Issue 4 - The Technical Analyst

Transcription

may 2004
www.technicalanalyst.co.uk
Brent Crude
The slippery climb upwards
Andrews’ Pitchfork
Talking to Ralph Acampora
The day-of-the-week effect
Why traders should look at
what prices aren’t doing
On getting TA the
respect it deserves
Monday blues and
sunny Fridays
Advertisements - STA and Paritech
WELCOME
Editor: Matthew Clements (MSTA)
[email protected]
Managing Editor: Jim Biss
Editorial Board:
Mikael Bask, Umea University, Sweden
Tai-Leung Terence Chong, The
Chinese University of Hong Kong
Marketing: Vanessa Green
Sales: Christopher Leigh
Design: Paul Simpson
The Technical Analyst is published by
Clements Biss Economic Publications Ltd,
10-12 King Edward's Road,
London E9 7SF
Tel: +44 (0)20 8533 3025
Web: www.technicalanalyst.co.uk
Email: [email protected]
SUBSCRIPTIONS
Subscription rates UK: £275 per annum
Rest of world: £325 per annum
For information, please contact:
[email protected]
ADVERTISING
For information, please contact:
[email protected]
PRODUCTION
Art, design and typesetting by
all-Perception Ltd.
Printed by The Friary Press
ISSN(1742-8718)
C y c l e s a r e c e n t r a l t o t h i s i s s u e o f T h e Te c h n i c a l A n a l y s t . S o m e
c y c l e s , s u c h a s t h e d a y - o f - t h e - w e e k , a r e e a s y t o ta k e o n b o a r d .
A ft e r a l l , i t ' s n o t m u c h o f a s u r p r i s e t o f i n d o u t t h a t , i n g e n e r a l , t h e
m a r k e t s a r e i n a b e t t e r m o o d o n F r i d a y t h a n o n M o n d a y.
Researchers from the National University of Singapore find evid e n c e f o r t h i s e ff e c t i n t h e s t o c k m a r k e ts o f A s i a , b u t a l s o g o f u r ther in trying to identify whether some weeks are gloomier or happier than others and in determining what the relationship is
b e t w e e n t h e M o n d a y a n d t h e p r e c e d i n g F r i d a y.
S o m e c y c l e s i n v o l v e a r h y t h m t h a t m a y h o l d s o m e ps y c h o l o g i c a l
a n d e n v i r o n m e n ta l u n d e r p i n n i n g , s u c h a s t h o s e i m p l i c i t i n t h e f o u r w e e k r u l e , b u t w h i c h a r e h a r d e r t o e x p l a i n f u l l y. I n o u r a r t i c l e o n t h e
four-week rule, the author provides some practical advice on how to
shore-up this general phenomenon with practical rules and complim e n ta r y a n a l y s i s .
F i n a l l y, w e ta k e t h e c y c l e t o t h e f r o n t i e r s o f c o m p r e h e n s i o n w i t h a n
article on Bradley's siderograph, a means of using astrology to
forecast mass market highs and lows. Despite any natural scepticism for this subject area, we include this article because the publ i s h e d r e s u l ts a r e i m p r e s s i v e a n d s h o u l d b e s c r u t i n i s e d f u r t h e r.
Also in this issue: a practical article on using Andrews' Pitchfork
a n d n e w r e s e a r c h o n n o n - l i n e a r i t y i n t h e s t o c k m a r k e ts a n d f o r e i g n
exchange rates of South-East Asia. In the latter article, the authors
characterise the type of non-linearity present and go on to suggest
w h i c h TA t e c h n i q u e s s h o u l d w o r k b e t t e r t h a n o t h e r s a s a r e s u l t .
I f y o u h a v e a n y c o m m e n ts o n a n y a s p e c t o f T h e Te c h n i c a l A n a l y s t
or just have something to say on the subject, please email them to
[email protected].
© 2004 Clements Biss Economic
Publications Limited. All rights reserved.
Neither this publication nor any part of it
may be reproduced, stored in a retrieval
system, or transmitted in any form or by
any means, electronic, mechanical, photocopying, recording or otherwise, without
the prior permission of Clements Biss
Economic Publications Limited. While the
publisher believes that all information contained in this publication was correct at the
time of going to press, they cannot accept
liability for any errors or omissions that
may appear or loss suffered directly or
indirectly by any reader as a result of any
advertisement, editorial, photographs or
other material published in The Technical
Analyst. No statement in this publication is
to be considered as a recommendation or
solicitation to buy or sell securities or to
provide investment, tax or legal advice.
Readers should be aware that this publication is not intended to replace the need to
obtain professional advice in relation to
any topic discussed.
Matthew Clements
Editor
April 2004
THE TECHNICAL ANALYST
1
CONTENTS
04
30
41
43
44
48
2
Product News
The Technical Analyst Talks To...
Ralph Acampora, Managing Director, Prudential
Equity Group
Book Review
Advanced Swing Trading by John Crane
Letters
The dangers of data snooping
Commitments of Traders Report
Training & Events Diary
April 2004
MAY 2004
06
12
32
Market Views
06
A head-and-shoulders pattern in USD/JPY
08
Brent crude
10
Making waves in the US dollar, Nikkei and
fixed income markets
Techniques
12
Andrews’ pitchfork – the price failure rule
14
The four-week rule
18
Using the McClellan Oscillator
22
Volume spikes and index reversals
26
The inverse fisher transform
28
Astronomy and the Dow Jones
Subject Matters
32
Monday blues and sunny Fridays
34
Why have the returns to managed futures
funds decreased?
38
Nonlinearity favours nonlinear TA techniques
May 2004
3
Product News
MarketWatch.com launches
real-time ser vice with eSignal
ChartFilter
announces integrated Stock
To o l s s o ft w a r e
ChartFilter has told The Technical Analyst
that its new analytical software, Stock Tools,
will be available by June 2004. It will comprise six integrated tools that can be downloaded and installed from their website
MarketWatch.com, Inc., a leading
multi-media publisher, has teamed
up with eSignal to launch CBS
MarketWatch LIVE, an online
streaming, real-time news and market data product that is also compatible with portable handheld devices.
This new service provides tick-bytick quotes from more than 75 global markets, integrated news and
commentary, streaming charts, and
market depth data. Subscribers also
have use of more than 15 analytical
studies as well as market scanners
with customizable searches that
screen the market for buy and sell
opportunities.
(www.chartfilter.com). These are: advanced
technical charting; an alert system that allows
you to create complex, staged alerts using
technical indicators and fundamentals on
multiple stocks; screening and backtesting;
fundamentals to export or print; and a portfo-
CBS MarketWatch LIVE is available
at www.marketwatch.com/cbslive
from $14.95 per month, plus
exchange fees and fees for any addon services. A free 30-day trial is
available.
lio manager. "We wanted to make it easy for
technical analysts to get the most out of their
own analytical abilities and imaginations,"
said the company CEO, Doug Hubscher.
The monthly or yearly subscription includes
end-of-day data (ComStock) as well as auto-
www.marketwatch.com
www.esignal.com
matic upgrades.
eSignal and FXCM create
integrated FX trading solution
eSignal and Forex Capital Markets
(FXCM), a leader in online FX trading, have come together to offer a
new feature that allows traders to
monitor market activity, identify trading opportunities and execute trades
from a single, integrated platform.
The feature integrates eSignal's
streaming market data, charting
package and customizable formulas
with FXCM's direct access execution abilities, thus eliminating the
need to switch between applications
to monitor and execute Forex transactions.
With eSignal and an FXCM account,
4
eSignal users may now: execute FX
trades on FXCM's trading platform
via a direct link from eSignal's market data and charting application;
send market orders directly to
FXCM's trading platform; obtain
split-second trade executions from
FXCM; and monitor open positions
via a direct link from eSignal to
FXCM's trading platform
"FX trading offers 24-hour trading,
transparent pricing, and low transaction costs. As a result, the active
trading community is embracing FX
trading in record numbers and looking for high-quality brokerage and
analytic products," said Drew Niv,
May 2004
chief executive officer, Forex Capital
Markets. "Our relationship with
eSignal gives FXCM subscribers a
consistent means for not only back
testing diversification strategies, but
also for charting, analyzing and
facilitating execution of timely Spot
FX trades."
Access to FX trading through FXCM
is available to eSignal clients who
have a brokerage account with
FXCM. Interested clients should visit
www.esignal.com. To open a brokerage account with FXCM, visit
www.fxcm.com
Product News
Pronet Analytics.com
set sails for China
Pronet Analytics.com Limited has
agreed terms for distribution of its
service to clients of FXCM Asia Ltd
in Greater China. Under the terms
of the deal, FXCM Asia will offer
Pronet's FX research service to new
and current clients in the region. The
commercial terms are structured so
that FXCM Asia will continue to offer
what are believed to be the most
competitive dealing spreads in the
region and the client will pay no
additional fees for access to
Pronet's research.
Shane Smith, Group CEO of Pronet
Analytics.com, told The Technical
Analyst:
"Access to the China market is a
very exciting development for the
Group. China presents an opportunity for many US and European
firms generally, but it is worth enumerating how large that opportunity
might be for Pronet Analytics.com:
FXCM is the leading non-domestic
FX online broker for mainland
China, serving also Singapore and
Taiwan, and since opening its office
in Hong Kong in March 2003, has
seen month-on-month growth in
accounts of 150%. China is now the
fourth largest trading nation in the
world, which itself creates a huge
FX requirement, but in addition
overseas funds are being repatriat-
ed at an unprecedented rate, putting
pressure on the Yuan. Irrespective
of whether the Yuan itself eventually
floats, individuals and companies
are anyway allowed to trade foreign
currencies, and also have a famous
propensity to speculate. Internet
penetration is growing, and with 80
million current subscribers China is
second only to the US and is on
course to overtake it as the most
wired nation on earth. All this points
to a large and growing demand for
web-delivered research services to
assist in better decision-making."
www.pronetanalytics.com
U pd a ta
Proquote supplies
500 screens to TD Waterhouse a n n o u n c e s
c o m pa ta b i l i t y
w i t h M y Tr a c k
The London Stock Exchange has
announced that Proquote, its trading
and market data business, has
secured its biggest deal to date with
an agreement to supply 500 screens
to TD Waterhouse. TD Waterhouse,
one of the UK's largest execution-only
brokers, has ordered the screens to
support the launch of ProTrader™, its
integrated market data and dealing
service for retail investors.
After announcing Bloomberg
Compatibility last month,
Updata has announced that its
Technical Analyst software is
now compatible with the
MyTrack data feed.
To coincide with this, Updata
has also launced a US site for
both Bloomberg and MyTrack
users.
The new order follows one for 125
screens in September 2003, and
means TD Waterhouse will be
Proquote's largest customer, accounting for just over a quarter of all
Proquote's installed screens.
www.updataTA.com
May 2004
5
Market Views
A HEAD-AND-SHOULDERS PATTERN
IN USD/JPY by Kevin Edgeley
he recent US dollar recovery has prompted concern of a
more significant reversal of the long-term dollar bear trend.
A look at the dollar index chart (Figure 1) shows the market
probing above the long-term bear channel line from early
2002. As yet, this potential long-term trend reversal is not mirrored in the USD/JPY chart (see Figure 2), but there are certainly signs of further near-term corrective strength before the
dollar resumes its path lower. Other major crosses (EUR/USD
and USD/CHF) have traded dollar positively through their 200day moving averages. Weekly momentum indicators (stochastic and MACD) for USD/JPY have turned higher showing
divergence at the recent lows. Whilst this pattern is a warning
sign of loss of trend strength, one would normally want to see
a trendline break to confirm a reversal. In USD/JPY the dom-
T
Figure 1.
6
May 2004
inant trendline from February 2002 is some way above current
prices.
The long-term pattern in USD/JPY shows a head-and-shoulders which developed from the left shoulder low in September
2001 at 115.50 through the head in January 2002 at 135.20 to
the extended right shoulder base formed with a long-term
descending triangle from July 2002 to September 2003. The
flat neckline was broken in September of last year and was followed by a large weekly breakaway gap after the Dubai G7
meeting. Subsequent price action formed a declining wedge
over the ensuing 5 months. The Bank of Japan was actively
intervening during this period to limit yen strength which eventually led to a sharp recovery and a move back towards the
Market Views
Figure 2.
Price action since the financial year-end has been dollar positive, so the BOJ has not had call to intervene to counter any
yen strength. Renewed weakness in the dollar may attract the
authorities back into the market but if the Japanese stock market sustains the bull run, and deflationary fears continue to
recede, then any intervention is likely to be of reduced magnitude and be used to smooth, rather than halt, yen strength. It
will be interesting to see their involvement if the 101.25 lows
from late 1999 are tested. The performance of the Nikkei, particularly the inflow of foreign investment, will be decisive for the
future path of USD/JPY. While the Nikkei posted a bearish
outside week in mid-April, the bull trend remains intact and the
longer term moving averages remain supportive for further
