Issue 23 - The Technical Analyst

Transcription

july/aug 2007
The publication for trading and investment professionals
www.technicalanalyst.co.uk
Trading Time
Waiting for the mega trend
Markets
Software
Interview
Outlook for
EUR/USD
Algorithm backtesting
with Progress
Aaron Brown of
Morgan Stanley
There’s been
something missing
in the UK.
Until now.
The new Lyxor ETF
FTSE All-Share
Bloomberg code: LFAS LN<Equity>
The new Lyxor ETF FTSE All-Share is the first Exchange
Traded Fund to track the whole FTSE in one trade. So
now you can access instant and diversified exposure to
the UK stock market, through a single investment traded
in real time on the London Stock Exchange.
We are also pleased to announce the launch of a further
two Lyxor ETFs in our FTSE range:
Lyxor ETF FTSE 100 – Bloomberg code:
L100 LN <Equity>
Lyxor ETF FTSE 250 – Bloomberg code:
L250 LN <Equity>
Lyxor Asset Management is one of Europe’s largest ETF
providers and is 100% owned by the Société Générale
Group.
For more information
visit www.lyxoretf.co.uk
or email [email protected]
This advertisement is issued in the UK by Société Générale authorised by the Banque de France and regulated by the Financial Services Authority for the conduct of UK business. Lyxor ETFs
are open-ended mutual investment funds established under French Law and approved by the Autorité des Marchés Financiers. The funds are UCIT III compliant however only those funds
recognised under S.264 of the Financial Services and Markets Act 2000 may be promoted to retail investors in the UK. A list of recognised funds with prospectus is maintained on
www.lyxoretf.co.uk. Any investment in Lyxor ETFs carries with it certain risks set out in the Prospectus’. Lyxor ETFs are not suitable for all investors, it is recommended that potential investors study
the Prospectus and seek their own independent financial advice before making any decision to invest in Lyxor ETFs. SG Option Europe may be the only market maker. Investors’ capital is at risk.
WELCOME
The importance of being aware of the technical picture in all time frames is often
emphasised by successful traders. For example, if you are trading short term
then major technical levels may exist out of your time frame that should be considered. In this issue, we look at one approach taken by Shaun Downey at CQG
as to how best to use ‘Time’ as part of an effective trading strategy.
We hope you enjoy this edition of the magazine
Matthew Clements, Editor.
CONTENTS 1 > FEATURES
JULY/AUG
Trading Time
In an excerpt from his new book, Shaun
Downey explains the importance of using time
as part of an effective trading strategy
Interview
Aaron Brown of Morgan Stanley
talks poker and the markets
Software
John Bates of Progress explains the
requirements, challenges and approaches that
should be considered in backtesting
algorithmic trading
© 2007 Global Markets Media 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 Global Markets Media 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.
July/August 2007
>20
>30
> 34
>>
THE TECHNICAL ANALYST
1
presents
Automated Trading
Strategies for Building an
Automated Trading System
11 October 2007
1 Wimpole Street,
2007
London W1
A premier event for trading and investment professionals
Automated Trading 2007 is the essential one day
conference for traders and investment managers
looking to develop and build their own automated trading system. Bringing together the
best international experts from banks and hedge
funds, the conference will provide delegates
with the perfect opportunity to learn about and
discuss all areas of model development, system
building and backtesting.
Who should attend:
Topics Covered:
+ Proprietary traders
+ Fund managers
+ Hedge funds
+ Dealers
+ Quantitative analysts
+ Algorithmic trading managers
+ System developers
+ Testing & optimisation
+ FX model strategies
+ Stock market trend models
+ Performance measures
+ High frequency trading
+ Uncorrelated models
+ Troubleshooting panel
Speakers include:
Max Dupont
Quantam
Raspal Sohan
Rathbones
Melanie Schmidt
Updata
Rami Habib
Kyte Group
Charles Morris
HSBC
Luc Van Hof
Aim Trading
Delegate fee: £445 + VAT
Register by 31 July for £100 early bird discount = £345 + VAT
Register Today!
Telephone: +44 (0)20 7833 1441
Web: www.technicalanalyst.co.uk Email: [email protected]
24
41
CONTENTS 2 > REGULARS
Editor: Matthew Clements
Managing Editor: Jim Biss
Consultant Editor: Trevor Neil
Advertising & subscriptions:
Louiza Charalambous
Marketing: Vanessa Green
Events: Adam Coole
Design & Production:
Paul Simpson & Thomas Prior
The Technical Analyst is published by
Global Markets Media Ltd
Unit 201, Panther House,
38 Mount Pleasant,
London WC1X 0AN
Tel: +44 (0)20 7833 1441
Web: www.technicalanalyst.co.uk
Email: [email protected]
INDUSTRY NEWS
04
MARKET VIEWS
Euro STOXX 50: Bull trend remains intact
EUR/USD: Bearish reversal signals?
US Interest Rates: Changing perceptions
07
09
11
ROUNDTABLE
Bond market outlook
14
TECHNIQUES
Technical analysis by numbers
Trading time
Candlestick signals: The J-hook pattern
Portfolio testing
18
20
24
27
INTERVIEW
Aaron Brown, Morgan Stanley
30
SOFTWARE
Algorithm backtesting, Progress Software
34
BOOKS
Trading Time by Shaun Downey
37
RESEARCH UPDATE
38
AUTOMATED TRADING SYSTEMS
Programming and Interoperability
Strategy spotlight: Stein Investment Management
41
44
SUBSCRIPTIONS
Subscription rates (6 issues)
UK: £160 per annum
Rest of world: £185 per annum
Electronic pdf: £49 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)
July/August 2007
3
Industry News
EX-CITY ANALYSTS LAUNCH PIA-FIRST RESEARCH
Three former City analysts have
recently launched a new technical
analysis research boutique to serve
investment houses, trading desks and
hedge funds. 'PIA-First' has been set
up by Max Knudsen and his colleagues, Steve Lucas and Alan Collins.
All three were formerly technical analysts at Dresdner Kleinwort in
London.
Knudsen says of the service's
launch, "Over the past seven years
working for an investment bank,
clients remarked on the uniqueness of
our research and its ease of use, simplicity and accuracy. It was this support and encouragement from clients
that prompted us to make the move
and develop PIA-First as an independent trading and trade timing service.
We are very encouraged by the
response we've had in the first six
weeks."
PIA-First provides ideas and advice
on trade timing across 26 markets,
(seven interest rate, 13 FX and six
equity indices) producing calls for 3
separate time frames: daily, weekly and
3 months. These are all published on
the PIA website as well as by email
bulletins. In addition, emails throughout the day update the calls in real
time. For a free 1 month trial go to:
www.pia-ffirst.com
CQG AND TULLETT PREBON
UNVEIL NEW DATA SERVICE
Tullett Prebon Information and CQG
have announced the launch of TPI's
new website for the resale of historical
market data, www.tphistory.com. The
site, built using CQG Data Factory
infrastructure and TPI's extensive historical market data content, will also
become available via CQG's Data
Factory website at:
www.CQGDataFactory.com. The content available for TPI's website currently includes sovereign debt, money
markets and interest rate derivatives.
Historical data is offered through the
site in tick-based, hourly, or end-of-day
frequencies.
The PIA-First
research team
Tick Data and CBOE Release
Options Data
Tick Data and the Chicago Board
Options Exchange (CBOE) have released
the industry's first commercially available
research-ready historical tick database for
the US equity options market. The new
database contains all US equity options
data from the consolidated Options Price
Reporting Authority (OPRA) feed dating
back to July 2, 2004.The database is
designed for the building, testing and validation of algorithmic options trading
models. The data can be used for preand post-trade analysis and optimization
of execution strategies.
UPDATA ANNOUNCES ESIGNAL COMPATIBILITY
Updata has announced the integration
of the eSignal data feed into its Updata
Technical Analyst software. In a move
which targets the high end 'Updata TA'
system at the lower tier professional
market and high net worth private
traders, Updata is aiming to extend its
reach, particularly in the system testing
and coding arena.
Commenting on this latest development, Updata's David Linton says
"Integrating eSignal presents us with an
opportunity to make our advanced sys4
tems available to a wider audience.
There is already a lot of eSignal users
out there globally, and a number of the
smaller funds and trading outfits see
this as a great way of being able to run
our systems on an affordable data feed
that still has a great breadth of coverage."
The eSignal compatibility is part of
Updata's latest software release which
includes divergence scanning, customised columns and multiple timeframe chart reporting.
July/August 2007
David Linton,
Updata
Free Bollinger
Ok, so we aren’t referring to the champagne,
but we’re still giving you a reason to celebrate.
