Debates between different forecasters and analysts about the methods

that they use usually boil down to a disagreement about what angle the
telescope through which they observe the markets and related events
should be tilted. But in his work Bob Prechter quite simply upends the
telescope; so that instead of events governing mood it is the mood that
governs events. Over the years Bob has established an international
reputation as a top technical analyst but more recently he has applied
the techniques that he uses to forecast trends in the financial markets to
predict a whole range of cultural and social trends. In April of this year
students at the LSE asked Bob to come over and talk to them about what
he calls socionomics. Unfortunately Bob was not able to schedule in a
talk to the STA but instead he has written a brief outline of some of the
ideas covered in his talk. It should be stressed that this is only an outline.
Anyone interested in pursing what is a very wide-ranging subject should
note that the Society has recently bought two books that deal with the
subject in depth, The Wave Principle of Human Social Behavior and
Pioneering Studies in Socionomics both these should be available in the
Barbican Library shortly. Alternatively, there is a lot of interesting
material on the Socionomics Institutes website (www.socionomics.org).
Annual Joint Meeting of the STA and
Society of Business Economists in Edinburgh
The STAs 3rd annual joint meeting with the Society of Business
Economists (SBE) was held in Leith, Edinburgh this year with John
Calverley from the SBE and Murray Gunn from the STA speaking. The
challenge was for the STA to find Value in investments and for the SBE to
Cut Losses in their trading. Murray robustly defended chartists ability to
find value by showing the benefits of using momentum indicators in
conjunction with trend analysis, while John highlighted the forward
looking nature of economic modelling and investment methodology,
stressing that if events change economists should be open to changing
their minds as well! Both were well received by the mixed audience and
the chair (Gerry Celaya) declared even honours.
STA members in Scotland should contact Murray Gunn for details of
planned meetings later in the year as we look to meet in Glasgow at
some stage, and are open to ideas of meeting further afield as well.
Elliott Day
Held in the Morris Lecture Theatre in the Robin Brook Centre at
St. Bartholomews Hospital on Friday, 2 July 2004, the Elliott Day was
a new venture for STA Education. It was designed as a serious look at
Elliott Wave Theory by leading analysts and academics for practitioners
and those who already have a grasp of the principles.
Bill Adlard opened with an in-depth look at the Dow. The relevance of
the 1929 Crash was immediately demonstrated and a clear argument
that the price action since then consists of five one degree lower sub
waves was convincingly made. A rigorous examination followed; 129
charts were presented. Finally, consideration of an even longer term
picture, from South Sea Bubble to date, completed the analysis.
Ron Giles of Queen Mary University, London opened the next session
with a brief review of current academic thinking. Andrew Pollock of
Glasgow Caledonian University demonstrated the use of directional
empirical probabilities in foreign exchange markets to conclude that
they worked better with Dow Theory than Elliott. Richard Ramyar
reported on research into Proportional Cyclical Dependence. The
terminology was very different from that used by technicians being
based largely on the concept of return rather than price action.
Peter Goodburn of Wavetrack International reviewed Gold. Again,
starting from the 1930s, a thorough analysis and precise use of
Fibonacci Relationship models led into forecasts up to the second
decade of this century.
Tony Plummer considered cycles, showed how bridging an energy gap
gives an Elliott count and illustrated cycles in US Economic Confidence
and US Output Cycle.
We hope to have a write-up of most of the sessions in the next issue of
the journal.
It was with great sadness that we learnt of the death of Elli Gifford
earlier this year. As well as being an immensely perceptive analyst, she
was a stalwart of the STA for many years. Her obituary appears on page
two of the Journal.
IN THIS ISSUE
Obituary Elli Gifford . . . . . . . . . . . . . . . . . . . . . . . . 2
R. Prechter Jr. The science of socionomics . . . . . . . .3
D. Watts Bytes and pieces . . . . . . . . . . . . . . . . . .6
J. du Plessis Optimisation from a technical
analysts point of view . . . . . . . . . . . . 7
R. Griffiths The Turning Tide . . . . . . . . . . . . . . . . 10
S. Griffiths Risk/reward trading using
Elliott Wave . . . . . . . . . . . . . . . . . . . . . 12
COPY DEADLINE FOR THE NEXT ISSUE
30th August 2004
PUBLICATION OF THE NEXT ISSUE
October 2004
FOR YOUR DIARY
8th September Monthly Meeting
13th October Monthly Meeting
10th November Monthly Meeting
8th December Christmas Party
N.B. The monthly meetings will take place at the
Institute of Marine Engineering, Science and Technology
80 Coleman Street, London EC2 at 6.00 p.m.
July 2004 The Journal of the STA
Issue No. 50 www.sta-uk.org
MARKET TECHNICIAN
MARKET TECHNICIAN Issue 50 July 2004 2
Elli Gifford, Fellow of the Society of Technical Analysts (UK), passed away
on April 15 2004.
Elli was one of the foremost technical analysts in the UK. She first
became known as a technical analyst after she went to work for
Investment Research (IR) in Cambridge in the mid 1960s, under the
leadership of Alec Ellinger and alongside other technical analyst
luminaries of the time, David Damant, John Cuningham and Harvey
Stewart. Previously she had passed Chartered Secretary examinations
and worked for Buckmaster and Moore, then a stock broking firm. She
briefly went to live in the USA in the late sixties, but returned to England
and IR in the early seventies. At that time she was also actively involved
in the production of their well known annual conferences, and was
instrumental in finding technical analysts from the USA to speak at those
conferences.