stock market gains.
nated asset. In short, it is a combination of moving averages
of varying time periods and with varying lead and lag times.
The "cloud" which is the shaded area between two of these
averages (plotted with a 26-period lead time) reflects the trend
of the market. The weekly USD/JPY chart shows how the
market often accelerates on a break of the cloud and that it can
also act as important support and resistance. Note the impulsive break down from the head-and-shoulders top through the
cloud support in June 2002 and the failure of subsequent rallies, within the descending triangle, at the cloud low. The
weekly cloud base is currently within the September gap,
which along with the long-term bear trendline from early 2002,
is reinforcing this as an area of major resistance. Despite the
short-term recovery from the March lows and potentially bullish dollar indications from the dollar index chart, the head-andshoulders pattern remains intact and gives scope for a longerterm move towards the reversal pattern objective at circa
95.50. We would expect the 100-101 area to be a significant
interim sticking point on this decline.
Ichimoku Kinko Hyo is a Japanese technical system that is
worthy of study in this market and indeed in any yen denomi-
Kevin Edgeley CFA is executive director and technical
analyst at Goldman Sachs
September gap. Crucially this move failed to enter, let alone
fill, the gap and the market declined again to new lows. The
latest rally should also be capped ahead of that major gap
area (112.70-113.60).
May 2004
7
Market Views
BRENT CRUDE
by Cliff Green
Brent crude prices could come under
renewed downward pressure in the shortterm. The longer-term outlook, however,
remains very positive.
hile the medium-term trend structure for Brent crude is
clearly upwards, the past 29 months' bull cycle still looks
to be a component of a major consolidation pattern which
started back in early 2000. As can be seen in Figure 1 prices
are now approaching the upper boundary of this broad trading
range with strong resistance anticipated in and around the
$34.50-$35.50 region. With oscillators showing the market to
be rather overbought, this should add a degree of potency to
this anticipated area of supply increasing the chances of a
period of correction if not a fresh downward leg within the prevailing sideways pattern.
W
However, underlying technical studies favour eventual breaks
above these historically important levels which would complete
an accumulative platform capable of supporting advances
towards the 1990 peaks at around $41.00 and possibly a marginal new high with targets of $45.00 clearly readable.
uptrend support waiting at $28.00 likely to damage the overall
positive tone. This may trigger more serious falls setting values on course to challenge the more important $22.50-$23.00
zone.
On the individual delivery months, Figure 2 clearly shows how
values are moving higher within the parameters of an upward
slanting channel with immediate rally attempts likely to again
meet strong resistance towards the upper boundary around
the $34.50 region (basis June 2004). This market appears to
be increasingly vulnerable to more serious corrective weakness with a fresh test of pivotal support around $32.30 likely in
the shorter-term. A decisive breach of this level would trigger
falls towards uptrend support waiting in the $30.00-$30.50
region with only a clear and sustained break beneath this likely to damage the medium-term positive outlook and trigger
falls closer to the $26.00 area.
Immediate corrective pullbacks should uncover support at
$30.00 initially with only a clear and sustained break beneath
Cliff Green has been a technician since 1971 and was previously a senior technical analyst with Merrill Lynch in
London and a partner at Trend Analysis Ltd. He is now an
independent consultant specialising in the Commodity
Markets. ([email protected])
Figure 1.
Figure 2.
8
May 2004
MTPredictor
TM
The software solution for complete trading excellence
Designed exclusively to find, assess and manage only the
very best trades in stocks, currencies and commodities
This is the type of trade MTPredictor can automatically uncover for you….
A Profit of approximately 7x the initial risk required to take the trade, ignoring
slippage and commissions, in the UK stock GKN (October 2003)
End-of-Day and Real-time programs with automatic routines for:
·
·
·
·
·
·
Ideal trades: Find exceptional set-ups with outstanding Risk/Reward prospects
Ideal trade size: Control your position size
Ideal trade management: Display the exit stop strategy on-screen
Ideal trading psychology: Consistent, logical trading, time after time
Systematic Elliott Wave software: Avoid the pitfalls of standard Elliott analysis
Advanced strategies: Expert trade opportunities and management plans
TAKE CONTROL OF YOUR TRADING WITH THE IMMINENT
LAUNCH OF THE NEW MTPREDICTOR 4.0 SERIES!
MTPredictor Ltd www.mtpredictor.com [email protected] Tel +44 (0) 208 9776191
Market Views
MAKING WAVES
APPLYING ELLIOTT WAVE THEORY TO THE US DOLLAR,
NIKKEI AND FIXED INCOME MARKETS by David Murrin
The US Dollar
The long-term view on the dollar is interesting. My general
view is that the US equity market leads the decline of the dollar by about 12 months, from the 2000 peak to the recent
February lows. The February lows are viewed as important
and will probably remain in place for the next 12 months or so,
allowing a long-term correction to dollar weakness to unfold.
Once this is complete somewhere in 2005 we would expect
the commencement of a second bear move in the dollar.
The overwhelmingly negative sentiment on the dollar matched
by massive bearish positions across the market place should
always provoke serious questions as to the sustainability of a
trend. For a clear picture, it's important to look at a number of
Figure 2.
in any market that is highly analysed and traded, and is the
market's way of adding a degree of uncertainty to the process.
That aside, the February lows in the dollar were marked by
very short five of five waves, a very clear signal that the market had lost momentum. USD/CHF has now broken its declining trend line at 1.3040 confirming the low and is set to rally
back to the 1.41 level in the next three months. The euro however has still to break its trend line at 1.1660, though it seems
that this is likely to occur and that we will see a move to the
1.06 level in the next three months.
Fixed Income
We believe there is a significant change in the relationship
between the US and European economies (which are on a
secular decline this decade) and the Eastern economies such
as China and Japan (which are on a secular rally). In this con-
Figure 1.
currency pairs. For example, USD/CAD (see Figure 1) shows
a very clear five-wave decline from its 2002 highs and the clarity of the internal five-wave structure suggests the January
2004 low was the end of the move. We have since been rallying as part of a correction to the 1.41 target in the A wave. Of
further note was the way this pair respected the trend break
shown on our chart as the correction unfolded. This was the
first pair to show the dollar's decline had finished.
The next two pairs of note are USD/CHF and the EUR/USD
(Figures 2 and 3). Both of these show a completed five-wave
decline, but the image is slightly complicated by a large third
wave unfolding with internal corrections of the same magnitude as those of one higher degree. This is now very common
10
May 2004
Figure 3.
Market Views
Figure 4.
Figure 6.
Figure 5.
Figure 7.
text, and coupled with the chart in Figure 4, we believe the US
T-bonds are declining to the 98 region in the C-wave associated with last year's rapid bear market. However at this stage we
still favour the scenario that the move is a big three-wave
decline, at the bottom of which sentiment will lean strongly to
a sequence of rate rises. Nevertheless, with our view of the
decade being one of US deflation, we think rates will stay low
and that the end of the C-wave will offer good buying levels.
The NIKKEI
Meanwhile, in Japan, if we are correct about the secular rally
in the fortunes of the Japanese economy then long-term rates
are on the rise and we have seen the lows for many years to
come. The JGB chart (see Figure 5) shows clear three-wave
rallies and five-wave declines that support such a view. Our
first target is the 132 region and our basic strategy will be to
sell three-wave rallies and buy back five-wave declines until
proven otherwise.
Consistent with our overall analysis, we are bullish over the
long-term for the Japanese stock market. This view is supported by the clear completion of a three-wave decline from the
1989 high to the 2003 low shown in Figure 6. This conclusion
is supported by the clear internal count showing wave five of
five of big C in March last year. Figure 7 shows the details of
the rally with five-wave advances and three-wave corrections,
suggesting the market will continue through the 12,000 level to
14,000 in the next few months. It is no surprise that short JGBs
and Long NIKKEI are in effect the same trade with some shortterm variance.
David Murrin is chief investment officer of Emergent
Asset Management Ltd
May 2004
11
Techniques
ANDREWS' PITCHFORK THE PRICE FAILURE RULE
by Gordon DeRoos
Those who have studied Dr. Alan H.
Andrews' trading techniques in depth
know that he taught ways to expect a
change in market sentiment, a change
that often results in a price move that
catches many traders by surprise. During
one of his seminars Andrews said, "a lot
of traders spend a good deal of their
time following the markets to see what
prices are doing. I suggest they would be
better off if they spent more time
observing what prices are not doing."
A
core Andrews technique that deals
directly with what prices are not doing
is called the "price failure rule." It comes
into play when prices don't reach the median line. Before we look at the technique,
let's first consider the broader interpretation of price action vs. the median line.
The median line as a price magnet
The median line technique is Dr. Andrews'
best-known work. It is the basis for the
trading tool commonly called "Andrews
Pitchfork" that is found on many charting
software programs. The primary feature is
the high probability that prices will reach a
median line and then reverse. This aspect is
well known among traders. Less known is
that the technique goes a great deal further
than that.
Figure 1 is a typical example of how the
median line acts as a price magnet, drawing
prices towards it. The discovery of this
phenomenon prompted Dr. Andrews to
develop supplemental techniques that
would help confirm the outlook of trade
positions already on the books, or alert him
to a potential shift in market sentiment as
prices interacted with the median line. The
price failure rule is one of those techniques.
12
Figure 1. A, B and C are the three points which determine the shape of the pitchfork. Point
A is placed at the end of the previous trend (and is the "handle"), B at the top of the next
trend and C at the bottom of the trend (together, the base of the "fork").
The price failure rule
To deal with those times when prices
change direction before reaching the median line, Dr. Andrews developed a special
method called the price failure rule. It is an
easy to use, two-step process that prepares
the trader for a trend change when new
buying or selling interest appears to be surfacing.
The rule relates to Dr. Andrews' comment
quoted earlier in the article regarding the
importance of observing what prices are
not doing; i.e. not continuing to press on
until the median line is reached. Price failure to reach the median line raises a couple
of interesting issues: market sentiment is
more than likely changing and it is not
May 2004
unusual to see a large countermove take
place following a price failure.
Confirming a price failure
Figure 2 shows prices steadily moving lower
after the pitchfork was drawn, well on the
way to the median line. An Andrews trader,
bearing in mind his observation that prices
will reach a median line more often than
not, would have had good reason to look
for prices to continue the move down to
the 15 area. However, that outlook would
have been less clear when prices reversed
and penetrated the upper parallel line of the
pitchfork. That was a warning that a price
failure was likely, calling for Hagopian's
rule, a trendline adaptation Dr. Andrews
Techniques
highs. It slopes away from the median line,
which is the preferred arrangement according to Andrews. In the example, a buy signal was given when prices broke through
the Hagopian line, completing the setup for
the price failure rule.
Traders looking for additional trade signal
confirmation ought to find that their
favorite technical indicators would merge
quite well with this technique. Oscillator
divergence is one approach that can tie in
effectively. Volume analysis is another. The
range of indicators is considerable, and
while each can be helpful to some extent,
the real power comes from the clues generated by the price movement itself as it
interacts with the median line. The test is to
recognize those clues, and that is where the
price failure rule can lend a hand.
Figure 2.
Figure 3.
named after one of his early course members.
when prices cross a trendline they were
moving along before reversing."
Hagopian's rule
In figure 3, the pitchfork median line is "the
line at which probability indicates such a
reversal could start." The "trendline they
were moving along before reversing" is the
Hagopian line.
The Hagopian line drawn in Figure 3 is a
downtrend line drawn across two previous
Here's how he described the rule in his
original course: "When prices reverse trend
before reaching a line at which probability