Throughout September, Patsystems is offering
free integrated charting for one month, to all
Patsystems J-Trader and Pro-Mark users.
The advanced functionality is simple to use,
with multiple technical indicators – including
Bollinger bands – and flexible representations,
efficient and fast backtesting, and an intuitive
programming language.
This could be just the taster you need.
For details call us on +44 (0)20 7940 0470 or
visit our website at www.patsystems.com.
Market Views
EURO STOXX 50
BULL TREND REMAINS INTACT by Cyril Baudrillart
B
ack in October 2006 the
consensual technical outlook for the equity markets
was for a continuation of the bull
trend which later proved to be
pretty good advice. The technical
outlook today is not very different.
The trend remains our friend and
investors have continued to buy
the dips quite aggressively during
setbacks.
The goal of this analysis is not simply
to study the index's trend and targets
but also to evaluate the current shape
of European equity markets via intermarket analysis and market breadth
indicators. But let's start with a basic
trend analysis.
Figure 1.
Trend following methods
Over the last few years, I have been
monitoring developments of the bull
market through objective trend-tracking methods using weekly and monthly
prices. The methods include point &
figure charts, Ichimoku, Daryl Guppy's
multiple moving averages and, of
course, classical simple moving averages such as 200, 100 and 50-day moving averages. According to most of
these methods, the primary trend
turned from bearish to bullish in the
second half of 2003 and, so far, this
bull trend has remained intact (Figure
1).
These trend following indicators are
intended to help us trade in the direction of the primary trend. They will
never identify market peaks, but that is
not their goal. Before studying a few
market-timing indicators, let's focus on
the next upside targets for the medium
term. The index has now retraced more
than 76.4% of the 2000-2003 downtrend on a semi-log scale. This increases the probability of retesting the alltime high of 5522, attained in March
2000. According to some point & figure charts, the next short-term target
could be in the 4758-4800 area. A break
through the 4560-area, which corresponds to the upper end of the ascending triangle forming since early June,
would increase the probability of
reaching this zone. Another potential
target is located at 5332 points, which is
2.618 times 1847 points, the intraday
low of March 2003 (the next Tom
DeMark Absolute Retracement™).
July/August 2007
TD Combo™
Now let us focus on market-timing
indicators. As markets never move in
straight lines, technicians must always
hold a few contrarian indicators in their
toolbox to identify overbought zones.
In sustained bull trends, classical
oscillators such as RSI and stochastics
do not help as they can stay overbought
or build multiple bearish divergences
for a long time. In such trends, most
contrarian indicators give mixed results;
the exceptions include Tom DeMark's
TD Combo™ and TD Sequential™,
which help to identify low-risk sell
areas, or at least to limit losses on contrarian trades. The daily TD Combo™
correctly identified the peaks of
November 2006 and February →
7
Market Views
such as auto and chemicals, the index's
market breadth has deteriorated significantly since early June. The percentage
of Euro STOXX 50 stocks trading
above their 200-day MA has declined to
70%, from 90% a month ago. The
Bullish Percent Index calculated on
STOXX 600 stocks also highlights this
deterioration. This indicator measures
the percentage of the index's components that are bullish according to a 1%
x 3 point & figure chart, i.e., where the
“…THE PRIMARY TREND
TURNED FROM BEARISH
TO BULLISH IN THE SECOND HALF OF 2003 AND
HAS REMAINED INTACT.”
Figure 2.
2007. A new TD Combo™ 13-sell
countdown was completed on 31 May.
This signal has not been disqualified as
it would require a daily close above
4576, followed by an open above the
previous day's close. As long as a break
through this level does not occur, a
continuation of the current consolidation period will be favoured according
to this method.
Weaker dollar and bonds
A deterioration of the macro environment supports the possibility of mixed
performances on the equity markets in
the coming months. First, the mid-term
trend on bond yields has reversed significantly worldwide. In Germany, 10year government bond yields may reach
4.90% in the coming months.
Second, more recently European corporate spreads have been widening.
The technical bounce of the iTraxx
Europe index (Figure 2) is now more
significant than that of late-February
but, unlike in March, this rise has not
triggered a significant correction of
equities. Nevertheless, the trend of
implied volatility indices such as the
VDAX and VSTOXX in Europe has
already turned upwards. For the first
8
time since the late-1990s, stock markets
are advancing with implied volatility rising, an indication that they have entered
a speculative phase.
Third, the euro dollar's trend remains
positive and the odds are still in favour
of an upside breakout through the
1.3680-resistance zone with a possible
target in the 1.40-area.
Oil and gas support equities
In light of this mixed macro outlook,
we have to admit that the recent stability of the Euro STOXX 50 index is
quite impressive. A study of sector
trends helps to explain the absence of a
sharp reaction by the index to the
recent rise in corporate spreads.
This robustness is mainly due to the
strong re-rating of the oil & gas sector,
which has fully overshadowed the
weakness of financials. It is also worth
noting the persistent weakness of the
healthcare sector. We have to go back
to the late-1990s, i.e., near the end of
the TMT bubble, to find a similar sustained under-performance of this
defensive sector. Investors are capitulating again.
Excluding commodity-sensitive
stocks and a few other cyclical sectors
July/August 2007
last signal is at least a double top buy.
According to this index, the percentage
of stocks that are on a bull trend has
declined sharply, from 85% in May to
below 60% in early-July. Nevertheless,
the index is holding above 50%, which
signals that the overall trend remains
bullish.
Conclusion
The index has entered a short-term
consolidation period that may eventually continue through the summer in the
event of weak bond markets and dollar.
However, a drop below the 4330-support zone would be required to confirm the risk of a more pronounced
setback between 4200 and 4100.
Alternatively, a break through 45604576, which includes the upper end of
an ascending triangle, could restore the
bull trend with the next targets at
around 4800 in the near term and
maybe 5332-5522 in the long term.
As the overall trend remains positive,
we cannot anticipate a major reversal in
the trend at this stage, even though the
equity markets have probably entered a
more speculative phase.
Cyril Baudrillart is European equities technical analyst at Exane BNP
Market Views
EUR/USD
ELLIOTT WAVE SUGGESTS BULL MARKET
IS NEARING COMPLETION by Andrew Chaveriat
E
lliott wave analysis suggests the
October 2000 EUR/USD bull
market is drawing to an end.
The bull market appears in its final fifth
wave, targeting completion ideally
between 1.3925-1.4225 during late
August to mid-September 2007. This
wave five high should complete the bull
market and mark the beginning of a
new bear market targeting a long-term
decline towards the 1.2485 October
2006 low and eventually the 1.1640
November 2005 low, representing support from the fourth wave of one less-
er degree (wave IV low). See Figure 1.
The current wave V rally off the
1.1640 November 2005 low is subdividing into the requisite five wave pattern. The wave 3 of V rally ended at the
1.2980 June 2006 high. Wave 4 of V
consisted of the June-October 2006
horizontal trading band between
1.2980-1.2485. Wave 5 of V originated
off the 1.2485 October 2006 low and is
forming a fifth wave extension. Waves
i of 5 (1.2485-1.3370), ii of 5 (1.33701.2865), iii of 5 (1.2865-1.3685) and iv
of 5 (1.3685-1.3265) are finished.
The wave v of 5 of V rally is now in
progress (off the 1.3265 June 2007
low). Overlap between the wave iv low
(1.3265) and wave i high (1.3370) indicates the October 2006 rally (wave 5 of
V) is forming a diagonal fifth wave triangle. This pattern, also known as rising wedge, portends a swift decline
once the current wave v of 5 of V rally
ends. As Frost and Prechter* note, "A
rising wedge… is usually followed by a
sharp decline retracing at least back to
the level where the diagonal triangle
began"; in this case the 1.2485 →
Figure 1. EUR/USD Weekly - Long-term Elliott wave count
July/August 2007
9
Market Views
Figure 2. EUR/USD Daily - Medium-term Elliott wave count
October 2006 low (see Figure 2).
The market top
As noted earlier, the probable target
zone for completing the wave v of 5 of
V rally is 1.3925-1.4225 during late
August to mid-September 2007. This
Elliott price target zone includes projections from waves of three different
degrees including the 127.2% retracement of wave IV (1.4220), 423.6%
retracement of wave 2 of V by waves
3-5 of V (1.3945), and the wave iii of 5
= wave v of 5 measured move (1.4085).
Finding targets from multiple time
frames grouped closely together
between 1.3925-1.4225 increases the
probability that the 2000 bull market
will terminate in this area.
The 1.3925-1.4225 target zone
includes testing the top of the October
2006 rising wedge now near 1.3850 and
projected to lie between 1.4015/50 during late August to mid-September 2007
when wave v of 5 of V is expected to
peak. That includes seeing the June
2007 wave v of 5 rally persist until it
10
measures 161.8% of the duration of
wave i of 5, and 100% of the duration
of wave iii of 5.
