) movements between the trend lines and patterns appearing on form of price movements of one individual
any two consecutive closing prices as an the application lead to other advantages: issue (security), or—at least—during a
absolute zero value. In that sense, any certain time period of price movement
difference resulting only in an absolute • The precision of breakout points history of the issue. To reiterate from
(0.00%) value is regarded as a flat move, from trend lines, channels and pattern part three: this application was primarily
and accordingly any other result would be formations lead to significant (and valid) intended for the major market indices and
regarded as otherwise. buy/sell signals. averages that basically reflect the overall
As a result of that feature, a flat (or a major market trends so such cases seldom
no-change) movement between any two • Accuracy in measured and projected occur and are rarely observed on the major
closing values is rarely observed through targets market indices or averages.
the pre-manipulated data (after the
quantization process). Consequently, flat • A filtering method or technique provided
movements are seldom witnessed when the during periods where false breakouts
application is visualized graphically (on a (violations) or indecisive penetrations from
chart). critical levels would appear on the price
Through the different cases chart (such whipsaws would not be visible
previously presented in part three, on the application).
several advantages were observed on
the application. Some of which led to Limitations were also found using the
a visualization enhancement of price application:
movements on charts. Others provided a
confirmative view to price movements.

page 68 IFTA.org
 IFTA Journal 2008 edition

4.2 Is the Application a Meisels Indicator is to graphically display Bibliography

Trend-leading or when the ∆ closing prices are considered
Trend-following Device? to be overextended after a particular 1. Murphy, John J., Technical Analysis of the
Financial Markets, Prentice Hall, Revised
Throughout the cases previously presented move. However, this indicator regards Edition, 1999
in part three, a rather interesting feature all changes in closing prices also in terms
was noted on the application. It was of direction only. Therefore at a primary 2. Dorsey, Thomas J., Point and Figure
observed that—generally and in most stage, the Meisels Indicator assigns positive Charting, Wiley, 1995
cases— the application was considered to and negative values for each up and down
3. Hurst, J.M., The Profit Magic of Stock
be trend-leading in nature. However during trading day respectively. As a result, this Transaction Timing, Prentice Hall, 1970
other cases the application appeared to stage becomes identical to the process
be trend-following or perhaps sometimes of quantization of the ∆ closing prices of 4. Gifford, Elli, The Investor’s Guide to
trend-confirming in nature. the application. Then a moving period Technical Analysis, Pitman Pub. Ltd., 1995
The application appeared to be trend- (usually 10-days) is suggested in which
leading during cases where: the sum of newly assigned positive or
negative values within each moving period
• Earlier breakout signals were generated is calculated. Finally, the result for each
from the TA tools used on the application, period is plotted as either a positive or
as compared to the actual breakout of negative value on the chart.
prices, when such tools were applied By doing so, the Meisels Indicator
on the price chart. These early signals created a by-product. This by-product
provided early warnings and pre- produced graphically an interesting and
confirmations to the price breakouts accurate oscillator that signals when price
occurring on the price chart. moves have reached their extremes (with
respect to direction). Thus, this oscillator
• Positive (or negative) divergences completely eliminated the visualization of
occurred between the price chart and the trends appearing on the price chart. From
application, providing early warnings for a that, the basic discrepancy between both
reversal in price trend direction. the application and the Meisels Indicator
lies within the final representation of each
The application appeared to be trend- one graphically. IFTA
confirming or trend-following during cases
where:
• Breakout signals generated from the
TA tools used on the application occurred
during the same time, or later than the
actual breakout of prices, when these tools
were also applied on the price chart (given
the same time period under study).

4.3 The Application vs. the

Meisels Indicator
This section provides the results of a
comparison between the application and
the Meisels Indicator. The basic points
of comparison include the goals to be
accomplished by each of these two tools, as
well as the methods of calculation used for
each. Within this comparison, a few points
are worth mentioning:
The goal of the application is to visualize
the price movement over time in terms of
direction only, disregarding the amplitude
of the direction but most importantly,
without eliminating the visualization of the
trends appearing on the price chart.
On the other hand, the purpose of the