indicates such a reversal could start, proper
action may be taken in buying or selling
May 2004
A question often arises regarding pivot
selection for the pitchfork. Users wonder if
there is a way to determine which set of
three pivots (e.g. A, B and C in Figure 1)
would be the best to use in any given situation. Dr. Andrews addressed that very question during one of his seminars. Here is
how he responded: "Usually it seems best
to start with the most recent three alternate
pivots of the time frame you're interested
in. That will give you the current outlook.
But don't stop there, because any three
alternate pivots can be used, and no matter
which set is chosen, each resulting pitchfork
will add something to the outlook as it tells
its own story." The price failure rule is just
one part of the pitchfork story.
Gordon DeRoos is an ex-US Army officer (24 years) and commodity broker (18
years). Although retired, he has been
teaching Dr. Alan H. Andrews'
action/reaction trading methods for the
last seven years. www.pitchforkprimer.com
13
Techniques
THE FOUR-WEEK RULE
by Alex Martin
The Four-Week Rule is a basic
method that may not appear glamorous in the company of Fibonacci
numbers and Japanese candlesticks.
Yet despite its simplicity and its obvious shortcomings as a trend-following system (it works well in up or
downtrends, but not sideways trends),
in the right hands it can be a powerful
and profitable tool.
The charting rules
The trading rules
The price channel generates the following
signals (see Figure 1):
1. When the price is at its highest in a four
week period, buy long and cover short
positions.
2. When the price falls below the lows of a
four week period, sell short and liquidate
long positions.
3. This last rule only applies to future
traders, which is “to roll forward, if
necessary, into the next contract on the
last day of the month prior to expiration.
1. buy signals are produced when the price
closes above the upper band of the price
channel.
2. sell signals are generated when the price
closes below the lower band of the price
channel.
Box 1. The rules
The rules according to Donchian
Developed by Richard Donchian in the
early 1970s for commodities and futures (it
is also known as the "price channel" or
"Donchian channels") The Four-week Rule
is a method that includes a set of charting
rules that are generated from the price
channel as well as a set of trading rules. The
mistake that is often made it to use the price
channels without the trading rules. But it is
the application of both sets of rules that
make the method effective. (See Box 1)
As you can see from Figure 2, trend following systems react to movements rather than
attempting to predict them. The trend
breaks before the price closes below the
lower band of the price channel.
When interpreting the price channel on
charts, buy signals are generated when the
price channel has closed above the upper
band as shown in Figure 3. The price channel tends to create quite a few signals during the course of the up trend.
For those who use technical stock screeners, use a screen with a rising close condition where the price closes higher than the
day before for three days, as well as a price
that closes above the upper band. When we
include a three-day rising close as well as a
14
price channel breakout, the number of false
signals is reduced as can be seen in Figure 4
below. The stock used in all of the chart
illustrations was found using the following
stock screen:
price-channel buy, where the price penetrates the
upper band, as well as the condition that the close
for the last three days was higher than the day
before it.
(The reverse does not apply during sell conditions, three consecutive days down is not
the best pattern to wait for.)
Donchian, in conjunction with the Fourweek Rule, to create combined signals that
help you determine if the price has really
generated a strong trend. Note: The rules in
these two systems do not conflict with one
another.
The 5- and 20-day moving averages
method
The 5- and 20-day moving averages method
includes several general and supplemental
rules. These rules were initially intended for
currency markets but can also be used to
analyze stocks.
Complimenting the Four-week Rule
The method consists of the following rules:
So what can you do to increase the effectiveness of the Four-week Rule so that you
don't miss opportunities due to the lagging
indicators? And equally as important, how
can you ensure that you aren't going to lose
money in a volatile or sideways-trending
market due to false signals?
One way to add certainty to the Four-week
Rule is to use complimentary indicators or
methods to generate additional signals that
provide a warning or confirmation.
For example, you can use another trend-following system, the 5- and 20-day Moving
Averages Method, also developed by
May 2004
Basic rule A: Act on all closes that cross the
20-day moving average by an amount
exceeding by one full unit the maximum
penetration in the same direction of any
previous closing when the closing was on
the same side of the moving average.
Basic rule B: Act on all closes that cross the
20-day moving average and close one full
unit beyond the previous 25 closes.
Basic rule C: Within the first 20 days after
the first day of a crossing that leads to a
trading signal, reverse on any close that
crosses the 20-day moving average and
Techniques
Figure 1.
Figure 2.
Figure 3.
May 2004
15
Techniques
Figure 4.
Figure 5.
Figure 6.
16
May 2004
Techniques
Summary: Getting the Four-week Rule to work
1.
Apply both sets of rules (trading and charting)
2.
Buy and sell strictly according to the rules
3.
Compensate for its shortcomings through complimentary
analysis
closes one full unit beyond the previous 15
closes.
Basic Rule D: Sensitive five-day moving
average rules for closing out positions and
for reinstating position in the direction of
the 20-day moving average are:
1. Close out positions when the currency
closes below the 5-day moving average for
long positions and above the 5-day moving
average for short positions, by at least one
full unit more than the greater of either the
previous penetration on the same side of
the 5-day moving average, or the maximum
point of any penetration within the preceding 25 trading days. Should the range
between the closing price in the opposite
direction to the Rule D closeout signal be
greater than the prior 15 days than the
range from the 20-day moving average in
either direction within 60 previous sessions,
do not act on Rule D closeout signals unless
the penetration of the 5-day moving aver-
age exceeds by one unit the maximum
range both above and below the 5-day moving average during the preceding 25 sessions.
Figure 5 generated by the 5- and 20-day
method, we can see that signals are generated earlier on in the trend than the price
channel shown in Figure 6.
2. Reinstate positions in the direction of
the basic trend (a) when the conditions in
paragraph 1 are achieved, (b) If a new Rule
A basic trend is given, or (c) if new Rule B
and Rule C signals in the direction of the
basic trend are given by closing in a new
low or new high ground.
To better interpret the signals generated by
the 5- and 20-day method, it is advisable to
include an MA cross system such as
Japanese Crosses.
3. Penetrations of two units or less do not
count as points to be exceeded by Rule D
unless at least two consecutive closes were
on the side of the penetration when the
point to be exceeded was set up. (Richard
Donchian, December 1974 Futures article),
as quoted by Cornelius Luca in Technical
Analysis Applications in the Global
Currency Markets, 1997.
When we look at the charting signals in
May 2004
Combining the 5- and 20- day moving average cross system with the Four-week Rule
can help to confirm information about the
potential trend change. These modifications
are not intended to replace basic trend-following techniques, but to provide more
information about the trend when price
channel signals are generated.
Alex Martin is Chief Technical Officer
at ChartFilter.
www.chartfilter.com
17
Techniques
USING THE
MCCLELLAN OSCILLATOR
by Tom McClellan
In the last issue of The Technical
Analyst, Tom McClellan gave an
introduction to The McClellan
Oscillator. In this article, Tom
takes us through some of the
more advanced ways in which it
can be used to help forecast market direction.
T
he McClellan Oscillator is a tool which
measures the acceleration in daily
Advance-Decline (A-D) statistics by
smoothing these numbers with two different exponential moving averages, then finding the difference between them. The
Oscillator's most basic indication is its position relative to the zero line, which is the
Oscillator's neutral level. The market is
nearly always accelerating or decelerating, in
one direction or the other, and rarely has a
neutral acceleration condition. A positive
Oscillator reading is an indication of
upward acceleration, while a negative reading is a sign of downward acceleration. But
there is much more that the Oscillator has
to tell us.
Overbought/Oversold
When the McClellan Oscillator reaches an
extreme level, either high or low, it indicates
an extended condition for the market. In
this respect, it is like many other overbought/oversold indicators, and like the
others, an extended McClellan Oscillator
reading is no guarantee that the extended
market condition has to end right away.
Oversold readings on the McClellan
Oscillator offer us some additional insights
when interpreted properly. First of all,
deeply negative readings tend to indicate
the conclusion of a down move, whereas
extremely high readings tend to show initi-
18
Figure 1.
ation of a strong new up move. Also, a
deeply negative Oscillator reading which
comes along after a long period of quiet is
a harbinger of more trouble to come.
We see great examples of all of these principles in Figure 1, showing the Oscillator in
1998 and 1999. Point 1 in this chart was a
deeply negative reading (-271) which came
after a long quiet period. As such, it gave us
warning of the weakness that arrived later
in 1998 when the "Asian Contagion" hit the
markets. Points 2 and 3 in this chart were
also very low, but rather than being indicative of future weakness to come they were
the fulfillment of the weakness forecast by
point 1. They also marked the end points of
strong down moves, with prices either
reversing or at least moving sideways for a
while as the bears gathered more strength.
For several months prior to point 3, there
had been no strong up moves accompanied
by very high Oscillator readings. The postings above +200 beginning in September
1998 were a sign that the bulls were going
to be rushing back in and that they had
enough money to push prices higher for a
sustained period of time. These high post-
May 2004
ings differed from the very low readings
because low readings are indicative of the
conclusion of a down move, whereas the
high readings tend to occur at the very
beginning of a strong up move. We almost
never see the price move higher on the
highest Oscillator reading. So when one
sees a very high reading, it may be a sign
that a brief pullback is needed, but it is also
a sign that higher prices should be expected
following that pullback.
In Figure 2, we see a great example of a
conclusive indication from a very oversold
Oscillator reading. This bottom was not followed by any strong positive readings for a
long time and the result was a choppy,
range-bound period for stock prices.
Some sources on technical indicators will
prescribe specific Oscillator values that represent overbought and oversold levels, but
we discourage people from following such
guidelines. A wide variety of factors can
affect the amplitudes of Oscillator moves at
various times, including market volatility,
the strength of price moves, and changes in
the number of issues traded on the
exchange. So an Oscillator value that might
Techniques
Divergences
To the extent that the Oscillator's movements diverge from price action, it can signal an impending change in direction for
prices. This is where it helps to understand
that the Oscillator serves as a measure of
acceleration for the market breadth statistics. Measuring the acceleration can be helpful to signal an impending change in trend
direction.
Figure 2.
indicate an extreme condition during one
period may only be a routine high or low
during another period. One way to adjust
for this is to calculate a "Ratio-Adjusted"
McClellan Oscillator. Using Ratio-Adjusted
McClellan Oscillator values does indeed
adjust for the changing number of stocks
on the exchange but it does not adjust for
other factors such as fluctuating market
volatility or changes in the diversity of
issues represented which may produce
greater or lesser Oscillator swings.
For periods of less than 2 years, we believe
that it is fine to use the conventional
McClellan Oscillator. Rather than focusing
on the specific numerical value, an examination of the chart pattern will give much
more information about what the
Oscillator has to tell us. Certain chart structures and behavior can be enormously
revealing.
Figure 3 shows several divergences between
the price action in the NYSE Composite
Index and the McClellan Oscillator. Notice
that these divergences tend to occur more
often at tops than at bottoms, which is due
in part to the way that the US stock market
tends to have more rounded tops and
exhaustive (spike) bottoms. This is not to
say that no divergent bottoms can be
found, just that divergent tops are much
more frequently seen.
Congestion Zones
A congestion occurs when the Oscillator
fluctuates by very small increments over
several days. One or two days of small
changes is not enough, it has to be a sustained period. The Oscillator value area