Later in that decade she left IR and joined Eurocommodities, where she
continued her increasing specialisation in the commodity and futures
markets. Disaster struck in autumn 1979 when the companys US parent
failed, but she was instrumental in the sale of the research department
to the large commodity broker, Rudolf Wolff. There she rose to be the
Director of Research, responsible for both the technical and
fundamental teams.
However, her old firm proved irresistible. In 1986 it had taken in new
investors and been incorporated as Investment Research of Cambridge
Limited, and its new MD, David Charters, asked Elli to return. It was an
offer she did not want to refuse, as East Anglia was her home area since
girlhood. There she wrote the Commodity, Futures and FX service, with
consultancies and frequent client visits in London. She told many stories
of her fellow travellers in the bar of the commuter trains from Liverpool
Street back to Cambridge in the evenings!
In all this time Elli was a well known and hardworking member of the STA
and, previously, ACTA, and a regular speaker at meetings. When the STA
was moving on from being the Association of Chart and Technical
Analysts to its current incorporated STA status, Philip Gray, then
Chairman, recruited her to the Committee to help with this transition.
This was a masterly move, for not only was she a well-respected analyst
but her Chartered Secretary experience was invaluable in dealing with
the process of incorporation. In further work for the STA, she was one of
the driving forces behind the IFTA conference held in London in 1989.
Although Ellis particular speciality was the futures and commodity
markets, she followed equities, spot currencies and interest rates as well.
Her all round knowledge led to two books on technical analysis. Her first,
Money Making Matters, is frequently mentioned by members as their
first introduction to literature on the subject. Her second, The Investors
Guide to Technical Analysis, for some reason never sold well. This is a
mystery to me as I regard it as one of the most comprehensive works
that has been written in the UK.
As a conference and seminar speaker Elli was much sought after, talking
about her methods and market views at venues all round the world.
Conferences she addressed included the STA and IFTA, amongst many
others, usually on her favourite subject of commodities and futures. In
addition, Elli was an active teacher of technical analysis, both in the UK
and elsewhere in Europe. Many people have contacted us at the STA to
say that she opened the door to technical analysis for them. There have
also been comments about how talking to her, or listening to her views
on the markets, helped to increase their interest. Both for her analytical
skills, and for the work she did to further technical analysts, both in and
out of the STA, she was created a Fellow of the Society of Technical
Analysts in 1989.
Those who knew Elli will also remember her whole hearted approach to
life, her slim tall figure, always elegantly clothed, and her immense sense
of fun. Technical analysis in the UK has lost one of its stars, and a
well-loved one. She will be sadly missed.
Anne Whitby, FSTA
CHAIRMAN
Adam Sorab,
Deutsche Asset Management,
1 Appold Street, London EC2A 2UU
TREASURER
Simon Warren. Tel: 020-7656 2212
PROGRAMME ORGANISATION
Mark Tennyson d'Eyncourt.
Tel: 020-8995 5998 (eves)
LIBRARY AND LIAISON
Michael Feeny. Tel: 020-7786 1322
The Barbican library contains our collection. Michael buys new books for it
where appropriate. Any suggestions for new books should be made to him.
EDUCATION
John Cameron. Tel: 01981-510210
George Maclean. Tel: 020-7312 7000
EXTERNAL RELATIONS
Axel Rudolph. Tel: 020-7842 9494
IFTA
Anne Whitby. Tel: 020-7636 6533
MARKETING
Gerry Celaya. Tel: 01561 340358
Simon Warren. Tel: 020-7656 2212
Kevan Conlon. Tel: 020-7329 6333
Barry Tarr. Tel: 020-7522 3626
Richard Raymar. Tel: 07890 821619
David Sneddon. Tel: 020-7888 7173
MEMBERSHIP
Simon Warren. Tel: 020-7656 2212
REGIONAL CHAPTERS
Robert Newgrosh. Tel: 0161- 428 1069
Murray Gunn. Tel: 0131-245 7885
SECRETARY
Mark Tennyson dEyncourt. Tel: 020-8995 5998 (eves)
STA JOURNAL
Editor, Deborah Owen. Tel: 020-7278 4605
WEBSITE
Gerry Celaya. Tel: 01561 340358
Simon Warren. Tel: 020-7656 2212
Deborah Owen. Tel: 020-7278 4605
Please keep the articles coming in the success of the Journal depends
on its authors, and we would like to thank all those who have supported
us with their high standard of work. The aim is to make the Journal a
valuable showcase for members research as well as to inform and
entertain readers.
The Society is not responsible for any material published in The Market
Technician and publication of any material or expression of opinions
does not necessarily imply that the Society agrees with them. The
Society is not authorised to conduct investment business and does not
provide investment advice or recommendations.
Articles are published without responsibility on the part of the Society,
the editor or authors for loss occasioned by any person acting or
refraining from action as a result of any view expressed therein.
Networking
WHO TO CONTACT ON YOUR COMMITTEE
Elli Gifford
July 1944 April 2004
Issue 50 July 2004 MARKET TECHNICIAN 3
Men have tried for millennia to forecast human events. In the long
history of social forecasting, the chronic propensity for immense error
has resulted from linear thinking, the extrapolation of current trends
into the future. This nearly ubiquitous approach is a result of the
assumption that laws governing billiard ball behaviour apply to
human behaviour. Most people believe that markets and societies
share the property of an object in motion i.e. it will continue along a
calculable path until some new outside influence a force or
obstruction alters its trajectory. Successful anticipation of future
events is possible. However, it is possible only with the knowledge
that human behaviour changes as a result not of external forces but
of internal ones. Understanding this dynamic is at the core of
socionomics.