IFTA.org page 69
 IFTA Journal 2008 edition

Author Profiles

Technical analysis
has been developing Dr Benard Belletante Mohamed El Saiid
into more and more Bernard Belletante is Dean and Professor of Mohamed El Saiid is vice president and head of
rigorous, quantita- Corporate Finance of the Euromed-Marseille Ecole the technical analysis desk for HC Securities and
tive and scientifically de Management in Marseille, France. Dr.Belletante Investments (HCSI), Cairo, Egypt. He began his career
based approaches received his doctorate from the University of Lyon. working for Momentum Wavers, Ltd., a Middle East
and research meth- He has published numerous books and articles and technical analysis firm. Mohammed was awarded
ods over recent years. has headed the Financial Observatory of Medium- the Masters of Financial Technical Analysis (MFTA)
Gandolfi et al., p. 14 Sized companies in France. certification in 2004 and in the same year joined HCSI
as an Associate/Lead Technical Analyst. Later, he
Prof Stefano Caselli joined Unifund, a Geneva-based international private
Stefano Caselli holds a PhD in Banking and Finance fund as the Chief Technical Strategist and co-fund
Part one introduces
and is a Professor of banking and Finance at Bocconi manager to the Middle East Investments. Mohammed
a method which
University, Milan. He is a Director of Masters in is a technical analysis instructor for the Egyptian
proposes a new
International Management (MIM), Center for Capital Market Association (ECMA) and a member of
technical analysis
European Management Schools and he is a Director the Egyptian Society of Technical Analysts (ESTA).
(TA) tool. This TA tool
of Executive Education Custom Programs for the
is primarily applied
Banks and Financial Institutions Division at SDA Prof Gino Gandolfi
to market indices
Bocconi School of Management. Dr. Caselli has Gino Gandolfi, has a PhD in Business Administration
and averages, and
published papers and texts on private equity and from the “L. Bocconi” University-Pavia University,
is generally used for
venture capital, corporate finance and family and is a full Professor in the Economics Department
market forecasting.
business finance. at the University of Parma where he teaches
El Saiid, p. 32
the economics of the financial markets and
Paul F. Desmond asset management. He is also an instructor in
Paul F. Desmond has been the President of Lowry’s the fields of banking, financial institutions and
If you are not using Reports Inc, since 1972. Over the past 40 years he insurance companies at SDA Bocconi School of
a volume indicator has earned the distinction of being regarded as the Management. Gino has extensively published on
or are looking for a Dean of Supply/Demand Analysis. Paul Desmond was asset management, the financial services market and
unique new charting the recipient of the Charles H. Dow Award in 2002, competition in the banking sector.
method, then the for his original research entitled “Identifying Bear
concept of using Market Bottoms and New Bull Markets.” He is a Past Stewart Gault
color variations to President of the Market Technicians Association, and Stewart Gault holds a Bachelors degree
represent changes in a Founding Member of the American Association of in Engineering and a Master of Business
indicator values may Professional Technical Analysts (AAPT), wining its Administration. He has been awarded a Master in
seem practical. first Annual Award for “Outstanding Contribution to Financial Technical Analysis (MFTA) and is a Fellow
Gault, p. 32 the Field of Technical Analysis.” He has been featured of the Financial Services Institute of Australasia.
in interviews in Barron’s, The Wall Street Journal, CBS He is the Director of Operations at Olympus
MarketWatch, Money Magazine and a wide variety of Funds Management, a member of the Investment
other financial publications. Committee, and plays a lead role in the due diligence
and selection of investment managers. He is also
a member of the management committee of the
Australia India Business Council. Prior to joining
Olympus, Stewart Gault was a Research Analyst at
Lonsec Limited; and earlier to this position he spent
four years in Japan with a boutique consulting firm.