where a congestion occurs (called the congestion zone) usually forms above the zero
line. We seldom see them form at extended
negative values.
The basic rule to remember is that a congestion zone is something to drop out of.
Figure 4 illustrates a few examples of congestion zones. The common characteristics
of each are that they show several days of
postings with the Oscillator in a relatively
small range. Once the Oscillator breaks
down out of that range the market begins
to decline sharply. A couple of these examples even have the congestion zone forming
Figure 3.
May 2004
19
Techniques
form, although that weakness may not be
manifesting itself during the period that the
simple structure is formed.
For example, the Oscillator could be chopping up and down below zero, implying that
the bears are strong and then it might move
briefly above zero as the bulls try to regain
control. But if (in this example) the
Oscillator moves straight up through zero
and then turns around and moves straight
back down through zero again, it is a sign
that the bulls do not really have the strength
to carry on their mission for more than a
brief period and the bulls cede control back
to the bears.
Figure 4.
at or below zero, but the result was still a
drop down out of the congestion zone.
Looking at one day's Oscillator value would
not convey this information; it takes a chart,
and someone to interpret that chart, to
notice behaviour like congestion zones and
divergences developing.
Complex Versus Simple Structures
When the Oscillator moves up and down
over a period of days on one side of the
zero line, we call that a "complex structure".
Complexity of a structure implies strength
for the side (of zero) upon which it forms,
whether positive or negative. A "simple
structure" is one in which the Oscillator
crosses zero in one direction in a move lasting from one day up to a few days, and then
turns around and heads directly in the
opposite direction without forming any
complex structure. Simple structures imply
weakness for the side upon which they
Figure 5 shows a few examples of each type
of structure. Where a complex structure
forms, it implies more strength to come for
that side of the market corresponding to
the side of the zero line where the structure
formed, i.e. complexity above zero is bullish, and below zero is bearish. That strength
may be temporarily interrupted while the
other side tries to exert its influence, but
where complexity has formed we have the
expectation that more strength will be manifested in that direction. Often we will see
trending price moves, either upward or
downward, made up of a succession of
complex structures that are interrupted
only briefly by simple structures. When
such a succession of complex structures
gives way to a simple structure, it can mean
that the trending side of the market is ready
to give up control for a while, and the
opportunity is there for the other side to
pick up the ball. Sometimes, neither side
will form a complex structure, meaning that
both the bulls and the bears are equally hesitant to take charge.
Oscillator Trendlines
One interesting feature of the Oscillator is
that it forms trendlines just like price charts
Figure 5.
20
May 2004
Techniques
homogeneous way. In a narrow sector like
gold mining stocks, for example, it is typical
to see all of them go up one day and then
all go down the next day. Other industry
groups and sectors show this same effect to
a greater or lesser degree. By narrowing the
focus to small groups like this, we end up
losing the key indication given to us by
looking at breadth statistics. By examining
the behavior of a diverse collection of
stocks, we can see if there is a different
indication from what we see in prices alone.
Figure 6.
do, but the Oscillator trendlines will usually
be broken before the corresponding price
trendlines are broken. Figure 6 shows a few
examples of Oscillator trendlines, and in
each case the breaking of the trendline signaled a reversal of the prevailing short-term
trend. And also in each case, the trendline
break in the Oscillator preceded the trendline break on the equivalent price chart.
advancing and declining issues, and so it
does not exist as an intraday indicator.
However, it is possible to take the intraday
values for the number of advances and
declines and calculate a "what if" value for
the Oscillator that assumes those A-D values are the closing ones.
Additional Points In Conclusion
It is also possible to use other data to calculate McClellan Oscillators. We calculate and
employ in our analysis breadth versions of
the Oscillator which are derived from A-D
data on the Nasdaq market, the stocks in
the Nasdaq 100 Index, the 30 stocks in the
Dow Jones Industrial Average, the corporate bond market, plus a subset of the
NYSE breadth data for the "Common
Only" stocks (filtering out preferred stocks,
rights, warrants, and closed end funds). It is
even possible to create a McClellan
Oscillator out of any other breadth statistics you might think of such as a subset of
the market that includes all of the stocks in
a particular sector.
The McClellan Oscillator is based on the
daily closing values for the NYSE's totals of
The problem with subset breadth statistics
like this is that they tend to all behave in a
It is important to be careful when drawing
such lines, and more importantly, when
drawing conclusions from them. Generally
speaking, trendlines which span longer periods become less meaningful, and it is better
practice to stick to the steeper trendlines
which span 3-6 weeks. As with price trendlines, it is not unusual for the Oscillator to
break out above a downtrend line and then
go back down to test the top of that line
before continuing higher.
May 2004
Breadth statistics are valuable because they
give some of the best indications about the
health of the liquidity that is available to the
stock market. A small amount of money
can be employed to make a handful of
stocks go up or down and if they are the
right stocks then even the major market
indices can be moved. But to affect the
breadth numbers, which measure all of the
stocks on the exchange, requires major
changes in the liquidity picture. The available money has to be so plentiful that it can
be spread far and wide in order to make the
majority of stocks close higher, and especially so in order for the market to show
positive breadth for several days.
By measuring the acceleration in the
breadth statistics, which is what the
McClellan Oscillator does, one can gain
important insights about impending trend
direction changes for prices.
Tom McClellan is the editor of The
McClellan Market Report. He is the son
of Sherman and Marian McClellan who
created the McClellan Oscillator in
1969.
www.mcoscillator.com
21
Techniques
VOLUME SPIKES AND INDEX REVERSALS
by Steffen Norgren and Andrew von Stuermer
In Volume Analytics, volume data
plays more than just a minor supporting role - it is the principal
variable used to forecast reversals
in stock exchange indices.
T
he extensive body of knowledge associated with volume analytics has given
rise to a view of the markets that is proven,
time-tested, highly consistent, and profitable. Its basic premise is that volume and
index behaviors are closely interrelated and
that the trading patterns of an index can be
predicted, or at least anticipated, from a
proper understanding of the unfolding volume patterns. The technique provides the
trader with an elegant way of monitoring
and analyzing the volume behaviour of a
particular index and allows him or her to
heed one of the golden rules of trading,
"Do not play against the market."
But why apply volume analytics to indexes
and exchanges, rather than to individual
stocks? Indexes best describe the mood of
the market as a whole. Regardless of what
you trade, a particular index or sub-index,
stocks, options, futures, most of these trading vehicles tend to move in concert with
the broad market. As a rule, the market dictates the direction of a particular security,
never the other way around. It therefore
makes sense to get a good grasp on what is
happening at the index or stock exchange
level, and we have found volume analytics
to be an excellent vehicle to make that
determination.
Terminology
Every trader is familiar with moving averages of securities prices, perhaps the most
frequently used technical indicator. We simply apply the concept to volume, rather
than to price, and plot Volume Moving
22
Averages (VMA) that range in duration
from as short as a few minutes to as long as
several months.
However, there is a slight twist to this.
Volume activity typically follows certain
predictable patterns throughout the trading
day, with high levels prevalent immediately
after the open, lower values around noon,
and increased levels once more toward the
close. We call this pattern the "time factor". Unfortunately, the time factor provides a rather distorted picture of the daily
volume activity. It makes it difficult to differentiate those volume events, which are
truly significant, from those that are simply
part of the normal daily fluctuations. We
have solved the time factor issue by
normalizing volume data before charting
it. Charting normalized volume allows a
much clearer determination of whether or
not volume levels are spiking above normal
levels, an aspect that is at the core of our
methodology.
We are particularly interested in the appearance of large peaks ("spikes") in the VMA known as VMA spikes - and how an index
reacts when they are generated. Sudden
VMA surges are indicative of bursts of significant buying or selling activity. As such
spikes occur, we determine whether the
index is moving up or down at that time. If
the direction is up, we call the associated
volume surge a resistive VMA spike; if the
index direction is down, we label the spike a
supportive VMA spike. In the absence of
distinct volume spikes, we still call any volume generated as the index is moving up
resistive volume, and as it moves down,
supportive volume.
Basic principles
The most basic premise of volume analytics is we can always anticipate an index will
May 2004
react to (significant) volume spikes - as a
rule, resistive volume spikes will force a
downward move in the index; supportive
volume spikes will generate upward index
momentum. This basic assertion must be
qualified by two key questions:
1. What determines the extent and characteristics of an anticipated move: Will it be
short-lived or have staying power over the
mid- to long-term? Will it be gradual or
sudden?
2. What determines when an anticipated
move will most likely occur: Will it happen
immediately or will there be a certain time
lag (a "delayed volume reaction")?
Our research shows the answers to these
questions vary considerably, depending on
(a) the general market context, and (b) the
technical characteristics of the actual volume spike(s) being analyzed. Therefore, in
order to get the most value from volume
analytics, it must always be placed in the
proper context:
Market context: Where in the larger market picture do supportive / resistive VMA
spikes appear: During short-term pullbacks
within a larger uptrend? As part of shortterm upside corrections within a larger
downtrend? At the presumed end of a
weakening long-term trend? At the beginning of a new trend or somewhere in its
middle? During distinct trend runs or in
markets with choppy sideways trading
action (i.e., in support / resistance corridors)?
Technical considerations: When analyzing a VMA spike, consider its magnitude,
both vertically (the height of a thrust) and
horizontally (its width or breadth).
Comparatively larger and/or wider spikes
obviously carry more weight. Caution must
be exercised when analyzing volume spikes
Techniques
on a short time frame, as their potential
impacts on mid- or long-term trends can
easily be misjudged. A noteworthy spike
appearing on a 5-minute chart could well
affect an index in the short-term, but it may
not necessarily have much of an impact on
the prevailing long-term trend.
Practical considerations
We suggest using only real-time intraday
index charts and applying volume analytics
to highly liquid indexes that reflect not only
the US economy, but also the world economy, such as the NASDAQ 100, the
S&P500, and the Russell 3000. Place volume spikes in a broader market context by
consulting several charts with different settings; we suggest a chart range from intraday to at least 2-years. Compare current volume events with those of the past. Finally,
it is essential to use only volume data that
has been normalized, so that the spikes you
observe are not distorted by the time factor.
Figure 1.
Chart examples
Index values will always (sometimes immediately, sometimes with a delay) react to volume spikes, and the greater the magnitude
of a spike (or series of spikes), the stronger
the ensuing reaction. (The many complex
reasons why sudden volume surges take
place are beyond the scope of this article).
For example, at the end of 2002/beginning
of 2003 the long market downtrend in the
S&P 500 finally reversed and switched to a
steady up-trend (Figure 1). A volume analysis chart provides us with fresh insight.
Three volume spikes (two large ones in July
and October 2002, as well as a smaller
VMA peak in February 2003) correspond
with a distinct long-term trend change for
Figure 2.
May 2004
23
Techniques
“INDEX VALUES
WILL ALWAYS
REACT TO
VOLUME SPIKES.”
October 10, 2002 and that the January 2003
move to retest the recent lows was just a
mid-term correction of the new up-trend.
Figure 2 (a 30-day chart) clearly shows how
each volume spike was followed by an index
reversal, whereas Figure 3 shows that the
relationship between volume spikes and
index reversals applies equally well to the
short-term.
Steffen L. Norgren and Andrew von