Weekly momentum
Weekly momentum is surging following
its bullish crossover during late June.
The strength of bullish weekly momentum is reminiscent of that during the
powerful October-November 2006
rally (8-weeks/+8.85-cents; wave i of
5). This implies scope for a sizable
medium-term spot advance: if the current June rise off 1.3265 matches the
October-November 2006 advance,
EUR/USD will hit 1.4150 in August
2007 reaching the 1.3925-1.4225 Elliott
target zone.
We suspect weekly momentum -- now
at 74% on the 8-week modified stochastic -- will rival the overbought conditions of December 2006 (84%) and
form bearish divergence with the April
2007 extreme (90%) in the weeks ahead
as EURUSD posts a major top.
July/August 2007
Bearish reversal points
Given EUR/USD has been rallying for
nearly seven years, it will take a clean
break of key support in order to confirm that a bear market is underway.
Initial signs of a bear market will likely
include a bearish weekly reversal signal
occurring in the favoured 1.39251.4225/late August to mid-September
2007 target zone, sparking a decline
that breaks daily support from the
recent 1.3415 June 27 low. Additional
confirmation of a bear market would
include a sustained break of the
February 2006 uptrend (now near
1.3250) triggering a long-term bearish
trend reversal and a break of pivotal
weekly support from the 1.3265 June
2007 low.
Andrew Chaveriat is a technical
analyst in the foreign exchange
department of BNP Paribas in New
York.
* Frost and Prechter: Elliott Wave Principle,
New York: New Classics Library, 1985 (5th
ed.), pp 30.
Market Views
INTEREST RATES
CHANGING PERCEPTIONS
by Ron William
I
nterest rate fever swept the market
after the yield on 10-year US government bonds registered its
biggest jump in years. The sharp rise
has now pushed above a twenty-year
downtrend and signaled a potential
long-term advance in rates. Technical
projections offer an initial target of
5.50%, followed by the psychological
6% level. Such a move could have global implications as other key government
yields also climb higher, fuelled by
strong economic data, and in places,
fear of inflation. Moreover, historical
trends in the supply of money and a
study of the relationship between commodity prices and interest rates provide
further evidence for a sustained period
of rising inflation.
Long-term trends
In 1981 US interest rates peaked near
16%, after an extended period of rising
inflation. Following this peak, interest
rates declined for just over twenty
years. These alternating long-term
trends, otherwise known as secular
moves, reflect generational economic
and social changes in society, (usually
lasting a minimum of two business
cycles). The low in 2003, at 3.10%, took
place in a deflationary bond buying
panic and marked the lowest level in
yields since the mid-1950s. It also
ended the secular decline. Since then,
rates have advanced and recently
pushed above the major trend-channel.
A sustained break above this area
would fuel a secular advance.
Money supply
When central banks grow the supply of
money faster than the general economy
Figure 1. US Long-term interest rates advance from overstretched half a century lows
and break above the major trend-channel. Source: Bloomberg L.P.
Figure 2. Historical trends in the supply of money, highlighting the most recent rise.
Source: Bloomberg L.P.
is growing (measured by GDP growth),
relatively more money chases fewer
goods and services, producing inflation. The chart below illustrates historical trends in the annual growth rate of
money supply - (measured using M3 the broadest definition of money). In
the 1970s the western world experienced double digit rates of inflation
and money growth, associated with a
July/August 2007
rising commodity bull-market. This was
then followed by a major decline as the
then Fed chairman Paul Volker
attempted to curb inflation by targeting
M3. The result was massive disinflation, with M3 rising, but at a slower
rate, relative to the growth of the
1970s. The 1990s was a period of deflation, registering a negative growth of
M3. Thereafter a new uptrend in →
11
Market Views
Figure 3 – CRB Commodity Index shares a unique relationship with US Long-term interest rates. However since 2001, the third greatest commodity bull-market in modern history has generated significant divergence with interest rates. (KEY: White – CRB Index,
Orange – US Long-term interest rates) Source: Bloomberg L.P. {Type HS <GO> to analyze the spread and correlation of two selected securities}.
M3 began to reflate financial assets and
the stock market boom, running
through to 2000. Recently, the trend
has altered direction and risen higher
into April of 2006, which is when the
Fed stopped reporting M3.
Rising commodity prices
Figure 3 illustrates the unique relationship between the CRB Commodity
Index and US interest rates up to 2001.
Since 2001, however - during the third
greatest commodity bull market in
modern history - the CRB Commodity
Index and US Interest Rates have
diverged significantly. This is almost
certainly unsustainable. Traditionally,
commodities are the basic source for
goods and services produced in the
economy and higher prices eventually
lead to a rise in the general cost of living.
The last time there was a strong negative correlation between the CRB and
10-year yields was more than two
decades ago in the early 1980s, when
these two markets shared a low correlation of -0.34. Interestingly, the only two
strong negative correlations occurred
around major long-term trend changes
on the CRB Index. Psychological
changes in perception could explain
12
why interest rates lagged commodity
prices during these two instances. A
twenty-year secular trend is such a large
fraction of an adult's professional life
that investors have a tendency to
believe that prices only ever move in
one direction. This rear view mirror
conditioning tends to be strongest at
major turning points, when the longterm view is being challenged.
The negative correlation in the early
1980s happened after the CRB Index
reached its all-time high of 335, following a decade-long advance in commodities. Most investors viewed this period
as the wave of the future and very few
believed that a change in trend was possible. Interest rates continued to rise
after the peak in commodity prices and
it was only after a 20% drop in the CRB
that inflationary fears started to recede
and interest rates finally began declining. Today we are faced with an even
larger psychological change in perception as most investors continue to
expect interest rates to hold around
overstretched half century lows,
despite the fact that commodity prices
almost doubled by early 2006.
Conclusion
Interest rates have broken above the
July/August 2007
long-term trend-channel, favouring an
advance to 5.50% and the psychological 6% level. It is worth remembering
the move originated from overstretched half century lows and still
maintains significant divergence from
commodity prices. Once a critical mass
of investors realize the economic
impact of a secular bull-trend in commodities and rising inflation, this psychological perception will be overcome.
The most recent phase has seen a six
year lag between the upturn in the commodity trend and interest rates. One
key reason for this is the disinflation
effect of low priced goods manufactured from emerging markets, notably
India and China. However, this globalization dividend and the cyclical impact
from excess capacity are now starting
to unwind and with central banks vigilant on inflation, interest rates will likely continue to rise.
Ron William is a Technical Analysis
Specialist at Bloomberg, LP.
The views and analysis presented here is not a recommendation to buy, sell or hold any security nor
are they to be relied upon for any investment decision. The views and analysis expressed here are
solely those of the author and do not neccesarily
reflect those of Bloomberg, L.P.
INTERMARKET ANALYSIS
Following the recent rise in longer end US bond yields, we bring together four leading market analysts to discuss the technical outlook for bonds
and the likely impact of higher yields on the global markets.
Sponsored by:
Chair: Matthew Clements
Editor,
The Technical Analyst
David Sneddon
Director in the Fixed Income division and
Technical Analyst for the global fixed income markets,
Credit Suisse
Clive Lambert
Director,
FuturesTechs
Tom Hobson
Chief Global Technical Analyst and
Head of EMEA Fixed Income Strategy,
Merrill Lynch
Max Knudsen
Director,
PIA-First
14
July/August 2007
“[BUND TRADERS] USE
CANDLESTICK AND MARKET
PROFILE CHARTS AND SO
PRODUCE VERY ‘WELL BEHAVED’
TECHNICAL SIGNALS”
- CLIVE LAMBERT
The 10 year US Treasury
Is the recent breakout of the 10 year US Treasury yield from its long
term downtrend the real thing or could it still be a false break?
Max Knudsen: If you look at the price action it's pretty convincing. The cash market is hovering around 5.25% and this
reflects the fact that real money knows yields are going higher. The breakout in the futures market is basically a continuation of what has been happening for the past four years.
There has been a sharp decline in bearish momentum during
the past two weeks but this has been prop traders picking up
cheap paper.
Tom Hobson: I think it's all a continuation of the price
action since 2003. We are in the middle of a corrective
process from very low yields that's accelerating. Therefore, I
do think it's a sustainable breakout and that we are now in a
longer trend towards higher yields. The era of low yields is
over. For me twos and fives have already reversed over a year
ago. It's the long end that is important now and directional
leadership in bonds is also about to switch back to the US.
What are the main price projections that can be made from this breakout and what are they based on?