What is socionomics?
Socionomics is the science of history and social prediction. It
examines and forecasts market and social trends from the perspective
that political, economic, cultural and financial trends are all the
product of the collective human psychology which stems from an
unconscious herding impulse in the pre-rational portion of the brain.
Understanding socionomics requires comprehending the contrast
between the following postulations:
1. The standard presumption: Social mood is buffeted by economic,
political and cultural trends and events. News of such events
affects the social mood, which in turn affects peoples penchant
for investing.
2. The socionomic hypothesis: Social mood is a natural product of
human interaction and is patterned according to the Wave
Principle. Its trends and extent determine the character of social
action, including financial, political and cultural trends and
events.
The contrast between these two positions comes down to this: The
standard presumption is that in the social setting, events govern mood;
the socionomic hypothesis recognises that mood governs events.
The Wave Principle
Underpinning socionomic
analysis is the Wave Principle.
This is an endogenous human
social dynamic that generates
a specific sequence of
progress and regress that
regulates the complex system
of social mood interaction.
It is not in the social nature of
mankind to accept and be
content with stasis. If there is
one constant regarding social
mood, it is its continuous flux.
However, the fact that social
mood is ever-changing is not,
as many would assume, an
impediment to forecasting; it
is the key to it. Investigation
by R.N. Elliott in the 1930s
and 1940s yielded the crucial
knowledge that social
behaviour changes not
randomly but according to a
hierarchal fractal pattern.
Social mood does have
constancy but it is a dynamic
one. While the extents of social mood, experiences and conditions
vary from time to time and place to place, the patterns of behaviour
that lead to a reversal in trend do not. In order successfully to
anticipate changes in society reliably, one must understand the
consistent pattern of societys internal dynamics.
The dynamics of the social mood
Many people assume that mood forms a feedback loop with events,
which in turn reinforces the mood. Some reflection reveals this idea
to be erroneous. If events formed a feedback loop with mood, then
social trends would never end. Each new extreme in mood in a
particular direction would cause more reinforcing actions and those
actions would reinforce the same mood, and so on forever. This is an
untenable idea.
The only feedback loop that must occur involves the synchronization
of Elliott waves through human minds. In order to participate in
social moods, individual minds must interact with others. All levels of
communication media, from face-to-face discussion to satellite
television, serve to effect this interaction. This interaction creates the
trends of social mood, which stimulate social actions, which are
reported as events. These actions and events are end results with no
consequences of their own in terms of the waves. This must be so,
because Elliott waves exist. Events affect minds to the extent that
they may shape specific actions that owners of those minds take, but
they do not alter or affect the trends in aggregate mood.
The sociological dynamic unfolds regardless of whether humans
create a gauge of it for observation and reaction, but when a gauge
(such as the Dow Jones Industrial Average) is created and widely
observed, people incorporate the gauge itself into the process of
mental interaction. It is a conscious reference point for the individual
participants and an unconscious reference point for the social
dynamic. Consciously rendered changes in the components of the
gauge (such as occasionally replacing one or two of the stocks in the
Dow) are irrelevant to the more important fact that the unconscious
dynamic uses the gauge as a reference regardless of its components.
Chart 1: Elections
The science of socionomics
By Robert R. Prechter Jr., CMT (compiled by Deborah Owen)
MARKET TECHNICIAN Issue 50 July 2004 4
People do not notice the operation of this sociological dynamic
because they are not looking for it. Indeed, given the opportunity,
they typically reject the idea outright because such dynamics are
contrary to the natural assumptions that people make about how
societies (and more narrowly, markets) progress and regress. That
these processes must be unknown to, or rejected by, virtually all is a
prerequisite for their operation. Only people who are blind to the
principles that govern the social dynamic can behave so as to
produce it, because only then can they be passionate, active
participants in it.
Because social mood change, as revealed by the stock markets form,
is patterned according to the Wave Principle, we can propose a larger
socionomic hypothesis, that the Wave Principle ultimately shapes the
dynamics underlying the character of all human social activity. We
will now investigate some presumed outside forces from the
socionomic perspective.
Elections
The standard presumption is that election results are a key
determinant of the stock markets trends. As an election approaches,
commentators debate the effect that its outcome will have on stock
prices. Investors argue over which candidate would likely influence
the market to go up or down. If so-and-so gets elected, it will be
good/bad for the market we often hear. If this causal relationship
were valid, then there would be evidence that a change in power
from one partys leader to another affects the stock market. On the
contrary, there is no study that shows a statistically valid connection.
A socionomist, on the other hand, can show the opposite causality at
work. Examine chart 1 and observe that strong and persistent trends
in the stock market determine whether an incumbent president will
be re-elected in a landslide or defeated in one. In all cases where an
incumbent was rejected by a landslide, the stock markets trend was
down. In not one case did an incumbent win re-election despite a
deeply falling stock market or lose in a landslide despite a strongly
rising stock market.
The social psychology that accompanies a bull or bear market is the
main determinant not only of how voters select a president but also
of how they perceive his performance. Correlation with the stock
market, consumer confidence, economic performance and other
measures suggests that social mood is by far the main determinant of
presidential popularity.