Stuermer, Highlight Investment Group.
Figure 3.
the S&P 500. You could argue it was
prompted by the outbreak of the war in
Iraq. However, our volume analysis demonstrates the index was ready to move up,
24
given the large buildup of supportive volume, as evidenced by the two very significant volume spikes. It could also be argued
that the new uptrend actually began on
May 2004
We would like to acknowledge
MarketVolume for their analytical support and for the chart material presented in this article.
www.MarketVolume.com
Techniques
THE INVERSE FISHER TRANSFORM
by John Ehlers
The purpose of technical indicators is
to help time your trading decisions.
Hopefully, the signals are clear and
unequivocal. However, more often
than not your decision to pull the trigger is accompanied by crossing your
fingers. In this article, I explain a way
of making oscillator-type indicators
give clear black-and-white signals of
when to buy or sell.
I
n the past, I have noted that the probability distribution function (PDF) of
prices and indicators do not have a
Gaussian probability distribution. A
Gaussian PDF is the familiar bell-shaped
curve where the long "tails" mean that wide
deviations from the mean occur with relatively low probability. The Fisher
Transform can be applied to almost any
normalized data set to make the resulting
PDF nearly Gaussian, with the result that
the turning points are sharply peaked and
easy to identify.
The Fisher Transform is defined by the
equation
1)
e2 y − 1
x = 2y
e +1
The transfer response of the Inverse Fisher
Transform is shown in Figure 1 (with x and
y reversed to correspond to the usual definitions of input and output). If the input
falls between -0.5 and +0.5, the output is
nearly the same as the input. For larger
absolute values (say, larger than 2), the output is compressed to be no larger than
26
unity. The result of using the Inverse Fisher
Transform is that the output has a very high
probability of being either +1 or -1. This
bipolar probability distribution makes the
Inverse Fisher Transform ideal for generating an indicator that provides clear buy and
sell signals.
One of the more popular technical indicators is a Stochastic RSI. This indicator starts
by taking an RSI of price. Then, a
Stochastic of that RSI is taken to limit the
output to between 0 and 100. Translating
and scaling, this is mathematically the same
as varying between -1 and +1.
⎛1+ x ⎞
y = 0.5 * ln⎜
⎟
⎝1− x ⎠
Whereas the Fisher Transform is expansive,
the Inverse Fisher Transform is compressive. The Inverse Fisher Transform is found
by solving equation 1 for x in terms of y.
The Inverse Fisher Transform is:
2)
Figure 1. Transfer response of the Inverse Fisher Transform compresses the output to
between -1 and +1
Vars: IFish(0);
Value1= .1*(RSI(Close, 5)-50);
Value2= WAverage(Value1,9);
IFish=(ExpValue(2*Value2)-1)/(ExpValue(2*Value2)+1);
Plot1(IFish, "IFish");
Plot2(0.5, "Sell Ref");
Plot3( - 0.5, "Buy Ref");
Figure 2. EasyLanguage Code to Take the Inverse Fisher Transform of an RSI
May 2004
Techniques
Figure 3. Inverse Fisher RSI
But there is no reason to bludgeon the RSI
with a blunt instrument like a Stochastic.
Instead of picking an observation length
that is guaranteed to drive the Stochastic to
saturation, you can finesse the indicator
PDF using the Inverse Fisher Transform.
The Easy Language code to do this is given
in Figure 2.
The 5 bar RSI varies from a minimum of 0
and a maximum of 100. The 5 bar length of
the RSI was selected to provide good operation when applied to many price series.
The RSI period is certainly available for
optimization. By subtracting 50, the RSI
indicator is translated to range from -50 to
+50. Then, multiplying by 0.1 reduces the
range to be between -5 and +5 for Value1.
This is just the kind of maximum swing
suited to the Inverse Fisher Transform. I
used a 9 bar weighted moving average to
compute Value2 to smooth Value1 and ultimately remove some spurious trading signals. There is no magic in this average. It
could have fewer bars to have less lag or it
could be an Exponential Moving Average.
Its function is just to be a smoother. The
transform is calculated as the variable IFish
and then plotted. The code also plots output reference lines at -0.5 and +0.5.
The transformed RSI is applied to the
Exchange Traded Fund (ETF) QQQ in
Figure 3. The trading rules are simple. Buy
when the indicator crosses over -0.5 or
crosses over +0.5 (if it has not previously
May 2004
crossed over -0.5). Sell short when the indicator crosses under +0.5 or crosses under 0.5 (if it has not previously crossed under
+0.5). The trading signals are not only clear
and unequivocal, but they are also profitable.
The Inverse Fisher Transform can be
applied with equal success to virtually all
oscillator-type indicators and should provide greater confidence in deciding when to
enter and exit trades.
John Ehlers is an electrical engineer and
has been a private trader since 1978.
www.mesasystems.com
27
Techniques
ASTRONOMY AND THE DOW JONES
by Larry Pesavento
Published in 1948, Donald
Bradley's
'Stock
Market
Predictions - the planetary barometer and how to use it' was a 50
page booklet that sold for four
dollars. It received little attention
and it would be forty years before
Larry Pesavento and Arch
Crawford began publishing yearly
forecasts based upon Bradley's
formulas. The Technical Analyst
takes a step into the unknown and
asks Larry Pesavento to explain
what it's all about.
T
he Bradley model gives a chart (called a
siderograph) based on the classic
Ptolemic harmonic angles between any two
planets. Although the Bradley model can
sometimes "predict" the exact highs and
lows of the stock market it is far from infallible. However, the key turning dates in the
model are very useful. These can be used
for locating tops and bottoms in the Dow
Jones as well as other actively traded markets.
Figure 1.
What the Bradley model does do is make
the technical analyst aware that there must
be some correlation of prices to various
astrological planetary harmonics. But
Bradley warns, "At no time must the reader
gain the impression that a siderograph, as
such, is a prediction of what the stock market will actually do. Nevertheless, observations prove that basic reversals in collective
attitudes clearly predicted by the line are
inevitably mirrored in stock averages."
It is my opinion that the Bradley model
should be used in conjunction with other
technical tools such as pattern recognition
and wave ratio analysis.
28
Figure 2.
May 2004
Techniques
“OBSERVATIONS PROVE
THAT BASIC REVERSALS
IN COLLECTIVE ATTITUDES CLEARLY PREDICTED BY THE LINE ARE
INEVITABLY MIRRORED IN
STOCK AVERAGES.”
Notice the four charts using the Bradley
model overlay as a cycle tool for determining future price swings.
Figure 1 is a current Bradley model for
2004. Figure 2 illustrates the model’s results
for 2002/2003 and shows that correlations
between the actual stock prices and the
Bradley model have been quite accurate.
Figure 3.
Next we overlay the Bradley model over the
crude oil future (Figure 3) and gold future
chart (Figure 4). There is clearly some correlation in price to astrology. I checked
these correlations in the Bradley model
from 1876 to the present and it consistently produces results above 70%. Not too bad
for a model that can be produced 100 years
or more in advance.
Larry Pesavento is the author of seven
books on trading and is currently a private trader for a large hedge fund.
Charts provided by
www.ensignsoftware.com
Figure 4.
The Technical Analyst apologises for printing the wrong charts in the article "Fibonacci and harmonicity - a personal view" in the
April issue. Please note the correct charts are available via e-mail (in PDF format) from Jim Kane ([email protected]).
May 2004
29
Interview
THE TECHNICAL ANALYST TALKS TO….
Ralph Acampora
Ralph Acampora is managing director and the
director of technical analysis for Prudential Equity
Group. Ralph has taught technical analysis at the
New York Institute of Finance since 1970 and is a
Chartered Market Technician (CMT).
TTA: You are well known as a founding member of the
Market Technicians Association. What was your motivation
for setting this up?
RA: When I started working in the markets in New York
most brokerage houses and major financial institutions in
Wall Street had a technical analyst who spent a major part
of his day analysing charts. Even then, technical analysis
was an established subject but there was very little communication between market analysts. They simply didn't talk to
each other. This was largely because there was a lack of
any formal structure of communication for analysts. At the
time, I had an analyst friend called John Brooks and together we decided to establish an association for analysts
where we could meet on a regular basis and discuss our
subject. This is how the MTA came about.
TTA: Despite the ubiquity of technical analysis in financial
institutions, most houses still appear to devote relatively
few resources to this area of analysis compared to the
large number of economists and fundamental analysts generally employed. Why do you think this continues to be the
case in the major financial centres across the world?
RA: Technical analysis still suffers to some extent with a
credibility problem among heads of the institutions to which
you refer. This is a perennial battle for me and has been
since the start of my career. In my view, the value and legitimacy of technical analysis has been proved again and
again going back to the stock market crash of the twenties.
However, fundamental analysis still reigns supreme among
the banking establishment. I think the reason this attitude
persists is because of the approach to financial analysis
propagated by universities and business schools in the US.
Postgraduate courses such as MBAs have traditionally
30
April 2004
devoted almost no time to technical analysis with preference instead being given to fundamental analysis. For
example, share valuations using traditional mathematical
methods still dominate rather than looking at a chart to
establish if a particular share is overbought or oversold. I
really believe that a sea change in academic circles
towards technical analysis is required before the hierarchy
of our banks and brokerage houses attach more importance to TA. Furthermore, in our academic institutions the
random walk theory still predominates as a method of
describing the long-term path of financial markets, whether
they be stocks, currencies or bonds. This theory has a history of acceptance within the academic community and is
somewhat revered as a valid theoretical model. It is a moot
point where exactly technical analysis stands in relation to
this theory but there are fundamental divergences in the
two methods that have established them as opposing views
of market determinants.
TTA: Is there any evidence that this situation is beginning
to change, at least as far as academia is concerned?
RA: I'm very pleased to report that things have begun to
change, but slowly. There is now a wealth of research into
technical analysis and charting methods emerging from
universities across the globe. The scepticism within the
academic community towards the subject that has existed
for so long is at last beginning to change as academics'
own research and testing of established techniques used
by analysts and traders have been shown to be empirically
robust. I'm pleased to report that Andrew Lo at MIT has
shown interest recently in establishing a chair for technical
analysis at the school. I'm confident that the prejudice
against technical analysis within the academic community
will continue to decline but it won't happen overnight.
Interview
TTA: Do you think the failure of fundamental analysis to
anticipate the crash of 2001 has done much to boost the
profile of technical analysis, especially as so many TA indicators and signals were strongly bearish before the crash?
RA: Undoubtedly, but it should be remembered that this
has happened before. In 1929, the traders who made
money from the crash were those relying on charting methods and got out before the fall. However, the small analyst
community failed to take advantage of this in order to raise
the profile of the subject and so the advantage was lost for
another 30 years or so. However, many of the most influential books on technical analysis and new charting techniques emerged in the 1930s as technicians hurried to put
their proven techniques that allowed them to avoid the
crash into print.
TTA: Are you still actively involved in the activities of the
MTA?
RA: My goal at the moment is to get those who hold the
MTA qualification (CMT) exempted from taking the CFA
qualification. This is currently a big issue among analysts
on Wall Street, many of whom face the prospect of taking
new exams as SEC regulations have changed. From a personal point of view, I don't want to have to start revising for
examinations at age 62 after more than 30 years in the
market. Economists and strategists are exempted even
though they have no formal trade qualification. There is an
injustice here that stems from the lack of acceptance of
technical analysis within the US financial establishment. If
this exemption is something that I manage to achieve I'll
look forward to becoming Saint Acampora among the TA's
on Wall Street!
TTA: Has the use of technical analysis changed in the US
over the last 10 years?
RA: Technical analysis has enjoyed a greater profile in
recent years as the private investor and day-trader community has increased in size. It must be said that television
has played a crucial role in increasing the exposure of the
subject. Charts now feature regularly on news bulletins and
experts such as John Murphy and John Bollinger are now