David Sneddon: For the 10 year Treasury, there is a whole
cluster of key levels around 5.40%-5.50% and how these are
dealt with will be the next big test for the US bond markets.
These level projections are based upon retracements from
the yield base at 4.91%-4.40% and a weekly resistance line
going along the top of the highs of the last few years at
5.43%. I expect to see buyers trying to defend this area so the
outcome from this level will be a key signal. Also important
are old yields highs from 2001 and 2002 and Fibonacci
retracements. It could turn out that the rise in yields since
2003 is a continuation wedge and so the upward move in
yields will stall. However, my projection is for yields to go to
6.0% by the end of the year followed by the possibility of a
July/August 2007
Clive Lambert
15
long sideways move.
What we need to look out for in the fixed income markets
which will bring our consolidation phase to an abrupt end is
a rise in 10 year JGB yields about 2.05%. Beyond that they
could go explosively higher in the long end and there is no
support after that until 2.50%. A big sell off in Japan will
bring consolidation to an end in the US and Europe.
Tom Hobson: It depends on the time frame. My retracement target from the high is 7.94% and so it's difficult to
come up with a technical reason why yields shouldn't go to
7% in the next five years. I'm targeting 5.75% to 6.0% by the
end of the year.
The interest rate bubble that began in the early sixties and
topped in 1981 and bottomed in 2003 is a historical event of
huge magnitude whether you look at it from a technical or
fundamentals perspective. Does this mean however that we
are going to go back to yields above 10%? I think it's very
unlikely.
Do bonds remain a leading indicator of the stock market and how
should the dollar be reacting?
David Sneddon: Trying to find a stock market sell signal
using the bond market just isn't working at the moment. The
evidence for changing sentiment in the equity markets just
isn't there now. Also, it should be remembered that the 2 year
Treasury hasn't moved much and even long end yields are not
high enough to dent equities at the moment. The stock market has had plenty of opportunity to go down over the past
month but has been supported
The resilience of US stocks is extraordinary and I believe
there is plenty of way to go even taking into account higher
10 year yields. But the dollar is reacting although not in the
way that may have been expected. Dollar weakness can be
seen against the yen, euro and Swiss. Dollar/Swiss is interesting because we may be approaching levels below 1.20 that are
major supports and if these fail to hold then we will see
much more entrenched dollar weakness.
Max Knudsen: I look a lot at the dollar index future and the
last few weeks have seen further entrenchment of bearish
dollar sentiment, despite higher yields. Importantly, key support levels continue to be broken including the previous 2004
low of 80.48 following an evening star pattern formation in
June. The next major level is the 1995 low of 80.14. I'm on
calm alert at the moment because if 80.14 were breached
then we are looking at 78.95, the 1992 low. Following each
bounce in the index, the sell off has been progressively more
aggressive so the significance of 80.14 should not be underestimated. However, we may see some short term profit taking soon.
Clive Lambert: The gorilla in the room is equity markets.
While there's no sign of a turnaround at present, if things did
16
Tom Hobson
suddenly turn it would surely be swift and nasty and there
would be a flight to quality into bonds. This relationship isn't
as hard and fast as it used to be, but in these kinds of 'event'
situations it always comes to the fore again.
Tom Hobson: US stocks haven't had a 10% correction yet
which is what all the equity guys in the US are still worried
about as European stocks have already corrected last year.
On the currencies side of things, yen weakness and the
strength of European currencies is a bigger issue than the
dollar. The yen is going to get crushed in the months ahead
as it begins to capitulate against the European currencies and
Australian and New Zealand dollars. Whilst this might sound
like a fundamental view, it's actually technical in nature
because of the way central banks have responded.
July/August 2007
Intervention has taken place against channel resistance so as
we see the yen move to 177-180 verses the euro, the pressure
for intervention will become severe and this in turn will force
the Bank of Japan to raise rates.
What about the European bond markets? Are bund prices still following Treasuries?
Clive Lambert: If Bunds are not following the Treasury
market in the way they once used to, it is probably because
the ECB is now producing a clearer outlook on rates where
with the Fed, things remain very uncertain. Now many of the
big moves in the bund market happen in the morning before
T bonds open.
First of all can I point out my clients are all short term
traders so I very much concentrate on the price and don't
spend much time at all looking at yields and cross-market
comparisons. Bunds have been behaving very well lately from
a technical point of view. In December, again in March, and
once more in mid May we came out of a period of consolidation by breaking a short term uptrend line, then soon after
we saw the breaking the bottom of the Bollinger Bands on a
closing basis. On each occasion this signalled an extended
period of weakness. Bunds are in a consolidation phase at the
moment between 109.66 and 111.31. The next move I expect
is through the lower level.
“THE 10 YEAR TREASURY CASH
MARKET IS HOVERING AROUND
5.25% AND THIS REFLECTS THE
FACT THAT REAL MONEY KNOWS
YIELDS ARE GOING HIGHER.”
- MAX KNUDSEN
Max Knudson
Why have technical signals in the bund market been so reliable?
Clive Lambert: The great advantage of trading the bund
market is that the technical signals have been so clean recently. Up to 50% of trades done on Bund, Bobl and Schatz contracts along with euribor and short sterling are done by proprietary or 'local' traders who have a heavy reliance on technicals and trade very short term timeframes. They use candlestick and Market Profile charts and the effect is to produce
very 'well behaved' technical signals and patterns. Also in
Chicago, profile charts are very widely used. I guess this highlights how technically driven bond markets are!
David Sneddon and Clive Lambert are both board members at the
Society of Technical Analysts (STA).
David Sneddon
July/August 2007
17
Techniques
TECHNICAL ANALYSIS
BY NUMBERS
by David Linton
T
he ever increasing demands on
technical analysts and traders
are such that you need to assimilate more key TA data faster and in less
space on a screen. So you want to know
which instruments are doing what in
terms of your own technical criteria
without having to scroll through lots of
charts or run scans. With more and
more powerful computers and software
you can virtually put any value or
expression you want in a grid with a
'custom column'. You can do much of
this by writing complex formulas in
your Excel spreadsheets or utilising
functions within your market terminal
or charting system.
Ranking the RSI
Let's say you are looking at the curren-
Figure 1.
18
July/August 2007
cy majors and you want to know the
RSI position of each one without having to look at all the charts. Here we see
in Figure 1, that USD/GBP has the
highest RSI at 72.83 with Dollar Kiwi
also above 70. So if your trading strategy is to sell a move below 70 on the
RSI, you know these are the only two
you have to watch. Conversely, two of
the Scandis below 30 could provide
your next long trades. You may want
additional information such as some
confirmation of the state of trend. Just
add in the ADX value and you could
rank this column to see the instruments
trending from strongest to weakest. So
here we see Euro/Danish has the
strongest trend with Euro/Swiss having the weakest.
Having the numbers in tabular form
Techniques
Figure 2.
Figure 3.
on daily data is one thing, but if you are
looking at intraday charts where values
are changing quickly with minute bars,
tracking the TA numbers in a table like
this really counts. You could have values, a trading system or indicators
showing in multiple columns for weekly, daily and hourly for instance, giving
you an instant feel of the charts across
those time horizons without having to
look at all the charts.
Comparing stocks
In Figure 2 we see some stock market
indices with their Point and Figure targets for a 1% arithmetic box with additional columns for the stop level and
the risk reward ratio (RR). So ranking
by RR we see that Nasdaq has the highest upside vs. downside on this basis
and the FTSE100 the lowest.
Where this sort of analysis becomes
really valuable is for lists of large data
universes such as stocks. For instance
you might want to know the stocks that
have the best relative strength against
an index from a given date. In Figure 3
we see the Norm R/S column showing
BHP Billiton has the highest relative
strength (up 43.35%) from the chosen
date. This ability to cross compare is
one of the unseen advantages of normalising Relative Strength. Again you
can add in other values such as
Momentum to further support your
tabular view of the market.
Perhaps the biggest advantage of
viewing markets in tables is the ability
to have your own signals shown, thereby hiding the complexity of the formulae that got you there. For instance, you
can add actions such as BUY and SELL
and the number of days since the signal
was given as we see in the last two
columns of Figure 3. If you are producing quartile style tables or spreadsheets full of data such as pivot points,
watching them live in this way is really
the way forward.
While experienced technicians will be
able to visualise the charts from just
looking at the tables, there is still no
substitute for opening the chart and
seeing all the data you want in graphical
form. The real advantage of this tabular view of the world is that you know
which charts you want to open first.
→
David Linton CFTe MSTA is chief
executive of Updata.
July/August 2007
19
Techniques
20
July/August 2007
Techniques
TRADING TIME
by Shaun Downey
In an excerpt from his new book, Shaun Downey explains the importance of using time as part
of an effective trading strategy.