What a leader does is mostly acausal with respect to the publics opinion
of him. There are two reasons for this fact. First, his actions, despite their
endless analysis in the press, do little to affect his popularity. Second, his
popularity is dependent upon a social mood and economy over which
he can exercise no countertrend influence. If you are new to these ideas,
they may be hard to swallow. Arent some presidents fools or rogues
and others statesmen? Dont some presidents affect the economy for
good or ill? As to the first question, the answer is, certainly there are
presidents of high or low character and ability. However, that does not
affect their popularity. For example, President John Kennedy blew the
only military conflict in which he engaged the country, attacked the
steel industry out of pique to no result, and continually committed
adultery. He is revered. Why? Because the country was in a state of
euphoria for all but a few months of his term, euphoria that morphed
three months later into Beatlemania. The same may be said of President
Bill Clinton, whose escapades brought about impeachment yet whose
popularity stayed high despite his troubles, peaking along with the
stock market in the late 1990s.
As to the second question, Republicans claim that the laissez-faire,
low-tax policies of President Ronald Reagan were responsible for the
economic recovery of the 1980s. The Democrats claim that the social-
regulatory, high-tax policies of President Franklin Roosevelt were
responsible for the economic recovery that brought the US out of the
Great Depression. Can they both be correct? No, theyre both wrong.
The economy recovered each time because a rising social mood
drove the expansion. Policies did not cause the recoveries. Policies do
have effects, but they do not change the patterns of social mood.
Rather, changes in social mood affect policies.
War
The conventional view is that war affects social mood and the stock
market. But this assumption is unsupported by argument or history.
As to argument, many people assume that war is a dangerous
enterprise that would cause concerned investors to sell. Many
historians, on the other hand, argue that war is good for the economy,
which by conventional logic would make it good for the stock market.
As this reasoning is contradictory, so is the historical record.
By contrast, socionomics would argue that social mood governs the
character of social activity, so a persistently rising stock market,
reflecting feelings of increasing goodwill and social harmony, should
consistently produce peace while a persistently falling stock market,
reflecting feelings of increasing ill will and social conflict, should
consistently produce war. Chart 2 showing the US stock market
(spliced to the English one prior to 1789) over 300 years bears out
this thesis.
Major mood retrenchment produces war, as humans finally express
their collective negative mood extreme with representative collective
action. As with economic output, the size of a war is almost always
related to the size of the bear market that induces it. The three
biggest wars involving North Americans followed the three largest
stock market declines. The Revolutionary War began near the end of
the 64-year bear market in British stock prices that started in 1720.
The Civil War followed the 24-year bear market that ended in 1859
after posting a 74 per cent decline. World War II began six years after
the 89 per cent collapse in stock prices that bottomed in 1932.
Long rises in the stock market unerringly result in climates of peace,
while sharp declines result in major wars.
Demographics
What is the socionomic position on demographic causality? If social
mood determines the trend of the economy, politics and the
conditions for peace and war, might it not also determine
demographics?
Chart 3 shows the stock market prices plotted against birth rates
from 1909 to the present. There is a fairly noticeable correlation
between the two sets of data. Why would births and the stock
market trend together, if they do at all? Sometimes answers can be
found in subtleties. Notice that the deepest low in births this
century came in 1933, the year after the deepest low in the stock
market this century. Notice that the second most important low in
Chart 2: War
Issue 50 July 2004 MARKET TECHNICIAN 5
births occurred again in 1975, one year after the second most
important stock market low of this century. Why would there be a
one-year lag? Well, can you think of any activity that always
precedes a birth by about a year? If so, could this activity be
correlated directly with peoples moods and therefore the trend and
level of the stock market? We now have a tenuous basis for a
socionomic explanation for demographic trends. As people in
general feel more energetic, confident and happy, they conceive
more children. Conversely, as people in general feel more sluggish,
fearful and unhappy, they conceive fewer children. Thus social mood
determines aggregate procreational activity.
Films
The social mood percolates through to cultural trends as well. An
example of this can be seen in the production and success of certain
film genres.
The Walt Disney Company released its first feature-length cartoon in
1937, the year which saw the top of the roaring five-year bull
market that accomplished the fastest 370% gain in US stocks ever.
As can be seen from the titles listed on the top side of Chart 4, these
films stayed popular for 30 years, culminating with the ultra-sunny
Mary Poppins in 1964, and to a lesser degree, The Jungle Book in
1967. The end of this period of success was essentially coincident
with the great stock market top of 1966. For the next 16 years, as
stock prices fell along with social mood, most people thought
Disneys feature cartoons were silly and sentimental. Indeed, the
studios productivity fell by more than 50 per cent. Not one cartoon
film from this period is considered a classic. When the bull market
returned in the 1980s and 1990s, so did feature-length Disney
cartoons that have been both acknowledged classics and box-office
blockbusters. In the most recent 11 years of bull market, Disney
produced 10 feature cartoon films. In brief, Disney cartoons are bull
market movies, reflecting the shared mood of both their creators
and their viewers.
At the other end of the spectrum are horror movies. Some of the
most notable of these productions are shown on the bottom side of
Chart 4. Horror movies descended upon the American scene in
1930-1933, the very years that the Dow Jones Industrials collapsed.
Five classic horror films were all produced in less than three short
years. Frankenstein, Dracula and The Mummy premiered in 1931. Dr.