to be seen regularly on our screens giving their commentary on the markets. As such, technical analysis has succeeded in becoming part of mainstream financial news
reports.
TTA: Does TA remain more popular in the US than in
Europe or Asia?
RA: No, I'm not convinced of this although the use of TA is
mixed across Europe. I am often asked to give talks in various European cities and in some countries the use of technical analysis among both institutions and private investors
is as widespread as in the US. Switzerland and Ireland
stand out as being enthusiastic TA followers. When I go to
Zurich they can't find a hotel large enough to hold the
attendees but in London there is almost indifference
towards such talks. On the other hand, perhaps they just
don't like me in the UK!
TTA: In your book " The Fourth Mega-Market" you predicted the Dow would reach 20,000 by 2011. Do you still hold
to this view?
RA: The book was written in 2000 near the top of the market. At the time, my view was considered by some to be a
very conservative outlook in terms of its relatively small
annual return. Other forecasts and books from the same
period were looking at 36,000 and above for the Dow.
However, despite the crash I still hold to my original view
but don't expect the market to enjoy an uninterrupted rally
to this level. I still expect a long period of very sloppy activity over the forthcoming years and a fall back towards the
7000 level before we reach 20,000.
TTA: Have you seen any recent developments in technical
analysis that have attracted your attention?
RA: Not so much in new developments but more in terms
of what is being done with established techniques. I'm
thinking in particular of the research that is being done with
backtesting and the rigorous studies into pattern recognition. I feel that this work by academics and market professionals is helping to justify the last 40 years of my life! In
the 65 years since the Great Depression we are still playing around to get the subject accepted. Therefore, any
good research that validates the subject is still very much
welcomed.
Correction: In the April issue The Technical Analyst said that Anne
Whitby is vice-chairman of 4CAST. This is incorrect - Anne Whitby
is vice-chairman of the Society of Technical Analysts, not 4CAST.
May 2004
31
Subject Matters
MONDAY BLUES AND SUNNY FRIDAYS
IN THE ASIAN STOCK MARKETS?
by Professor Wing-Keung WONG and Nee Tat WONG
Do stock markets in Asia suffer
from Monday blues and sunny
Fridays, or what is more commonly known as the day-of-the-week
effect? Apparently yes. In agreement with previous studies in the
US and elsewhere (including Asia),
we find that the mean return on
Monday is indeed negative, and
the mean return on Friday is positive and generally the highest.
Y
et despite the mounting studies on the
day-of-the-week effect, researchers have
not been able to explain the causes conclusively. In this study, we attempt to shed some
light on the mysterious day-of-the-week effect
by examining whether the Monday (and
Friday) returns are concentrated in any partic-
32
ular week(s) of the month and whether the
low-return Monday is related to the (preceding) Friday returns, as in the US.
Day-of-the-week effect
Using daily stock index returns from 1986 to
2002, we find a cyclical pattern of stock
returns in the five Asian markets that we studied (see box "The data").
Consistent with the studies in the US and
other countries, the mean returns are negative
on Monday and highest on Friday. To substantiate the evidence for the day-of-the-week
effect, an appropriate statistical test (the Ftest, which is the ratio of two chi-square tests)
is used and results show that the F-statistics
are significant for all the markets in the full
period 1986-2002.
However, sub-period analysis shows that the
values of the F-statistics decline significantly
May 2004
The data
Sample: Daily stock index returns from
2 January 1986 to 31 December 2002
1st sub-period: 1986-1994
2nd sub-period: 1995-2002
Daily returns (Rit) are calculated as: Rit =
(Pit - Pit-1) / Pit
where Pit and Pit-1 are the closing values of
stock index i on days t and t-1 respectively.
Indices used: the Hang Seng Index (Hong
Kong), the KLSE Industrial and Commercial
Index (Malaysia), Manila Commercial and
Industrial Index (Philippines), Straits Times
Index (Singapore) and the SET Index
(Thailand).
Box 1.
from the first sub-period (1986-1994) to the
second sub-period (1995-2002) for most of
the markets. For Hong Kong and Malaysia,
the F-statistics turn insignificant in the second
sub-period 1995-2002. These results suggest
Subject Matters
“SELLING PRESSURE FROM
INVESTORS IS SUBSTANTIALLY
HIGHER FOLLOWING BAD NEWS
ON THE PREVIOUS FRIDAY.”
the day-of-the-week effect has generally
diminished in the Asian markets, and in some
cases disappeared.
the Friday returns across the weeks. Although
Friday returns are generally positive across all
the weeks, they are not significantly positive in
any particular week.
Monday effect
In testing for the Monday effect (and Friday
effect), we divide each month into five calendar weeks. The first week of the month is
defined as the week that contains the first
trading day of the month.
We find no significant difference in the
Monday returns across the different weeks.
This is in contrast to the recent studies in the
US, which show that the Monday effect is
concentrated in the last two weeks of a
month.
Friday effect
Relationship between Monday effect and
Friday effect
Monday returns tend to follow preceding
Friday returns. In particular, Monday returns
are significantly positive (negative) when previous Friday returns are positive (negative).
Thus, it seems that the selling pressure from
investors is substantially higher following bad
news on the previous Friday, as proxied by the
negative returns on Friday. For instance in the
full period 1986-2002, the mean returns on
Mondays following positive Friday returns are
0.219% to 0.535%. In contrast, the mean
returns on Mondays following negative Friday
returns are -0.528% to -0.810%.
Similarly, we find no significant difference in
May 2004
Conclusions
The study re-examines the existence of the
day-of-the-week effect in the Asian markets of
Hong Kong, Malaysia, Philippines, Singapore
and Thailand. Using eighteen years of data up
to 2002, this study provides some evidence for
the day-of-the-week effect (for the full period
1986-2002 and the first sub-period 19861994). However, sub-period analysis indicates
that the day-of-the-week effect has generally
declined and in some cases disappeared in
recent years.
The study also reveals that, unlike the US,
there is no weekly pattern of Monday and
Friday returns in the Asian markets. However,
consistent with the US evidence, we found
that Monday's returns are related to the previous Friday's returns.
Professor Wing-Keung WONG and Nee
Tat WONG, Department of Economics,
National University of Singapore.
33
Subject Matters
WHY HAVE THE RETURNS TO MANAGED
FUTURES FUNDS DECREASED?
by Willis Kidd and Wade Brorsen
Returns to managed futures funds
and Commodity Trading Advisors
(CTAs) have decreased dramatically. Since funds overwhelmingly use
technical analysis, the authors consider why it is that technical trading strategies have become less
successful and find evidence to
suggest that lower price volatility
is the most likely culprit.
D
uring the 1980s and early 1990s, investment in the managed futures industry
grew quickly. In recent years however, futures
fund returns have decreased and the value of
assets invested in managed futures has stagnated (Pendley and Zurla, 2002). The Barclay
Commodity Trading Advisor Index in Figure
1 shows a steady trend of decreasing returns
over the past twenty years. The causes of this
decrease in fund performance are not fully
known. Two possible explanations are:
(a) decreased market volatility (and therefore
profit opportunities) and
(b) price distortion caused by the growth of
the industry.
To date, no author(s) has examined possible
causes nor comprehensively studied a change
in daily return characteristics. Yet research is
needed to determine the ways in which the
market has changed, thereby allowing technical traders to adjust trading systems to
account for these changes.
Our study (which uses bootstrap resampling
techniques) tests the hypothesis that a structural change in price fluctuations has occurred
and that this may have affected the profitability of technically managed futures funds.
Evidence that technical trading returns
have decreased
Mean returns before 1991 and from 19912001 are presented in Table 1 for various
34
Figure 1.
indices of Zurich Capital Markets. All of the
five indices show a substantial decrease in
returns. In particular, the CTA trend-following index shows a dropoff of over six percentage points. The other indices include
some funds that either do not use technical
trading systems or use a mixture of trading
methods.
Brorsen and Irwin (1987) found that 13 of 21
advisors relied solely on computer-guided
technical trading systems, but that only two of
21 used no objective technical analysis.
Billingsley and Chance (1997) used a dataset
that let CTAs describe their trading approach.
As a result, Billingsley and Chance went on to
classify 80% of CTAs' returns as being from
technical trading. They also pointed out that
traders in their non-technical category do rely
partly on technical analysis.
Data we obtained from the Center for
International Securities and Derivatives
Markets ("CISDM", 2000) classify the CTAs
as 18.2% discretionary, 0.3% quantitative,
60.5% systematic, 17.4% trend-based and
3.6% trend-identifier. Several CTAs we know
that rely almost entirely on trend-following
systems are classified as systematic in the
CISDM database. Many of the discretionary
traders also use trend-following systems based
May 2004
on both charts and computers. Thus, the
broader indexes may be as representative of
trend-following returns as the narrower
indices.
The returns quoted in Table 1, however,
should be considered in the light of the following:
Fund returns are reported net of costs and
include interest returns. Both costs and
interest rates have declined over time
Industry sources have related to us that
some of the larger funds may have adopted trading methods that accepted a lower
return in order to get reduced risk, resulting in reduced volatility.
Estimating the decline in gross terms (i.e.
adding back fund costs based on historical records and reports): A conservative
estimate of the decrease in gross returns over
the study period is 5.15 percentage points
(Table 1). However, if the calculation uses
records that show a much larger decrease in
costs, the decrease in gross returns is more
dramatic - 10.16%.
The decline in interest yields: The decline
in interest rates can explain part of the
decrease in the net returns (interest received is
Subject Matters
Annual Continuous-Time CTA and Commodity Pool Trading
Return Statistics Before and After 1991
a
The source is the indices of Zurich Captial Markets (2002). The returns are calculated as the natural
logarithm of the ending index value minus the natural logarithm of the initial value divided by the
number of years.
The source is Brorsen and Irwin (1985). The costs are for public funds.
c
The source is Brorsen (1998). The costs are for commodity trading advisors.
d
Averages of the annualized returns for one-year U.S. Treasury bills by Federal Reserve Bank of St. Louis
reported and then converted to a continuous-time return.
e
The standard deviation of monthly returns was computed by year for each CTA that listed their trading
system as trend-following, trend-identifier, or mechanical. The statistics reported are the simple average
of these annual standard deviations.
f
Calculated using the CTA value-weighted return of 14.65 minus interest of 6.75 plus cost of 13.85. The
cost number is calculated as the public fund cost of 19.20 minus the difference in CTA value weighted
returns and public fund returns (14.65-9.30 = 5.35) and thus assumes the entire difference in CTA and
public fund returns is due to differences in cost.
g
Calculated using the CTA value-weighted return of 9.32 minus interest of 3.80 plus cost of 10.00. The
result of 15.52 is then multiplied by the ratio of standard deviations (8.05/7.54) to adjust for possible
lower leverage in the more recent period, which gives the result of 16.57
b
“DATA CLEARLY SUGGESTS
THE RETURNS TO MANAGED
FUTURES FUNDS AND CTAS
HAVE DECREASED BECAUSE
OF
THE
DIMINISHING
SUCCESS OF FOLLOWING
TECHNICAL
TRADING
STRATEGIES.”
Table 1.
Return Characteristics Before and After 1991 for the Two CTAs
That Were the Largest Before 1991
a
Variance of Daily and 20-Day Returns for Futures Prices
Variance of Daily Returns
Variance of 20-Day Returns
1975 or the first date in the time series.
Notes: Hypothesis tests were performed using the two sample stationary bootstrap with 1,000 repetitions.
Statistically significant increases are denoted by+at .10 level and ++ at .05 level.
Statistically significant decreases are denoted by * at .10 level and ** at .05 level.
Table 2.
a
Notes: Hypothesis tests were performed using the two sample stationary bootstrap with 1,000
+
repetitions. Statistically significant increases are denoted by at .10 level and ++ at .05 level.
Table 3.
6
May 2004
35
Subject Matters
Frequency and Mean of Breakaway Gaps in Futures Prices
a
Statistically significant increases are denoted by+ at .10 level and ++ at .05 level.
Table 4.
Mean and Variance of Close-to-Open Changes in Futures Prices
a
Statistically significant increases are denoted by +at .10 level and ++ at .05 level.
Table 5.
Skewness and Kurtosis of Daily Returns for Futures Prices
Variance and Skewness of Breakaway Gaps in Futures Prices
a
a
Table 6.
36
Table 7.
May 2004