page 70 IFTA.org
 IFTA Journal 2008 edition

Since the crash of

Manfred Hübner Prof Henry (Hank) Pruden 1987, the picture of
Manfred Hübner is Head of Behavioral Finance and Henry Pruden is a leading technical analyst with the rational investor
a Fund Manager at Deka Investment in Frankfurt, more than twenty years of active trading experience. faltered… the idea
Germany. He manages equity and bond funds He is currently Executive Director of the Institute of efficient markets
for institutional and individual investors and of Technical Market Analysis and President of the with rational inves-
has seventeen years experience in the markets. Technical Securities Analysts Association of San tors could no longer
Manfred is a member of the Swedish society of Francisco (TSAASF). Hank Pruden is a professor be supported. The
Technical Analysts (STAF) and a former chairman of at Golden Gate University in San Francisco, where paradigm change is
the German Society of Technical Analysts (VTAD). he has taught technical analysis for thirty years. accompanied by the
Manfred is a proficient speaker and is well known He has also served on the board of directors of the rise and develop-
as a behavioral finance expert and appears regularly Market Technicians Association (MTA) and serves ment of a new scien-
in print and television media. as vice chair of the Americas for the International tific area, behavioral
Federation of Technical Analysts. Prof. Pruden is finance.
Cynthia Kase a member of AAPTA, The American Association of Hubner, p. 26
Cynthia Kase is accredited as a Chartered Market Professional Technical Analysts-USA.
Technician (CMT) and was awarded a Masters in
Financial Technical Analysis (MFTA). She is an Monica Rossolini The Wyckoff Law
engineer turned energy trader and risk manager. She Monica Rossolini is a PhD candidate in “Banking and of Cause and Effect
is president of Kase and Company, Inc. .a hedging and Finance” at Tor Vergata University in Rome and has successfully passed
trading solutions firm, primarily serving corporate a Laurea Degree. She collaborates and is involved in its market test of
and institutional energy. She is an author and was teaching and research activities in the Economics DJIA of 7,200 + 10%
the winner of the Market Technicians Association’s Department at the University of Parma. or 14,400 + 5%. We
(MTA) “Best of the Best” award in 1997. With decades anticipate that the
of energy hedging experience, Cynthia Kase carries DJIA will continue
out extensive ongoing statistical and technical Ing Antonella Sabatini to conform to the
research, which has helped her to become known by Antonella Sabatini, has a Masters degree in Science “Wyckoff Count
many as the energy market’s leading forecaster. and is Chief Executive and Financial Officer of Guide” and mid
Finbest - Family Office. She is a graduate of the 2007 is a “Stop, Look
George McLean Massachusetts Institute of Technology with a and Listen Point”
George MacLean was a Director of European degree in Electrical Engineering. She has held with three different
Technical Analysis for Standard and Poors MMS short-term teaching positions at Centro di Studi, scenarios for the
where he analyzed the European Cash Bond and Bancari Lugano, Switzerland and Bocconi University future price trend.
Futures Markets from a technical perspective. Mr. Milan, Italy. Antonella is a board member of IFTA Pruden and
McLean contributes to the Society of Technical involved in the Membership Committee. She is Belletante, p. 13
Analysts (STA) distance learning course and was a noted for her research and development projects
lecturer on the STA Diploma Course and the LSE/STA across many fields including Portfolio, Project and
Diploma Course. He has been an examiner for the Risk Management, Asset Allocation, and Technical
The financial markets
STA Diploma and Level II of the Certified Financial Analysis. She has written papers for various US
act as a natural
Technicians course and has written a book on and European journals. Antonella is a frequent and
system, in that
Fibonacci and Gann applications. respected speaker at conferences.
not only do prices
have a relationship
Dr. Rolf Wetzer
with time and with
Rolf Wetzer has a PhD in Finance and works as
each other but also
a senior portfolio manager with Bank Sarasin
display a mathemati-
in Switzerland, and teaches at TU Berlin and
cal relationship to
ESC Toulouse. He has extensively published on
recent highs and lows
quantitative trading strategies with a focus on
(harmony) using an
position sizing algorithms. He has won the German
inverse decimal value
Society Technical Analysts (VTAD) award in 2006
of harmonic ratios
and 2007, and is a frequent speaker at national and
(the size of the musi-
international conferences.
cal interval from one
note to another).
MacLean, p. 48

IFTA.org page 71
 IFTA Journal 2008 edition

Directors and Board

The International Federation of Technical Analysts Inc.

The International Federation Of Administrative Committee Chairs Directors At Large

Technical Analysts Inc.
Timothy Bradley (Tsaasf), Usa Gregor Bauer (Vtad), Germany
157 Adelaide St West, Suite 314 Conference Advisory Julius De Kempenaer (Dcta),
Toronto, Ontario M5h 4e7 Netherlands
Canada Simon Warren (Sta), United Kingdom
Finance Taichi Ohtaki (Ntaa), Japan
Email: [email protected]
Telephone: 1-905-854-6337 Hiroshi Okamoto (Ntaa), Japan
Fax: +1-416-849-0074 Alex Douglas (Stanz), New Zealand
Jorge Perez (Aeat), Spain
Membership & New Development
Antonella Sabatini (Siat, Samt), Italy,
EXECUTIVE COMMITTEES Switzerland
Public Relations/ Marketing
Elaine Long Knuth (Samt), Switzerland Ulf Sandberg (Vtad), Germany
Chairperson Jerry Butrimovitz (Tsaasf), Usa Bill Sharp, Immediate Past Chairman
(Csta), Canada
Henry Pruden (Tsaasf), Usa
Adam Sorab (Sta), United Kingdom
Vice-Chairperson - The Americas
Wang Tao (Tass), Singapore
Axel Rudolph (Sta), United Kingdom
Vice-Chairperson - Europe & Africa

Minoru Eda (Ntaa), Japan

Vice-Chairperson - Pacific Region

Peter Pontikis, (Stanz), New Zealand

Treasurer

Bruno Estier (Samt), Switzerland

Treasurer

Core Business Committee Chairs

Henry Pruden (Tsaasf), Usa

Academic Interface

Larry Lovrencic (Ataa), Australia

Accreditation

Saleh Nasser (Esta), Egypt

Data and Information Management

Claude Mattern (Afate), France

Education and Body of Knowledge

Regina Meani (Ataa, Sta), Australia,

United Kingdom
IFTA Journal

page 72 IFTA.org
KNOW
MORE
Bloomberg is the leading source of market data, news and analytics for financial and
legal professionals, corporations, media organizations and private individuals around the
globe. The BLOOMBERG PROFESSIONAL® service and Bloomberg’s media services
provide real-time and historical pricing, indicative data, reporting, research, estimates,
analytics, portfolio-management tools, electronic-trading and order-management systems,
multimedia events and electronic communications to Bloomberg customers in more
than 150 countries, 24 hours a day.
Visit the Bloomberg Stand at the 20th Annual IFTA Conference.
www.bloomberg.com

London New York Tokyo

+44 20 7330 7500 +1 212 318 2000 +81 3 3201 8900

©2007 Bloomberg L.P. All rights reserved. 26262779 0907