Jekyll and Mr Hyde was released in 1932, and King Kong in 1933, on the
test of the low in stock prices and right at the trough of the Great
Depression. These are the classic horror films of all time. Ironically
Hollywood tried to introduce a new monster in 1935 during a bull
market but Werewolf of London was a flop. When film makers tried
again in 1941, in the depths of a bear market, The Wolf Man was a hit.
From the latter half of the 1980s through the 1990s horror films
became increasingly derivative, muted or comic, just as in the years
following 1942.
The effect of social mood on cultural trends is not just restricted to
the US. Chart 5 (below) shows the explosion of horror films that were
released in Japan during its bear market.
Chart 4: Films
Chart 5: Japanese film titles
Chart 3: Demographics
MARKET TECHNICIAN Issue 50 July 2004 6
Music
A collector of old 78-rpm records writing in The Wall Street Journal
noted that music reflects every fiber of life in the US. Chart 6 shows
that popular musical themes have been virtually in lock step with the
social mood, as reflected by the major trends in the Dow Jones
Industrial Average.
A useful perspective
An individual who can rise above the social trends that sweep along
his fellows so that he can observe their operation has an incalculable
advantage over all those who do not. He has a basis upon which to
anticipate the future, not every time and not perfectly, but well
enough to have immense value. Anticipating the social future has
always appeared to be a gift unavailable to humanity. That has been
true until now, not because the task is impossible but because it
requires detailed knowledge of the patterns of social behaviour and
an ability to think and act in a fashion contrary to standard
assumptions.
Armed with knowledge of societys patterned behaviour, a wise
individual can train his reasoning to recognise emotionally driven
social thought processes and use that knowledge to make decisions
about how to harness or avoid coming trends in social mood and
social action.
Robert R. Prechter Jr is Executive Director of the Socionomics Institute
(www.socionomics.org)
Anyone looking for Elliott Wave Software should consider the
following alternatives:
Elliott Wave Analyzer III.
This program provides counts and projections of the waves. When
there is no count, no projection or count is presented. An interference
engine has modified the classic wave counts to enhance the accuracy
of the software. Provides a scanner to find specific wave counts.
There is a review of this software at
www.futuresmag.com/library/jan02/0102softwarereview.pdf
There is a free trial: Download from
http://downloads-zdnet.com.com/3000-2066-0218739.html
ElWave 7.1 a.
The Prognosis Software Company developed a modula program
approach. Real-time and EOD software available.
There is a review available from:
http://www.prognosis.nl/web2002/elwave/reviews/index.html
The program is generally well regarded, provides wave counts and
projections.
Again get the demo. Website: http://www.prognosis.nl/web2002/
Advanced GET now by Esignal.
Tom Joseph developed the first software to do wave counts, and this
is it. When a clear wave is present it accurately labels the wave. AGET
was one of the best and most user-friendly programs released in the
90s. It has a host of unique studies that help the trader. It does
re-label waves when the price is updated. The count can change.
Esignal have a module for real-time scanning. Most buy this program
for technical studies not only the Elliott Wave counts.
See http://www.esignal.com/advancedget/default.asp
Alpha Omega Elliot Waves
An addin for Metastock. Most other programs build upon the work
of Tom Joseph to provide wave counts. Generally these addin
products do not work as fast as standalone programs.
http://www.equis.com/Products/Catalog/Product.aspx?pcid=25&pid=1
135 Two other programs worth a mention are:
Market Predictor which finds and trades specific Elliott patterns ABC
patterns as developed by Steve Griffiths. A real-time version is now
available. See http://www.mtpredictor.com/
Dynamic Trader: Bob Miners time and price program allows the
User to label waves, but gives a time projection histogram of the
likely turning points for the next wave.
See http://www.dynamictraders.com
Real-time Quotes and Charts.
Its worth mentioning again as this must be a steal. Yahoo now offers
real time quotes for the NYSE/Amex and Nasdaq exchanges and
streaming real time charts with a real time stock screener. All for the
unbelievable price of $9.95 per month
There is a two-week trial available at:
http://billing.finance.yahoo.com/ym/f/FinanceReal1?.
Chart 6: Music
ANY QUERIES
For any queries about joining the Society, attending one of the
STA courses on technical analysis or taking the
diploma examination, please contact:
STA Administration Services (Katie Abberton)
Dean House, Vernham Dean, Hampshire SP11 0LA
Tel: 07000 710207 Fax: 07000 710208
www.sta-uk.org
For information about advertising in the journal,
please contact
Deborah Owen
PO Box 37389, London N1 OES.
Tel: 020-7278 4605
Bytes and Pieces
By David Watts
Issue 50 July 2004 MARKET TECHNICIAN 7
When the IBM PC arrived 23 or so years ago, many technical analysts
thought that its power would unleash great riches. The ability to
optimise and back test was thought to be the holy grail we had all
been searching for. Sadly it was not the case and optimisation was
abandoned by many in the 1980s.
A wise man once said that optimisation is just curve fitting and that
optimisation simply allows you to forecast the past with 100%
certainty. He was right. Optimisation takes a set of data and then
works out what would be the best way to have analysed it with the
benefit of hindsight, and if you apply those results to the same set of
data you will have, by definition, fitted the curve. He also said that
anyone stupid enough to optimise should optimise one section of data
and then test the results of the optimisation on a different section and
finally apply it to a third section. He was right about that too. His
arguments are nothing new. Its been heard before and will be again.