Subject Matters
down 3% per year), but is not enough to
explain the overall decline in returns.
further evidence that reduced price variability
was the primary change in futures markets.
The adoption of less risky trading strategies by larger funds: In terms of size, two
funds operating in the early period clearly
dominated the others. Returns and standard
deviations for these two funds are reported in
Table 2. Risk does show a substantial decline
which supports the assertion that the larger
funds changed their trading methods. But for
both funds, the decline in trading returns was
much greater than the decline in standard
deviation. While a decline in risk may explain
a portion of the decline in returns, it cannot
explain all of it.
However, although many variables show evidence of change, there are still a few statistics
that remained generally the same. The average
size of close-to-open price changes and
breakaway gaps did not consistently change.
The skewness of returns, gaps and close-toopen changes changed in only a few commodities.
The data clearly suggests the returns to managed futures funds and CTAs have decreased
because of the diminishing success of following technical trading strategies.
Structural changes to price
movements
Any change in the way futures prices fluctuate
could change the returns to technical trading.
Technical trading systems developed prior to
the change may therefore be obsolete.
For example, three livestock commodities do
indeed show decreased frequency of limit
moves: pork bellies fell from 17% to 10%,
feeder cattle from 5.5% to 2.2% and live cattle from 5.2% to 1.4%. Tables 3 to 7 present
several other statistical measures and provide
Wade Brorsen is a Regents Professor and
Jean & Patsy Neustadt Chair in the
Department of Agricultural Economics at
Oklahoma State University.
(Full details of the study will appear in the
Journal of Economics and Business, MayJune 2004, pp. 159-176).
References
Conclusions
The two dominant changes are a decrease in
price volatility and an increase in the kurtosis
of price changes occurring while markets are
closed. These changes are consistent with the
reduced profitability of technical trading
being due to changes in the overall economy.
The results are not consistent with increased
technical trading having caused the structural
change, because in that case price volatility
should have increased. If economic conditions change so that futures prices become
more volatile, then presumably returns to
technical trading would increase.
Billingsley, R.S. and Chance, D.M. (1996). Benefits
and limitations of diversification among commodity trading advisors. Journal of Portfolio
Management, 23, 65-79.
Brorsen, B.W. and Irwin, S.H. (1987). Futures funds
and price volatility. The Review of Futures Markets,
6, 118-135.
Center for International Securities and Derivatives
Markets (CISDM) (2000). CISDM Database.
University of Massachusetts at Amherst.
Pendley, K. and Zurla, K. (2002). Managed futures:
Performance hinders growth but still outshines
equities. Futures Industry, July/August, 20-22.
Willis Kidd is a former graduate research
assistant in the Department of
Agricultural Economics at Oklahoma
State University.
May 2004
37
Subject Matters
NONLINEARITY FAVOURS
NONLINEAR TA TECHNIQUES
by Kian-Ping Lim and Venus Khim-Sen Liew
Understanding the characteristics
of nonlinearity in financial markets is crucial to the development
of technical analysis - it may help
us determine which TA techniques will work better than others. Using data from the SouthEast Asian stock and currency
markets, the authors describe nonlinearity, show evidence for its
existence and talk about its implications for TA.
S
tudies into the financial markets have been
dominated by the linear paradigm. The linear paradigm assumes that financial timeseries conform to linear models, which may be
best represented by a continuous straight line
on a graph. Moreover, it is widely believed
that (1) stock prices can be well approximated
by the Capital Market Pricing Model (CAPM)
and the Arbitrage Pricing Theory (APT)
Model; and (2) foreign exchange rate movements may be determined by the Purchasing
Power Parity (PPP), Interest Rate Parity (IRP)
and other fundamental models. All these are
commonly applied linear models.
However, there is no solid reason to suppose
that movements of financial prices must be
intrinsically linear. One can hardly find a continuous straight line on any graph of any
financial price. Rather, one normally comes
across combinations of wave-like or even U-,
S-, V-, W- and L-shaped patterns. In fact,
some recent empirical studies have shown that
financial variables exhibit nonlinear movements. (All movement that exhibits polynomial (e.g. square, cubic), exponential, logistic
and/or trigonometric (wave-like) functions
are nonlinear in character). Among others,
Nicholas Sarantis found empirical evidence
on the nonlinear behavior in the stock prices
of seven major industrial economies (G-7),
whereas Christopher Baum and others have
detected the presence of nonlinear adjust-
38
ment dynamics in developed foreign exchange
markets.
The potential existence of nonlinearity in
financial markets is due to the dynamic, asymmetric and heterogeneous behavior of market
players in their response to market information and conditions. Another cause is the nonproportional speed of adjustment in financial
prices. A relatively small rise in a financial
price, for example, may not be transformed
into arbitrage action if such profit is not large
enough to cover the trading cost. Thus, small
changes in prices will be left unadjusted.
However, once the price exceeds certain limits
whereby marginal profits meet the players'
expectations, they will certainly adjust. In such
case, the larger the price change, the faster the
market adjustment.
Our study
This study reports the empirical evidence for
nonlinear behavior in the stock prices and
spot exchange rates of the major developing
South-East Asian (SEA) financial markets.
Our stock price data consists of the daily
growth rates (from the close of one day to the
Growth Rates
next) of the Jakarta Composite Index (JCI),
the Kuala Lumpur Composite Index (KLCI),
the Philippines Composite Price (PCOMP),
the Singapore Straits Times Index (STI) and
the Stock Exchange of Thailand (SET). As
for currencies, we study the daily growth rates
of the spot US dollar based exchange rates of
the Indonesia rupiah (IDR), the Malaysia ringgit (MYR), the Philippines peso (PHP), the
Singapore dollar (SGD) and the Thai baht
(THB).
General stock markets characteristics
The results of our study on SEA stock prices
are summarized in Table 1. Table 1 shows that
on average the sum of daily positive growth
rates in stock prices are slightly more than
those of negative growth rates in the KLCI
(mean=0.002%) and the STI (0.005%). On
the other hand, the reverse is true for the JCI
(- 0.001%), the PCOMP (- 0.003%) and the
SET (- 0.038%). However, as the mean value
is susceptible to bias due to outliers (unordinary large or small values), the median value,
which is not affected by outliers, is more reliable. By this value, all stock indices exhibit
zero average daily growth rates, implying
JCI
KLCI
PCOMP
STI
SET
2/1/1990
31/10/2001
3087
– 0.001
0.000
13.128
–12.732
1.558
2/1/1990
31/10/2001
3087
0.002
0.000
20.817
– 24.153
1.715
2/1/1990
31/10/2001
3087
– 0.003
0.000
16.178
– 9.744
1.665
2/1/1990
31/10/2001
3087
0.005
0.000
14.868
– 9.672
1.358
2/1/1990
31/10/2001
3087
–0.038
0.000
11.350
– 10.028
1.889
0.485
14.195
0.000*
0.461
36.898
0.000*
0.5571
11.558
0.000*
0.201
14.070
0.000*
0.287
7.432
0.000*
0.006
0.000*
0.006
0.000*
0.028
0.000*
0.006
0.000*
Summary Statistics
Sample period
No. of observations
Mean
Median
Maximum
Minimum
Standard deviation
Normality Analysis
Skewness
Kurtosis
JBN Test (p-value)
Linearity Tests
HBL Test (p-value)
LST Test (p-value)
0.036
0.000*
Table 1. Summary statistics, normality analysis and linearity test results of growth rates of
major SEA stock prices
May 2004
Subject Matters
"trading rules based on techniques such as Elliott Waves
and head-and-shoulders, which already incorporate some
sense of nonlinearity, may be more reliable or profitable.
In contrast, reservations must be made for technical
analysis that uses linear autoregressive, moving average
or exponential moving average models."
break-even investment or arbitrage activities
in these markets in the long term, even though
profits or losses may be encountered by
investors or arbitrageurs in the short term.
By referring to the maximum and minimum
values in Table 1, we know that the positive
and negative daily growths rates in the SEA
region have recorded a peak of 20.817% and
24.152% respectively, both from the
Malaysian stock market. Regarding investment
risk, the standard deviation value shows that
the Thailand stock market is the riskiest (standard deviation=1.889) followed by the
Malaysian market (1.715), whereas the
Singapore stock market (1.358) has the least
risk.
Evidence for nonlinearity in stock prices
More interestingly, the skewness values in
Table 1 show that these financial prices have
asymmetrical distribution (0 in normal condition). It is therefore not surprising to observe
that the normality assumption (i.e. symmetrical bell-shaped distribution) has been strongly
rejected by the Jarque-Bera normality (JBN)
test (see Box 1).
However, this finding cannot be taken as statistical evidence for nonlinearity. Two formal
tests capable of distinguishing the linearity or
nonlinearity of our growth rates - the HinichBispectrum Linearity Test (HBL Test) and the
Lukkonen-Saikkonen-Teräsvirta Test (LST
Test) - are applied in this study. Briefly, the
former is able to detect the existence of nonlinear self-dependencies whereas the latter is
useful in capturing the potential nonlinear
adjustment of stock or forgien exchange market dynamics.
The results of the HBL linearity test (Table 1)
show that stock prices exhibit strong nonlinear dependencies on their own past records.
The LST test tells us more and suggests the
nonlinearity present can be characterized by a
type of nonlinear time-series model (the
Smooth Transition Autoregressive (STAR)
model) which captures bell- and S-shaped
nonlinear adjustments (exponential and logistic nonlinearity).
IDR
MYR
PHP
SGD
THB
16/11/1995
31/12/2002
1759
0.078
0.017
30.189
– 23.316
2.512
2/1/1990
31/8/2001
2179
0.020
0.000
7.196
– 9.157
0.695
16/11/1995
31/12/2002
1773
0.040
0.008
7.176
– 12.518
0.776
2/1/1990
31/12/2002
3267
– 0.003
0.000
2.762
– 4.144
0.356
2/1/1990
31/12/2002
3210
0.016
0.000
20.769
– 6.353
0.753
1.172
34.783
0.000*
– 0.084
43.141
0.000*
– 1.304
54.518
0.000*
– 0.908
20.905
0.000*
6.185
196.243
0.000*
0.006
0.000*
0.006
0.000*
0.010
0.000*
0.006
0.000*
0.006
0.000*
Summary Statistics
Sample period
No. of observations
Mean
Median
Maximum
Minimum
Standard deviation
Normality Analysis
Skewness
Kurtosis
JBN Test (p-value)
Linearity Tests
HBL Test (p-value)
LST Test (p-value)
Table 2. Summary Statistics, normality analysis and linearity test results of growth rates of
major SEA exchange rates
May 2004
39
Subject Matters
Brief note on the test statistics
p-values show how far we can reject a
null hypothesis. Conventionally, a
p-value of 0.10 or larger is taken as
evidence for not rejecting a null hypothesis. In the tests listed in Tables 1 and 2
the p-values are significantly less than
0.10 and result in the rejection of the
following null hypotheses:
JBN test: Normality does exist
HBL test: Nonlinearity does not exist
LST test: STAR-type nonlinearity does
not exist
Box 1.
General foreign exchange markets characteristics
Table 2 depicts the results of our study on the
SEA daily spot exchange rates. Table 2 shows
that, on average, the sum of daily positive
growths (depreciation) are either equal to or
more than those of negative growths (appreciation) in all exchange rates (mean/median
are zero or positive in value) except the
Singapore dollar, which recorded an average
mean value of - 0.003% (median=0, howev-
40
er). It should be noted that while the median
for other foreign exchange rates are 0 in value,
implying break even long term arbitraging and
hedging activities, there are two rates namely
the rupiah (median=0.078%) and peso
(0.008%) where investors holding short positions may be slightly better off. However, it is
worth knowing that the peso has the highest
exchange rate risk (standard deviation =
2.512, maximum = 30.189, minimum = 23.316; much riskier than its stock price index)
in the SEA region. As for the other exchange
rates, they have lower risk than their corresponding stock indices.
behavior, as found in this study, has crucial
implications for financial market researchers
and practitioners alike. We can no longer take
the linear assumption for granted. Hence, linear models like the CAPM and APT (for
stocks), and the PPP and IRP (for exchange
rates) may need modification before they can
be readily applied, at least to the South-East
Asian region. To this end, trading rules based
on techniques such as Elliott Waves and headand-shoulders, which already incorporate
some sense of nonlinearity, may be more reliable or profitable. In contrast, reservations
must be made for technical analysis that uses
linear autoregressive, moving average or exponential moving average models.
Evidence for nonlinearity in foreign
exchange rates
Similar to the findings in the stock markets,
the skewness and JBN values show that the
SEA stock exchange rates are asymmetrical
and thus non-normal in nature. On top of
that, the results from the HBL and LST linearity tests provide strong evidence of nonlinear
behaviour in all these exchange rates.
More reliable technical trading rules
The existence of nonlinear financial price