T
rading time - a double meaning
alluding to actually allocating
the time to trade and then
understanding the critical information
regarding where you are in time when a
trade is placed. This facet of time has
many definitions.
1. The timeframe of the chart that was used
and why?
2. How critical is the immediate price action
directly after the trade is placed?
3. How long is the trade expected to last?
4. At what point in time is the trade within
the trend or are we at the end of the trend?
5. How strong is the trend based on the time
it has existed?
6. What is the risk/reward in relation to
time?
Volatility throughout the day
When understanding variations of risk
throughout the day, there are many
potential problems. The extension of
trading hours and the ever lengthening
number of economic data events mean
that traditional technical analysis methods that measure momentum on a continual basis are facing increasing challenges as markets go through periods
of low ranges and a lack of direction,
followed by bursts of activity and short
term trends. Automated trading seems
to have moved into the very low timeframe, high frequency of trades model
to tackle this problem, but this is not an
option for the human trader. In the
same fashion that timeframes of charts
are often fixed, so are the variables
within the momentum-based indicators
that are used on charts. If a 10 period
moving average is placed on 30 minute
chart on Bunds at 11am the average is
likely to have flattened due to lack of
activity. This would be the same case on
the opening when it would have reflected the activity or lack of it in the
evening session of the day before.
However, come 4pm, the average could
display completely different behaviour
based on the number of statistics produced that afternoon. Therefore it is
very difficult to use momentum indicators in a predictive manner and so we
return to the inherent ability of the
good trader to ride the waves of →
These are all important questions but
in my experience of visiting thousands
of traders over the years, they are questions that are rarely asked and for a
large number they are never even considered. One of the first questions I
ever ask a trader when we first meet is;
what timeframe charts do you use? The
answer is always a variation on the
same theme. "Oh I use a 30 min, 60
min daily and weekly". Not one person
has ever said. "I use the timeframe
chart that is relative to my concepts of
risk, volatility and range"
For the great trader their success with
this somewhat random method is
proof enough of their inherent ability.
For the not so great trader this is a
recipe for disaster. Therefore obtaining
a true measure of expectation in any
one period of time is critical to improving the chances of success.
Figure 1.
July/August 2007
21
Techniques
volatility. If you accept the concepts of
continual fluctuation in range and the
occasional mutation of a market into a
different environment then the answer
must be to make that variable of the
average continually adjustable based
not only on the range of any particular
bar, but also the time of day that that
bar was created.
Volatility time averages and
bands
Volatility Time Averages treats each
individual bar only in connection to the
same bar the previous days.
The average range is computed over
a user defined range. Then the highest
and lowest value of range for that time
of day is computed over the last 1000
bars. The difference between the current ranges over n bars is recorded
against the highest range over the last
1000 bars and depending on the difference, an exponential moving average is
calculated. This average is given a user
defined minimum and maximum range
of average which defaults between a 3
and 21 period. The conclusion is that
if the range is narrow in relationship to
the history of that time of day then the
average slows but if range is large, the
average speeds up.
Figure 2.
22
July/August 2007
Removing the variable of the average
and replacing it with a variable that
looks at each specific time of day to
previous days enables a set of bands
that maintain their flexibility to market
changes. They are called Volatility Time
Bands. As soon as the bar opens, the
average range for that time of day is
computed and 1, 2 and 3 standard deviations are placed on either side of the
market. The use of the opening is critical in that it provides a predictive
framework as the values are fixed and
lead to the ability to analyze on a multitude of concepts.
One of the key criteria is being able
to understand what is the limit of range
within one aspect of time. Whilst 1
timeframe can be used in isolation,
extra power can be obtained with multiple timeframe confirmation. Figures 1
(Bunds) show a confluence of
extremes as the 30 minute chart has an
extreme 3rd deviation low at 109.90,
which is also the limit of range in the
15 minute chart and as low as the 5
minute chart. When this is used in
combination with true measure of support and resistance with Market Profile,
not only can day trading turning points
be found, but also major strategic turning points in trend. This is given even
more strength when the Kase Peak
Oscillator is showing an oversold scenario as seen in the 15 minute chart. At
such times for both the short term
trader and strategic players, risk can be
defined as low as 3 ticks on Bunds. This
is due to the connection between
macro picture supports and resistance
and micro picture limits of range. It
allows for far higher volume to be traded as position sizing and subsequently
risk reward ratios explode upwards. →
Techniques
market is in a strong trend. When 6
steps are in place we are in a mega
trend.
Figure 3.
Stochastic Steps
Once a trade has confirmed a major
turning point, the next major difficulty
is in knowing how to switch such a
micro timeframe trade into a position
that can be held if the trend then develops. This is one of the hardest skills in
trading and the development of what I
call Stochastic Steps logic attempts to
solve this problem. Past analysis shows
that there are some trends in stock
index's that began in a 15 minute chart
and are still valid 3 years later and for a
weekly chart, many thousands of
points later.
Stochastic Steps records each
crossover of the stochastic and states
whether it was confirming the continuation of the trend by doing so in a
higher or lower contract value than the
previous
crossover.
Therefore
Stochastic Steps will either step up or
down each time the stochastic crosses
depending on the comparison in price
to the last time the stochastic crossed.
Trend definition and divergence
However, closer examination of how
the Steps interact between the contract
value and the Slow Stochastic value
itself reveals how new concepts of
divergence can be built based on the
patterns and connections between
them. This remains beyond the scope
of this article but it is an important
consideration for those who want to
investigate the relationships between
the Stochastic Steps with that theory in
mind. This becomes clearer if two
more Step studies are created recording
the value of the stochastic itself when
they cross over.
Crucially, they also tell us what the
focus timeframe is when a trend begins
and if it develops, whether the focus
timeframe is moving higher. This
enables a trade that may have begun
with a short-term bias to become a
long-term trade. This is described
below.
Confirming the trend
Each time the market steps in the direction of the trend, the trend itself is
being confirmed. Once the relevant
indicator has stepped in the same direction 4 times consecutively - this is the
trending and focus timeframe - the
July/August 2007
The mega trend
If both Stochastic Steps are above 6
this indicates the strongest trend of all.
Figures 2 and 3 show the 15-minute
entering a mega trend. This is followed
by the 30 minute doing the same later
in the trend. This is an example of how
the focus timeframe can be moved up
and allow for trends to be ridden for
longer.
This is critical to trends developing as
they must move up timeframes in a
continuous fashion or the trend will
simply die. Most trends with low beginnings will often end long before the
focus timeframe moves up to a daily
chart. This is normally true of bond
and FX markets which rarely go
beyond a half day chart. Even so this
would entail a trend lasting for more
than 6 months in most cases. The real
power comes in stock markets where
trends can last years. The Dax rally
began in 2005 and Figure 2 shows how
it began with a mega trend in the 15
minute before moving up to the 30
minute. This trend moved up all the
subsequent timeframes and now is a
weekly trend in spite of the recent correction (Figure 3). Whilst these dips can
seem large, the fact that the trade had
such humble beginnings means that
risks can be wider. For those who
would want to maintain tighter risk,
they can use the many methods shown
in my book ‘Trading Time’ which look at
unique ways of qualifying swing patterns.
Shaun Downey is head of research
at CQG.
23
Techniques
CANDLESTICK
SIGNALS
THE J-HOOK PATTERN by Steve Bigalow
24
July/August 2007
Techniques
O
ne of the most powerful candlestick patterns is the J-hook.
A J-hook pattern is a variation
of a wave 1-2-3 price move and is an
easy pattern to identify with the use of
candlestick signals. A common challenge for traders is in knowing when to
sell after a price has made a strong up
move. For this, the J-hook pattern
demonstrates some easily identifiable
attributes. First, it starts with a strong
uptrend that usually produces stronger
than normal returns in a very short
period of time. This strong up move is
significant enough to create the normal
wave pattern, a reversal caused by profit taking followed by a declining trajectory of the pullback, then the continuation of the uptrend. The J-hook pattern is the description of the pullback
involving a rounding out of the pullback low followed by a move back up
forming a hook (Figure 1).
The pattern set-up and criteria
The first uptrend, which is usually a
powerful move, will show clear candlestick sell signals when the initial upmove comes to an end. The top will be
formed with the stochastics (or other
trend indicators) in the overbought
area. Because of the strong initial
uptrend, the first evidence of sell signals should be acknowledged. Even if
it is suspected that the uptrend could be
forming a J-hook pattern, why risk
remaining in the trade? When a sell signal becomes evident, take your profits.
What criterion makes a candlestick
trader suspect a J-hook pattern will
Figure 1.
form? Analysis of the market trends in
general will provide that information.