The problem is the wise man was a statistician, not a technical
analyst, and that is the traditional way that a statistician would look at
it. Statisticians do not understand the market or market charts. They
see price data as a time series of numbers. They do not understand
that the price data has life. They will only work on data that has no
autocorrelation (autocorrelation describes a condition where data
points in a time series are not independent of each other). If
statisticians find autocorrelation in a data series they will use
techniques such as first differences to try to eliminate it. They then
work on the resultant lifeless data.
Technical analysts on the other hand do understand the market and
market charts. They understand the charts are created by price data
and that the price data is created by human beings. They understand
that these human market
participants are subject to
human emotions which
affect the price and
consequently affect the
charts. Technical analysts
hope and believe that
autocorrelation exists
because if it did not and
todays price change is
independent of tomorrows
price, technical analysis
would be worthless.
Technicians realise that
because prices, and
consequently charts, are
created by humans, they
have human traits such as
trends and pattern
repetition. Technicians
realise that each instrument
has different characteristics
and so understanding the
movement of one, does not
mean that another is
understood. They
understand that the
characteristics of the price
movement are different
depending on whether the
price is in an uptrend,
downtrend or sideways
trend.
So, lets look at optimisation from a technical analysts point of view.
In doing so, lets stop to think for a moment what technical analysis is.
Technical analysis the study of price through the use of charts. Its the
study of the past in the belief that it can tell us something about the
future. Its the understanding that patterns in price and indicator
charts repeat. That is what chart reading is all about. Technical
analysts look for things in the chart that have proved reliable in the
past such has shapes, patterns, indicator movement and so on on
the premise that they will occur in the future and will therefore assist
us in making decisions. If you look back and inspect a chart and
notice that every time there is an inverted head and shoulders
pattern, the price has a significant rise, or every time the RSI breaks
through the 30 level and is rising its a good buy signal, isnt this
mental optimisation looking at past history and applying what you
have learnt to current data? Every time we look at a chart we are, in
effect, mentally optimising.
The problem with mental optimisation is that it takes a considerable
amount of time and it is possible that you could make a mistake. So
why not use mathematical optimisation instead? Its quicker and far
more accurate. To do this, you have to decide what you want to
achieve. Its a pretty good guess that profit is the motive, so you
could take a section of data and then test a range of conditions which
generate buy and sell signals and mathematically decide which set of
signals gives the greatest profit.
Lets take RSI for example. When Welles Wilder developed the RSI 30
years ago, he decided on a 14 day period. He also decided that a buy
would be indicated when the RSI broke above 30 and sell when it
broke down below 70. Presumably he did some mental (perhaps
even mathematical) optimisation in order to reach that conclusion.
Chart 1
Optimisation from a technical analysts
point of view
This is a summary of the talk given to the Society on 14th April, 2004 By Jeremy du Plessis , FSTA
MARKET TECHNICIAN Issue 50 July 2004 8
How else would he have arrived at those parameters? But do they
apply to all instruments all the time? What if the 12 day RSI gave
better results when breaking above 22 and below 75? How would
you know? Manual inspection of every chart would be tedious and
lead to errors.
You could try it mathematically, by trying every RSI period from, say,
5 to 50. Then for each
period, you could test
buying at a break up of
every level between 5 and,
say, 95 stepping up 1 at a
time. You could then test
selling at a break down
from every level stepping
down 1 at a time. Its a
complex test. For example,
the optimisation would
start with a 5 day RSI and
then record a buy when
the RSI breaks up through
5. Once it has done that, it
would have to test a
breakdown level starting
with, say 95, then 94, 93
and so on, recording the
gains or losses from each
level. Then it would move
to the next buy level and
perform the same tests
again, and so on. We would
have to test every buy level
and every sell level for
every RSI period and
eventually we would find
the combination that
yielded the greatest profit.
Unfortunately it doesnt
stop there. We have made
the assumption that buying
and selling on a break of a
level is the best way to read
the RSI, but maybe it isnt.
Perhaps it would be better
to buy on a break up of a
level but to sell when the
RSI crosses down through a
moving average. But what
moving average? We would
have to test all moving
average periods. Perhaps it
would be better to buy
when the RSI crosses the
moving average and sell on
a level. Perhaps it would be
better to buy when the
moving average of the RSI
breaches a level. We will
not know which RSI period
is best and which is the
best combination of entry
and exit signals until we
have tested all of them.
Having done many of these
tests, one thing becomes
very apparent and that is,
exit signals must be
guaranteed. Welles Wilder
suggested an exit when the
RSI breaks down through
the 70 level. It usually
generates a very good exit signal, but what if, after giving a buy
signal, the RSI never goes above 70 and the price falls. You have no
way of exiting your trade.
Chart 1 (on the previous page) shows the FTSE 100 index with a 14 day
RSI using Wilders buy and sell criteria. It works fairly well until 2000
when a buy signal was generated because the RSI broke up through
Chart 2
Chart 3
Issue 50 July 2004 MARKET TECHNICIAN 9
30. The index then fell over 40% before generating an exit in 2004
when the RSI did finally go above 70 and break down through it. In
fact during the bear trend of 2001/2002 the RSI fell below 30 and
generated repeat buy signals which were not exited until 2004.
Immediately the importance of guaranteed exit, no matter what the
conditions are, is paramount.