May 2004
Kian-Ping Lim is a lecturer in the Labuan
School of International Business and
Finance, Universiti Malaysia Sabah,
Malaysia.
Venus Khim-Sen Liew is a research assistant at the Faculty of Economic and
Management, Universiti Putra Malaysia,
Malaysia
Book review
ADVANCED SWING TRADING
A handful of books on swing trading have appeared in
recent years, in particular Marc Rivalland's book published by Harriman House. In contrast to Rivalland's
work, Crane's book assumes the reader has a good
working knowledge of charting techniques and takes a
more in-depth and systematic approach to the various
techniques available to the swing trader.
Swing trading is designed to take advantage of shortterm movements in longer-term market trends. This
allows more pro-active traders to identify and exploit
profit making opportunities in normal market volatility.
Crane calls such movements "reaction swings" and he
goes on to explain how they allow the trader to identify
not only price objectives and stop loss levels, but also
the timing of future moves or "reversal dates".
Advanced Swing Trading
Strategies to predict, identify, and trade
future market swings
By John Crane
Published by John Wiley
219 pages, £45.50
ISBN 0-471-46256-X
Reaction swings resemble flags in the middle of a trend
and are based on the theory of action/reaction. This is
the basic pattern of a trending market (action) which
corrects and then resumes its trend (reaction). The
analysis of this reaction swing allows the trader to anticipate subsequent moves and time future reversals in the
current price trend. The mid-point of the reaction swing
can be used to make projections by drawing a line from
the start of the main trend through the mid-point of the
swing. This accurately pinpoints both the price and timing of the next price reversal.
The reaction swing is familiar to technical analysts as a
flag or pennant. These usually indicate a mid-point of a
price trend but offer only limited information regarding
the timing of a price objective or future turning points.
The reaction swing is a simple method that can make
greater use of the information inherent in these patterns.
The reaction cycle represents the bigger picture allowing the trader to identify the start of a market trend, its
mid-point and the end of the major trend. In an uptrending market, identifing the first reaction swing is crucial in
establishing the beginning of a reaction cycle. Having
discussed the reaction swing and cycle, Crane goes on
to discuss the application of Gann lines in relation to
action/reaction theory and also presents a useful
overview of trading "hints" such as reversal price patterns, gapping patterns and trail days.
This book is written for the experienced trader who
May 2004
41
Book review
already has a solid understanding of technical analysis.
While swing trading is a relatively new technique for
trend analysis, it is not as groundbreaking as some
recent books have implied. Swing trading relies on the
application of simple and established TA techniques and
as such, this trading method is not as advanced as
Crane's book title suggests.
In the text, the author readily admits to being an inexperienced writer and this shows in his descriptions of the
various techniques he presents. He therefore relies on
numerous examples to illustrate the subject with little or
no reference to the market psychology that is behind the
behaviour of the reaction cycle. The problem with a
heavy reliance on market examples is that the reader
has no way of knowing how carefully chosen the exam-
ples were by the author. A more scientific and theoretical approach to the subject would greatly enforce the
arguments presented in this and numerous other technical analysis publications.
This is not an easy book to read and understand. The
principles that Crane attempts to describe in each chapter are not always as clearly explained as they should
be. Moreover, the time required to master the numerous
variations of the swing trading techniques described in
the book is probably more than most traders would consider, especially as he or she is doubtless already using
other techniques in their trading strategies. However,
the numerous charts are well presented and Crane's
publication adds to the relatively limited amount of literature on a valuable subject.
Letters
LETTERS TO THE EDITOR
SIR: I would like to draw your readers' attention to the dangers
of data snooping - a phenomenon that I believe should be
considered in all assessments of TA techniques.
Data snooping is the generic term for the danger that the best
forecasting model found in a given data set by a certain specification search is just the result of chance instead of the result
of truly superior forecasting power. Jensen (1967) already
argued that the good results of the relative-strength trading
rule used by Levy (1967) could be the result of survivorship
bias. That is, strategies that performed well in the past get the
most attention by researchers. Jensen and Benington (1969) go
a step further and argue, "Likewise given enough computer
time, we are sure that we can find a mechanical trading rule
which works on a table of random numbers - provided of
course that we are allowed to test the same rule on the same
table of numbers which we used to discover the rule. We realize of course that the rule would prove useless on any other
table of random numbers, and this is exactly the issue with
Levy's results."
Another form of data snooping is the publication bias. It is a
well-known fact that studies presenting unusual results are
more likely to be published than the studies that just confirm a
well-known theory.
The problem of data snooping was addressed in most of the
work on technical analysis, but for a long time there was no
test procedure to test for it. Finally White (2000), building on
the work of Diebold and Mariano (1995) and West (1996),
developed a simple and straightforward procedure for testing
the null hypothesis that the best forecasting model encountered in a specification search has no predictive superiority
over a given benchmark model. The alternative is of course
that the best forecasting model is superior to the benchmark.
Summarized in simple terms, the procedure bootstraps the
original time-series a great number of times, preserving the
key characteristics of the time-series. White (2000) recommends the stationary bootstrap of Politis and Romano (1994).
Next, the specification search for the best forecasting model is
executed for each bootstrapped series, which yields an empirical distribution of the performance of the best forecasting
model. The null hypothesis is rejected at the alpha percent significance level if the performance of the best forecasting
model on the original time series is greater than the alpha percent cut-off level of the empirical distribution. This procedure
is called White's Reality Check (RC) for data snooping.
Sullivan, Timmermann and White (1999, 2001) utilize the RC
to evaluate simple technical trading strategies and calendar
effects applied to the DJIA in the period 1897 to 1996.
Sullivan et al. (1999) take the study of Brock et al. (1992) as a
starting point and construct an extensive set of 7846 trading
rules, consisting of Alexander's (1961) filters, moving averages,
support-and-resistance, channel break-outs and on-balance
volume averages. It is demonstrated that the results of Brock
et al. (1992) hold after correction for data snooping, but that
the forecasting performance tends to have disappeared in the
period after the end of 1986.
For the calendar effects (for example the January, Friday and
the turn of the month effect) Sullivan et al. (2001) find that
the RC in all periods does not reject the null hypothesis that
the best forecasting rule encountered in the specification
search does not have superior predictive ability over the buyand-hold benchmark. If no correction were made for the
specification search, then in both papers the conclusion would
have been that the best model would have significant superior
forecasting power over the benchmark. Hence Sullivan et al.
(1999, 2000) conclude that it is very important to correct for
data snooping otherwise one can make wrong inferences about
the significance of the best model found.
Gerwin Griffioen, analyst,
Insinger de Beaufort Asset Management,
The Netherlands
May 2004
43
COMMITMENTS OF TRADERS REPORT
16 January 2004 – 4 May 2004
Non-commercial net long positions and spot rates
10-year US Treasury
10-yr Treasury
Spot
5-year US Treasury
Source: CBOT
-250000
4.80
5-yr Treasury
Spot
Source: CBOT
300000
3.80
4.60
-200000
3.60
250000
4.40
-150000
3.40
4.20
-100000
200000
3.20
4.00
150000
3.80
-50000
3.60
3.00
100000
0
2.80
3.40
50000
50000
2.60
3.20
100000
3.00
Jan-13
Feb-10
Mar-09
Dow Jones Industrial Average
Apr-06
DJIA
Spot
0
May-04
2.40
Jan-13
Feb-10
Swiss franc
Source: CBOT
-6000
11000
Mar-09
Swiss franc
Apr-06
Spot
May-04
Source: CME
20000
1.32
15000
-5000
1.3
10800
10000
-4000
1.28
10600
5000
-3000
1.26
10400
0
-2000
1.24
10200
-5000
-1000
10000
0
9800
1000
2000
9600
Jan-13
44
Feb-10
Mar-09
Apr-06
May-04
May 2004
1.22
-10000
1.2
-15000
1.18
-20000
Jan-13
Feb-10
Mar-09
Apr-06
May-04
Pound sterling
Pound sterling
Spot
Yen
Source: CME
25000
1.95
20000
1.90
15000
1.85
Japanese yen
Spot
Source: CME
80000
112
111
60000
110
109
40000
108
107
20000
10000
106
1.80
0
105
5000
104
1.75
-20000
103
0
1.70
Jan-13
Feb-10
Euro
Mar-09
Euro
Apr-06
Spot
-40000
102
Jan-13
May-04
Feb-10
3-month eurodollar
Source: CME
35000
1.30
Mar-09
3-month eurodollar
Apr-06
Spot
May-04
Source: CME
500000
1.12
400000
1.10
30000
1.28
300000
25000
1.08
200000
1.26
100000
1.06
20000
1.24
0
1.04
15000
-100000
1.22
-200000
10000
1.02
-300000
1.20
5000
1.00
-400000
0
1.18
Jan-13
Feb-10
Mar-09
Apr-06
May-04
-500000
0.98
Jan-13
May 2004
Feb-10
Mar-09
Apr-06
May-04
45
Nasdaq
Nasdaq
Spot
Nikkei
Source: CME
2200
-25000
Nikkei
Spot
Source: CME
2200
-25000
2150
-20000
2150
-20000
2100
2050
-15000
2100
2050
-15000
2000
-10000
2000
-10000
1950
1900
-5000
1950
1900
-5000
1850
0
1850
0
1800
5000
1750
Jan-13
Feb-10
Gold
Mar-09
Gold
Apr-06
Spot
1800
5000
May-04
1750
Jan-13
Feb-10
US dollar Index
Source: CEI
160000
Mar-09
US dollar index
Apr-06
Spot
May-04
Source: NYCE
430
-12000
117.00
420
-10000
116.00
410
-8000
115.00
400
-6000
114.00
390
-4000
113.00
380
-2000
112.00
370
0
140000
120000
100000
80000
60000
40000
20000
0
Jan-13
46
Feb-10
Mar-09
Apr-06
May-04
May 2004
111.00
Jan-13
Feb-10
Mar-09
Apr-06
May-04
Non-commercial
April 12
April 20
April 27
May 4
10yr Treasury
-100749
-100749
-191271
-165896
5yr Treasury
118310
118310
100899
130465
561
561
1203
- 222
Swiss franc
-5941
-5941
-11092
-7089
Pound Sterling
10648
10648
3732
4730
Japanese yen
24760
24760
204
2973
Euro
5743
5743
5570
6999
3m eurodollar
-67991
-67991
-312512
-367205
Nasdaq
-7263
-7263
-8909
-7205
Nikkei
3692
3692
-275
3850
138696
138696
57804
54663
-3835
-3835
-2735
-1543
DJIA
Gold
US$ index
Commercial
April 12
April 20
April 27
May 4
10yr Treasury
-100749
232784
-191271
278056
5yr Treasury
118310
9227
100899
-27468
561
2147
1203
2115
Swiss franc
-5941
17896
-11092
12794
Pound Sterling
10648
-9665
3732
-7257
Japanese yen
24760
-10506
204
-16933
Euro
5743
-14332
5570
-9741
- 67991
366523
-312512
532849
Nasdaq
-7263
18888
- 8909
21722
Nikkei
3692
-2214
-275
-5317
138696
-122430
57804
- 89398
-3835
304
-2735
-514
DJIA
3m eurodollar
Gold
US$ index
Charts and tables: Open interest (futures only)
All data provided by the Commodity Futures Trading Commission (CFTC) with permission
May 2004
47
Training and events diary
TRAINING AND EVENTS DIARY
22/23 June
4 June
9 June
Course:
Introduction to technical analysis
Organiser:
7City
Contact:
[email protected]
Event:
STA monthly meeting
Organiser:
Society of Technical Analysts
Contact:
[email protected]
Course:
An introduction to charting and
technical analysis
Organiser:
IPE
Contact:
[email protected]
23 June
7 July
19 July
Course:
Advanced technical analysis
Organiser:
The Oxford Princeton Programme
Contact:
[email protected]
Event:
STA monthly meeting
Organiser:
Society of Technical Analysts
Contact:
[email protected]
Course:
Organiser:
Quorum Training
Contact:
[email protected]
23 August
Late October
Late October
Course:
Organiser:
7City
Contact:
[email protected]
Course:
Technical analysis & charting
Organiser:
ChartWatch
Contact:
Market Directional Analysis
Tel: 020 7723 0684
Course:
Advanced t echnical analysis
Organiser:
ChartWatch
Contact:
Market Directional Analysis
Tel: 020 7723 0684
10 November
15 November
24/25 November
Course:
Organiser:
7City
Contact:
[email protected]
Course:
Organiser:
Quorum Training
Contact:
[email protected]
Course:
An introduction to charting and
technical analysis
Organiser:
IPE
Contact:
[email protected]
For training and events diary submissions please email us at: [email protected]
All venues are in London
48
May 2004

Issue 4 - The Technical Analyst

Transcription

Similar documents

News File

KERFORD INVESTMENTS (UK) LTD KERFORD INVESTMENTS

our Alpha API Brochure

WizeVest

Six Steps to Effectively Measure Analyst Performance

newismWhy more than ever, consumers lust after

7 Chart Patterns - Traders` Library

Ultimate Skills Checklist for Your First Data Analyst Job

why trend micro?

PDF Brochure

Tsunami-Trading - Traders World Magazine

High Accuracy Option Trading Prime Trade

the official magazine of technical analysis

Updated V1.3 of Retired CX TAI/CDI - Rae

Metastock Professional Brochure

TTT-based tests for trend in repairable systems data

Robert Half Technology Salary Guide 2014

Wounded Healers in Practice - AURA

Initial Operation of a Pulse-Burst Laser System for High

The Corporation As Insider Trader

Internalizing Outsider Trading - Yale Law School Legal Scholarship

Sep/Oct 2008 - Traders.com