For example, if a stock price had a
strong run up while the market indexes
had a steady uptrend and did not
appear to be ready for a significant pullback, then a strong stock move could
warrant some profit taking before the
next move up. The benefit of being
able to identify candlestick signals is
being prepared for buy signals after a
few days of pullback. These signals
would also alter the trajectory of the
stochastics that will be pulling back.
Witnessing Doji, Hammers, Inverted
Hammers or Bullish Harami after a few
days of a pullback becomes an alert
that the selling is starting to wane. If
the stochastics are flattening out during
that same timeframe then a set-up for a
J-hook pattern is taking place. Taking
profits when the first sell signals occur
in the initial uptrend eliminates the
downside risk with the sell signals indicating that it is time to get out of the
trade. Even though the strength of the
initial move would warrant suspecting a
J-hook pattern to form, there is no
guarantee that the pullback could not
retrace 20%, 40%, 60% or even greater
of the initial move up.
If after four days small candlestick
buy signals start forming, there is nothing wrong with buying back into the
position. The second entry of this
trade now has some targets that can be
clearly defined: the first target should
be the test of the recent high. This can
induce taking quick profits and getting
back out of the trade. On the other
hand if strong signals are seen as the
recent high is breached, that would be a
clear indication the high was not going
to act as a resistance level. A new leg of
the trend may be in progress.
Examples
Figure 2 shows that after the uptrend,
the J-hook formed when the price
pulled back for a few days. However,
the stochastics never reached the oversold area and they came down only part
way before hooking back up. The signals indicated buying before it pulled
July/August 2007
back too much showing that buyers
were going to test the high of the previous week. The gap above the recent
high indicated that the buyers were very
anxious to see prices go higher.
Recognizing this pattern and the elements that form it allows traders to
move decisively at the right points of a
trend.
Where will the pullback move to?
Sometimes that is obvious, sometimes
it is not. Yet there are indicators that
can at least provide a target for a J-hook
pullback.
In Figure 3 the 8-day moving average
becomes the obvious support level.
Although the stochastics have not
moved back down to the oversold condition, it becomes apparent with the
Morning Star pattern that the potential
of a J-hook pattern is starting. Buy signals occurring at a major technical support level, (even though the stochastics
are only part way down toward the
oversold area), should be recognized
for their potential. Buying in the 47.50
area should be done with the anticipation that the price could reasonably test
the recent high in the 50 area. Once
again, the benefit of candlestick signals
is being able to determine whether the
50 level will become a resistance level
or not. A break through that area then
becomes the next evaluation.
J-hook and position taking
The benefit of candlestick signals is
that they reveal when a pullback is not
occurring with great enthusiasm.
Seeing minor candlestick buy signals a
few days after a pullback has occurred
at least provides the inkling that the
pullback may just be profit taking. As
the downward trajectory of the pullback starts flattening out, watch for
more buy signals. When the trend
starts moving up, a new position can be
established.
After a strong rally a profit-taking
period is to be expected but a full-scale
reversal may have occurred. A candlestick strategy should involve deciding
whether to short heavily the market
or be prepared to re-establish long →
25
Techniques
“THE J HOOK IS
A ROUNDING OUT
OF THE PULLBACK
LOW FOLLOWED
BY A MOVE
BACK UP
FORMING A
HOOK.”
Figure 2. Garmin Ltd.
positions. Once candlestick buy signals
start appearing in a market index chart,
giving the indication that a J-hook pattern could appear, these prepare traders
mentally to move one way or the other.
If short positions were established at
the first sell signals in the trend, being
prepared for the covering of those
positions can be better executed when a
J-hook pattern formation is anticipated.
Individual stock positions have the
additional benefit of the market trend
Stephen W. Bigalow is the author of
several books on candlestick charting and is the principal of the
www.candlestickforum.com, a leading website on candlestick trading.
Figure 3. Inverness Medical Tech. Inc.
26
itself in evaluating the potential J-hook
pattern. If during the market uptrend a
stock price has moved up with greater
magnitude than the market trend in
general, then that becomes the first
alert. Simply, the stock trend is very
strong. A pullback occurring in that
stock, when the market trend appears
to be continuing, also gives rise to
watching for a J-hook pattern to occur.
July/August 2007
Techniques
PORTFOLIO TESTING
by Thomas Dorsey
Techniques for evaluating the robustness of a stock portfolio.
P
reparation in portfolio management means designing a trading
plan and testing it. Without test
results over a long period of time that
take into consideration all types of
markets you've got nothing. You must
evaluate the program in up, down, sideways, large cap, small cap, mid cap,
equal weighted, cap weighted, growth,
and value markets. If the market can
throw it at you, you better have tested
for it. Without this work you have a 10
high poker hand at a table of card
sharks. It may be the winning hand, but
what are the odds and how would you
know?
With test results, you've got peace of
mind and proof that you're trading plan
has a high probability of working as
you expect. Test results also make it
much easier to execute your plan
because you aren't racked with doubt
on each trade - you know it works on
enough trades to make up for the bad
ones. A brief period of unfavourable
results does not rattle you because
you've got conviction that can't be
shaken. You've got the data to back
your trading plan and you have effectively taken out the biggest deterrent to
profits in investing; "emotion".
Testing at Dorsey, Wright &
Associates
We spent a lot of time on portfoliolevel testing when developing the
PowerShares DWA Technical Leaders
Index (PDP) - an Exchange Traded
Fund (ETF) that trades on the New
York Stock Exchange (NYSE) - and the
Arrow
DWA
Balanced
Fund
(DWAFX), an open-end mutual fund
that we developed. We also spent many
tireless hours testing our Systematic
Relative Strength accounts, which are
run exclusively in accounts that we
manage at Dorsey, Wright, &
Associates Money Management.
In the PowerShares DWA Technical
Leaders Index (PDP) we complete a
stock to stock comparison to pull out
the strongest stocks. I'll use an example
of a comparison between Home Depot
(HD) and Lowe (LOW) to give you an
idea of how we might accomplish this.
Here are two companies which are in
the same industry group, but yet have
provided very different returns to
investors over the years. Using a Point
& Figure relative strength chart, we can
discern when it is best to buy Home
Depot (HD) and when it is best to buy
Lowes. To create this relative strength
chart, the price of one stock is divided
into that of another, then multiplied by
100 and the resulting value is plotted on
a Point & Figure chart (see Figure 1). It
is really only once it is plotted on the
Point & Figure chart that these readings
come to life and provide us with meaningful guidance. In this example we will
use HD as the numerator, and LOW as
the denominator. When the chart is on
a buy signal we would expect the
numerator (HD in this case) to outperform. When the chart is on a sell signal,
we would expect the denominator to
outperform (LOW in this case). At →
Figure 1.
July/August 2007
27
Techniques
the bottom of the chart historic signal
dates are listed so that you can easily
use the performance function to see
how effective those signals have been.
In the case of HD versus LOW, from
July 16th 1996 to October 12th 2000,
HD was clearly the place to be as it was
up 217%, versus 149% for LOW. Since
October 2000 however, HD has been a
substantial laggard, returning only
14.37% while LOW is up an astounding
235%.
An investor who has been in one of
these stores has likely been in the other,
and despite encountering different
color schemes the results of each visit
were probably very similar. You walk in
looking for a box of nails and you walk
out with a $400 nail-gun and a truck full
of lumber. In this sense we are all the
same, but in investing things are not all
the same. Using the powerful Point &
Figure relative strength tool, we can
clearly show when the tide turned in
favor of an investment in LOW at the
expense of HD shares. This comparison is multiplied thousands of times
over to achieve a final index of 100
stocks that comprises the PowerShares
DWA Technical Leaders Index (PDP).
We begin our process with a universe
similar to that of the Russell 3000
Index components, and finish our
process with a portfolio of 100 stocks
that we know provide us with good
odds of outperforming the broader
market.
Fund of funds
In the Arrow / DWA Balanced Fund
(DWAFX) we take a similar approach
utilizing relative strength comparisons,
however this fund is unique in that it
was the first "fund of funds" constructed exclusively with ETFs. As this is a
true Global Balanced fund there are
essentially four pieces that make up the
fund; US market based and sector
ETFs, Fixed Income ETFs, Alternative
Asset ETFs, and international (NonU.S.) ETFs. The first step in managing
this fund is a relative strength comparison process that allows us to determine
the size of each "slice," and thus
28
whether we overweight alternative
assets (commodities, currencies, REITs,
etc.), or International equities. For
instance, we can perform a relative
strength comparison of bonds to US
equities to determine which has the
upper hand. This type of comparison
would have had you overweight bonds
in the portfolio during several critical
junctures over the past seven years, but
largely underweight outside of those
stretches. Bonds would have been
overweighted in the portfolio from
April 2000 to April 2001 and then again
from July 2001 to November 2001. The
relative strength chart would have
switched once again to suggest an overweighting in bonds in April 2002 to July
2002, when the iShares Lehman
Aggregate Bond Index (AGG) was up
2.45% while the S&P 500 Index (SPX)
was down 16.5%. Perhaps more
importantly, however, is that since
March 2003 this relative strength comparison has had the portfolio move
away from bonds (within the confines
of a minimum weighting of 25% in
bonds because this it is a "Balanced
Fund"), and into overweighted exposure within equities. Since March 2003,
bonds have been essentially flat while
the SPX is up 70%.