The only exit that can be guaranteed is a trailing stoploss. But
knowing what percentage stoploss to use can only be done by
optimisation. So an optimisation exercise could be performed that
testing every single entry condition together with a range of
percentage trailing stoploss exits. Chart 2 shows Wilders standard 30
entry signal with a 19.5% trailing stop giving the best exit.
For a long term investor, a 19.5% stop loss is quite acceptable but not
for a short-term trader. Simply reducing the maximum allowable
stoploss to, say, 10% will not solve the problem, because patently the
best stoploss is greater than 10%. There is another technique
however which impacts on the stop loss percentage, that is signal
delay. Signal delay means that trading on a signal is delayed by a pre-
determined period in order to reduce the chance of a whipsaw. The
strategy is to wait a number of days after the signal and only act if,
after waiting the required period, the signal is still in place. This
avoids the 1 day spikes through the stoploss. Signal delays are
designated (t+1), (t+2), (t+3) etc., where t is the day of the signal. By
definition therefore it cannot be acted on until day (t+1) because by
the time the signal is generated on day (t), the market has closed.
(t+2) means do not act on day (t+1) and if at the close of day (t+1) the
signal has not been cancelled, then act on day (t+2).
Knowing what delay to impose is part of the optimisation too. It is to
do with the characteristic, some would say, the volatility of the
instrument. Taking the 14 day RSI again with a 30 level entry signal
and testing various stoploss exits with various signal delays, the result
is a 9% stoploss with a (t+3) signal delay as shown in Chart 3. Notice
the spike through the stop loss in March 2003 which did not trigger a
sell signal because the delay is (t+3).
But what if there is a better entry point than a break up through 30?
Going back and checking by optimisation will answer the question.
The point is, you have
absolutely no idea which
has been the best entry
method in the past without
running an optimisation.
Chart 4 shows an
optimisation conducted on
the FTSE 100, testing a
range of RSI periods with a
full range of break up levels
together with a range of
stoploss levels. The result is
a 5 day RSI breaking up
through 16 as the entry
and an 8% trailing stoploss.
Both signals have a (t+2)
signal delay imposed.
Mathematical optimisation
is good. It does tell you
things you would not
otherwise know, but what
about the second thing the
wise man said about
optimising and testing the
result on different sections
of data? An optimisation to
find the best entry and exit
signals could be run, but on
what data? Technical
analysts know that market
conditions continually
change and what worked last year in isolation may not work this year.
This writer believes that it is a waste of time to randomly select a
section of data, optimise it and then apply the results to another set of
data. There is little chance of it being consistent.
The market is a moving target, therefore it is important to take as much
of the latest data into account when optimising, but it is important to
re-optimise regularly as new data is added. Traditionalists (and
statisticians) will throw up their hands in dismay but, if it is done
sensibly and with a few rules, it does seem to produce the best
ongoing results.
The first thing to do, is to decide on ones time horizon. There is no
point finding which was the entry and exit that worked best over the
last 20 years if you are a short-term trader. The result is a compromise
which caters for all market conditions. If you are a long-term investor,
then of course, the more data, the better. Short-term traders need to
consider how much market information they require to assess a
signal. This may be six months, one year or two years. The shorter the
trading horizon, the less data should be used in the optimisation. This
is because short-term traders require the optimisation to squeeze as
much profit out of the data as possible, which means going for the
smaller trades. The next thing to consider is dealing costs or the
penalty for trading. If there was no penalty, the optimisation would
consider every gain, no matter how small, as a profit. For longer term
optimisation, something like 2% entry and 2% exit commissions
should be applied. For shorter term, these can be reduced. Finally
and most important, because the optimisation is conducted on the
latest data, it is important to ignore any open trades in the result. This
means that any open trade will be as a result of optimisations prior to
the trade and is therefore not included in the total profit which
determines the best technique.
Chart 5 shows an optimised RSI of BG Group. The result is a six day
RSI where entry is a break up through 20 and exit is a 9% trailing
stoploss on the price. Both are subject to a (t+3) signal delay.
Notice that the last trade is still open, which means the best entry and
exit conditions were determined without taking any data to the right
of the last entry signal into account.
Chart 4
MARKET TECHNICIAN Issue 50 July 2004 10
From an historical point of view, it is highly unlikely that the top to a
bull market in the US would form during the run-up to a presidential
election with the economy still strong. This four-year cycle is
powerful. The 10-year rhythm is even more dominant. A recession
always comes in the first two years of any decade. The recovery
follows in the third and fourth. It is relatively unusual to have a
major top in the fifth year it is normally later but this time we
believe it will happen. Markets are now about to run higher,
breaking above the
resistance levels in place
since March. The high of the
cycle, which bottomed in
October 2002, is due in
March 2005.
Gaining interest
The whole cycle is driven by
interest rates. These are
already trending higher but
will not yet cool the
economy. A series of
modest, or measured, moves
is likely to take some time
before it tips the market over
into the next bear period.
Between March and June
2005 we think that bonds
will become a better buy
than equities. As soon as it is
clear that the hikes have
worked, the Fed will
probably move to soft land
any recessionary tendencies.
The optimisations shown
here apply to RSI only, but
optimisation can be
extended to any technical
indicator and chart,
including Point and Figure. It
can be applied to all time
frames. It can be applied to
short trades as well as long
trades and it is interesting to
see that the best exit out of a
long trade is not the best
entry into a short trade.