Relative strength
In the second step, we focus on relative
strength analysis within each slice, so
once we've determined that a larger
equity slice is recommended, the question becomes how to gain that equity
exposure. We take this step with the
comfort of knowing that we have done
our homework. Extensive testing using
comparisons such as that between HD
and LOW as well as that of comparing
major asset classes to one another, over
a long period of time was vital because
we discovered a lot of things we didn't
expect to find, namely:
The best portfolio performance
comes from buying the highest relative strength (RS) stocks. Those
stocks can be volatile, but the data
shows that is where the best returns
July/August 2007
are.
A portfolio of high RS stocks has
smaller drawdowns than the market.
We thought that because high RS
stocks are volatile then drawdowns
might be higher, but the data shows
they are not.
A portfolio of high RS stocks has a
low R-squared. Most fund portfolios
have an R-Squared around 0.9. We
didn't expect to find R-squares below
0.6!
A portfolio of high RS stocks has a
beta that is less than the market.
Because of the volatility, we thought
the beta might be high but it turns
out that because the portfolio can
often move opposite to the market,
the beta is under 1.0.
The long-term alpha on a portfolio of
high RS stocks is surprisingly high.
The data shows that very high alphas
can be generated because out-performance is so large and the beta is so
moderate.
Tax efficiency is good. We figured
that cutting losses and letting winners
run would be fairly efficient, but we
were still surprised when data proved
85% of profits were categorized as
long-term capital gains and thus qualified for preferential tax treatment.
Turnover is very manageable. We
thought a high RS portfolio might
generate a lot of transactions, but the
data shows turnover is not much different than the average equity mutual
fund.
It's only because of the data generated
by extensive testing that we discovered
these things. An old basketball coach
once told me, "Somewhere some kid is
practicing his shot. If you're not practicing, when you go head-to-head he's
going to beat you." The market is no
different. If you're making excuses for
not doing the testing, you're looking for
the easy way to win and there is no
short cut.
Thomas J. Dorsey is president of
Dorsey, Wright & Associates,
www.dorseywright.com.
Energy Futures and OTC Markets. One Screen. ICE.
Introducing: NYBOT® Soft Commodity Contracts on ICE.
Commodities traders and risk managers rely on ICE.
ICE Futures’ suite of energy contracts, coupled with ICE’s
robust OTC energy marketplace, offers tighter spreads
and better fills – while increasing market transparency
and efficiency. So find out for yourself why traders
worldwide rely on ICE.
To learn more, contact [email protected]
or call +1 (646) 733 5000 or +44 (0) 20 7265 3782.
Visit us at www.theice.com.
Trading on the NYBOT is governed by specific rules as set forth by the Exchange. Those rules are subject to change. Contact a licensed broker for additional
information including commissions, fees and margin requirements.
Interview
Aaron Brown
INTERVIEW
Aaron Brown is an executive director at Morgan
Stanley, where he works in risk methodology, modelling the distribution of trading P&L at the firm-wide
level. He is highly regarded as a quant, trader, academic and serious poker play, and holds degrees in
applied mathematics from Harvard and finance from
the University of Chicago. His career in the financial
markets spans over 20 years and includes positions
with Prudential Insurance, JPMorgan, Rabobank,
Citigroup and now Morgan Stanley.
In his book "The Poker Face of Wall Street", Aaron
Brown explores the historical and conceptual links
between gambling and modern finance, and explains
why success in both depends on the art of taking
"incalculable risks."
30
TA: You are a key figure in the movement referred to in
the financial markets as 'the rise of the geeks'. Do you
think traders and fund managers need to be cleverer or
better educated now than say 10 or 20 years ago, or does
the same kind of person still succeed regardless?
AB: The technical bar has certainly been raised. Twenty-five
years ago if you could sum a geometric series and use a
spreadsheet, you were a rocket scientist. Today kids come out
of school and apply for entry-level jobs with far more mathematical and technical education.
Still, the determinants of success have not changed so
much. Cleverness is more important than I.Q.; the person
who is good at solving math problems, and enjoys them, will
do better than the person who can prove deep theorems.
Confidence, ambition, focus, pride, people skills and honesty
are still the main ingredients. Stochastic calculus is a nice
extra.
TA: Is the efficient markets hypothesis now discredited?
AB: A hypothesis is something you assume for the sake of
the argument, to see where it leads. If you don't assume efficient markets, you can explain any price. "He bought it for
that price because he thought it was worth more, she sold it
for that price because she thought it was worth less." If you
can explain everything, you explain nothing.
When academics started assuming efficient markets fifty
years ago, they had no idea how close to true the assumption
would be. Everyone was shocked at how efficient markets
July/August 2007
Interview
“IF YOU WAIT FOR EVERY RISK OF A
STRATEGY TO BE CALCULATED, IT WILL
HAVE BEEN COMPETED AWAY LONG
BEFORE YOU GET IN THE GAME.”
are: how few professionals made consistent money, how hard
it was to predict future price movements. What started out as
a simplifying assumption, like ignoring air resistance, turned
out to be almost true.
Yes, some persistent anomalies have turned up. But no one
ever found an anomaly without starting from the efficient
market hypothesis. Without it, there are no anomalies, everything is consistent with theory.
The efficient markets hypothesis was always meant to be
the beginning of inquiry, not the end. It has not been discredited for that purpose; in fact, no one has come up with a
credible alternative. The only surprise is that by starting with
efficient markets, you were ahead of 99% of the professionals who had studied finance for years.
TA: What do you think about the research coming from
behavioural finance? Have you found ways to quantify
any behavioural biases into profitable trading strategies?
AB: I think there is a lot of interesting work being done by
some behavioural finance researchers, but also a lot of nonsense. I am not personally interested in why people do things
or in finding predictable "biases" or irrationalities.
When someone does something with a predictable result,
I'm willing to assume they want that result. So I regard
behavioural anomalies as evidence that people want something different from classical economic assumption rather
than that people are bad at making decisions. I think behavioural results are a challenge to utility theory, not to financial
theory.
With regards to trading, my trading is heavily quantitative.
I consider the behaviour of market participants, but as institutional entities rather than psychological beings. For example, I think about how much demand there will be for a certain stock at various times and prices in order to cover
options positions, but I don't worry about what the holders
of those option positions are thinking.
In poker, I definitely think about what other players are
thinking.
TA: Do you think a trader has to be able to take "incalculable risks" to succeed?
AB: I think taking incalculable risks is the essence of real
trading. Certainly you can call yourself a "trader" and make a
living as an order-taker or by taking spreads; this has no more
incalculable risk than many other jobs. But if you wait for
July/August 2007
every risk of a strategy to be calculated, it will have been
competed away long before you get in the game.
TA: Can you provide an example of a trade/investment
whereby you exploited or entered into 'incalculable
risk'?
AB: Well, you almost always enter into incalculable risk.
Certainly when you trade something for the first time, there's
always the risk that you failed to understand some essential of
the market. When you throw the switch on a program trading system, you're never 100% sure it won't generate disastrous nonsense trades. When you hire someone, or start a
business or quit a job; there are aspects you cannot calculate.
And, in a sense, when you enter into incalculable risk you're
exploiting it, because it keeps a lot of the competition away.
One example of pure exploitation of incalculable risk was
my company eRaider.com. I started it in 1998. We were a
public mutual fund (Allied Owners Action Fund Inc.) that
bought 5% stakes in public companies, then used the website
(eRaider.com) to organize all company shareholders to force
positive change. I think in normal times we never would have
got approvals to open the fund; it posed issues to dozens of
securities regulations. If we did get it open, it would have
been buried with the techniques companies use for other dissident shareholders.
But in those days, no one knew how the Internet would
affect financial regulation and equity trading, anything was
possible. eRaider.com could be part of the problem or it
could be part of the solution. We got meetings with the SEC
commissioners, we announced our targets live on CNNfn
from the floor of the New York Stock Exchange, we got
instant media and company attention despite being a tiny
fund with a nutty idea. We were influential in the Council of
Institutional Investors, the Financial Accounting Standards
Board, the National Association of Securities Dealers and
the New York Stock Exchange. When the future seems highly uncertain, anyone willing to stake a bold claim gets attention.