Optimisation is not the holy
grail, but what it can do is
tell you how you should be
reading indicators. In doing
thousands of optimisations,
a number of valuable
lessons have been learnt.
The best exit is usually a
different type of signal from
the best entry. Exit signals
need to be guaranteed or
they may never be
triggered. A trailing stoploss
is one of the best and most
reliable exit signals and
imposing a signal delay to
both entry and exit signals
does improve the results
considerably.
Constant optimisation of the same indicator and instrument will
adjust the results as the characteristics of the price change, but
ignoring the last open trade ensures that the optimisation is not
meaningless. Optimisation can be dangerous in the wrong hands but
if used correctly and wisely, it can teach you a lot about the indicator.
Chart 5

Chart 1: World Equities
The turning tide
By Robin Giffiths, FSTA
Issue 50 July 2004 MARKET TECHNICIAN 11
US stock market
Borrowed money left the market, partly encouraged by the Fed and
partly by the cooling of the Chinese economy. This was a severe test
for the markets. It could have broken them into bear patterns, and
some thought that had happened but it was not so. On the charts
the test was passed and the continuing uptrend confirmed.
Accidents aside, the way to bet is that the broad market will rise
through the period of the presidential election and whoever wins will
have at least a brief honeymoon period. The odds are then equally
good that a bear will start next year, when interest rate hikes finally
bite and cause the US consumer to desist and pay back debt. Our
road map indicates that the S&P index could hit 1350 by March 2005.
Then a switch back into bonds which will be yielding much more
than now will be the preferred strategy. If it were not for
unquantifiable geopoliticial risks, we would be raising our targets on
the strong earnings numbers. As it is, we err on the side of caution.
The Nasdaq index is on the flat road map, in secular downtrend. We
do not expect new highs. Any rally from here will be short-lived and
then many hedge funds will be shorting it. The broad market indices
like the S&P, and especially the Value Line and Wilshire 2000, should
outperform.
With the economy running strongly, it would be very unlikely to find
that the US stock market had topped out in the run-up to a
presidential election, even given the undoubted geopolitical risks.
Robin Griffiths is senior technical analyst at HSBC

Chart 2: US treasury bond yields

Chart 3: S&P 500

Chart 4: Dow Jones


Chart 5: Nasdaq

Chart 6: Wilshire
Chart 7: Value line
MARKET TECHNICIAN Issue 50 July 2004 12
Many traders regard Elliott wave as a technique that only projects
future movement by knowledge of where in a current pattern you are
at any particular time. However, because of the unpredictable nature
of the markets, many Elliott wave forecasts have been less than
reliable, causing many people to question its ability to forecast future
movement.
However, it is my belief that this is looking at Elliott wave in the wrong
light. What you should be asking is whether Elliott Wave (or parts of it)
can be used to identify trade opportunities that allow the trader to
enter a trade with a small controlled risk in relation to the profit
potential on a new trade. This is very different from asking where will
a market be at some point in the future,
and then getting disheartened when it
does not get there!
I have identified one specific Elliott
wave pattern that allows the trader to
enter a trade with a small controlled
risk. This is the simple ABC correction.
This simple pattern has many other
benefits which I have discussed in prior
articles, so in this one I would like to
focus on the ability of this pattern to
allow trades to be identified in which
the initial risk is kept small in
comparison with the potential profits.
Over time this allows profits to be
larger than losses which, as all traders
know, is the fundamental building
block for a successful approach to
trading the markets.
Lets look at an example of a daily chart
of the US stock CCI (chart 1). Here is a
recent trade automatically found by
MTPredictor (the software I have
designed to identify these specific
trade set-ups), and was reported at the time in the
MTPredictor daily report. It has not therefore been
chosen with the benefit of hindsight.
As you can see, the end of the ABC correction was
identified allowing a trade entry just as the Wave C
low was ending, thereby keeping the initial risk small
at only 0.69 points (13.67 12.98). A few weeks later
CCI had reached the profit target where the profit
was 3.43 (17.10 13.67) or nearly 5x the initial risk
required to take the trade, thereby making the profit
large in relation to the initial risk required to take
the trade.
This is how you should view profits, not just in pure
dollar terms, but as a function of the initial risk that
was required to take the trade. In this way you can
control your trades.
Obviously not all trades are winners; it would be
foolish to think so. But keeping such losses small,
when they happen, is one of the most important
techniques that should be mastered when trading.
There are too many traders that focus on the
technical side of mastering analysis techniques and
forget that trading is all about money management.
I hope you are beginning to see how if you can only
identify trades in which the initial risks are small in relation to the
potential profits, then it puts you on a solid foundation for a
successful approach to the markets.
Chart 2 shows another example on a 5min chart of the Dow Jones:
Again, the same basic principle was applied, looking for trades where
the initial risk (or loss) is small in comparison to the potential (or final)
profit.
In this example the eventual profit was approximately 9x the initial
risk required to take the trade (ignoring slippage and commission).
The initial set-up was off the same ABC pattern as the CCI example
earlier. However, the most important point is the size of the profit,
again, not in dollar terms, but in relation to the initial risk required to
take the trade.
Steve Griffiths is managing director of MTPredictor Ltd, and has
developed his own unique software. For more information please visit
www.MTPredictor.com or email Steve@MTPredictor.com
Risk/reward trading using Elliott wave
By Steve Griffiths
Chart 1
Chart 2