TA: You say in your book that passive poker - most often
seen when players tend to 'call' rather than 'raise' or
'fold' - is not a winning strategy. Is passive investing
similarly a bad idea?
AB: There are lots of passive players in the market.
Obviously you have the index funds, but lots of funds that
claim to be actively-managed are really taking what the market gives them. Many participants want to execute at volumeweighted average price rather than take a chance on trading.
The difference between poker and the markets is there is no
point to passive poker, but passive investing and passive capital raising works pretty well.
TA: What kinds of trading strategies do you favour? →
31
Interview
AB: I'm a quant. I look for patterns in historical data that I
can exploit. I don't think I have better information than other
people or faster execution, I think I'm smarter about how to
process the information. I like to remove as much noise as
possible, so I tend to implement things in a market neutral
way, and diversify enough to make it a risk-arbitrage
approach. I also like to be liquidity neutral, I've never liked
pure momentum or pure convergence trading; I don't like
carry trades.
TA: If you were to set-up a new fund, what market(s)
would you focus on and what strategies would you
employ?
AB: I think the revolution started by
credit derivatives has a long way to run
yet, and there will be major opportunities in that area. One idea I toyed with is
to start a credit-crunch binary payoff
fund: the fund will be invested to stay
flat until the next major credit crunch,
then generate exceptional returns. I
think the fund-of-funds business has
been unimaginative; there is a lot of
room for new ideas there.
In a badly-crafted strategy, you can't get your execution and
missing it is fatal. I have never tried to make money with pure
technical analysis, but I think ignoring the technical factors is
courting disaster.
TA: Do old adages like "let your profits run and cut your
losses short" still have any bearing on trading success?
AB: There are good strategies that involve all four permutations of fast or slow profits and fast or slow losses. The key
is to have a strategy; the best way to lose money is to make
each decision as it comes up.
Still, most people have biases to take
profits fast and losses never. So the
adage is good advice relative to instincts.
TA: What principles do you employ
with regard to risk management?
AB: Once again, I'm a quant. I think you
can calculate these things. I design strategies very carefully to have precise return
distributions. I don't have much faith in
diversification beyond seven or eight. I
have no faith in big covariance matrices.
So I look for four or five factors with
low dependence and build on them.
TA: Does any market offer better
arbitrage / price anomaly opportunities than others?
AB: Not really. With efficient markets
like major currency FX and major equities you can trade faster and bettermatched, so you can generate your own
opportunities. With less efficient markets, the opportunities
are there naturally. Traders always push markets to the point
where there are opportunities.
TA: Is high frequency trading seriously reducing arbitrage opportunities?
AB: No, high frequency trading is eroding slightly less high
frequency trading opportunities. It should make long-term
arbitrage and risk arbitrage opportunities better.
TA: Do you use technical analysis?
AB: I do not use standard technical tools, but I do believe
that understanding how short-term supply and demand, and
potential supply and demand, work through the markets is
essential for trading.
Thinking through the technicals also makes execution
pleasant. In a well-crafted strategy, you find the market coming to you. You get good execution for what you want to do
and if you miss your execution, the market will come back.
July/August 2007
TA: To what extent do you think the
rise of automated trading and algorithmic execution strategies affect
technical analysis? E.g. will chart
patterns fail more often?
AB: In theory, automated trading should drive patterns out
of the market. In practice, it seems more like the opposite. So
it's still an open question.
TA: Looking ahead, in your book, you say "while our
financial models have become very good at pricing
securities, they require assumptions that clearly conflict
with how security prices actually move….and when [the
conflict] is solved it will reveal hidden worlds of opportunity". Do you see signs of this issue being addressed
and from which area of finance and/or field of study is
progress being made?
AB: No, I don't see any progress. Times are too good; people are making too much money. It will take a disaster before
people take this seriously again.
Aaron Brown's "The Poker Face of Wall Street" is published by John Wiley & Sons Inc. and will be available in
paperback this August.
33
Software
ALGORITHM
BACKTESTING
by Dr John Bates
The term 'backtesting' has been created to describe the process of using past market data as a tool to ascertain how a prospective trading strategy would perform
under various circumstances, before that strategy is deployed live in the market.
Backtesting algorithms can help to ensure that financial institutions are prepared,
as there are a number of requirements, challenges and approaches that should be
carefully considered.
Background: Why Algorithmic
Trading Needs Backtesting
Algorithmic trading is one of the most
discussed topics in capital markets.
Initially the term was used to describe
the automation of equity execution, i.e.
dividing large block trades into slices,
using some statistical measure, in order
to minimise market impact and achieve
a benchmarked price.
However, in the last few years the definition has expanded to include highfrequency trading, i.e. analysing market
data in real-time against statistical models in order to detect trading opportunities - and then executing those opportunities. An example of a high frequency algorithm is pairs trading, which
monitors correlated instrument pairs,
looking for aberrations in the correlated relationship that imply the ability to
buy one and sell the other at a profit
before they return to correlation.
Algorithmic trading has also spread
into asset classes beyond equities, such
as futures and options, fixed income
and foreign exchange. In each asset
class, the same basic principles of monitoring market data for trading opportunities and then automatically executing on the opportunities still apply.
However, each asset class differs in the
34
specific algorithms that are appropriate.
As algorithmic trading has developed,
several imperatives have emerged. The
first is that you have to identify opportunities first - before your competitors and build your own custom algorithms
to capitalize on opportunities. In many
circumstances, you can't rely on prebuilt algorithms purchased or leased
from vendors or brokers because if
everyone has access to the same algorithms, there is a reduced competitive
advantage. The second imperative is to
gain first-mover advantage by building
and deploying algorithms quickly,
ahead of the competition. Only those
that can quickly deploy are likely to harvest the benefits.
The markets are continually changing
and an algorithm that was highly profitable yesterday may not be profitable
today. So the third imperative is to continually evaluate the effectiveness of
existing algorithms and, if necessary,
evolve or decommission them.
Algorithmic trading is both fast moving and highly automated and, as a
result, is often associated with
increased risk. Many people fear that
should things go wrong, they will go
wrong so quickly that traders will be
unaware and unable to intervene in
July/August 2007
time to prevent damage. This leads to
concern that the damage may not be
constrained and that certain circumstances may cause the entire market to
spiral out of control. It is therefore of
paramount importance to ensure that
an algorithm has been tested under a
wide variety of circumstances and in as
many trading situations as possible to
mitigate such risks.
Backtesting techniques provide a way
of evaluating and tuning algorithms for
profitability and testing algorithms
under various circumstances to ensure
they perform as expected in exceptional, as well as normal, trading conditions.
Backtesting Principles,
Requirements and Issues
Backtesting uses historical data
sequences in order to simulate how an
algorithm would have performed if it
was trading at a particular point in time.
The theory is that by testing it under
normal and extreme conditions, performance can be ascertained and sensible responses to exceptional circumstances assured.
The most extreme example would be
to test an algorithm with data from
both a bull market year and a bear market year. Another example might be
Software
Sample Backtesting Configuration
Figure 1. Synchronised data capture in the production and test environments: In the production environment data is sent and received via adapters to various market data feeds
and trading venues. This data enters the trading engine and also the tick database for
storage. In the testing environment the engine is fed with historic sequences from the
tick database and sends trades to a market simulator.
testing a forex algorithm with data
from every non-farm payroll day in
2007. The sequences of data selected
may be hours, days, weeks, months or
years, depending on the requirements
of the particular algorithm and backtesting scenario.
Identifying and Backtesting
Trading Patterns
In order to create a new algorithm, historical data will most likely be used to
research and identify the patterns that
an algorithm can use to make money.
For example, if a strategist using historical research identifies a particular pattern (e.g. when the moving average of
instrument X exceeds the price by
threshold Y, the price will always rise by
Z), an algorithm can be created to capitalise upon this opportunity.
Historical data, in which the pattern
was originally spotted, can then be used
to backtest and evaluate whether the
prospective algorithm would identify all
relevant opportunities and take advantage of them.
Acquiring and Managing Years
of Data
In order to backtest algorithms, a store
of relevant historic data must be kept.
In equities, for example, this could
mean tick and quote data, with full
depth of book, going back ten years.
Where an institution is trading across
borders, there may be a requirement to
store data from exchanges in the US,
UK, Canada, Mexico and Japan.
Depending on the detail of data that is
stored, there could be several thousand
market events per second on an individual exchange - equating to millions
per day, and even billions per year. In
data storage terms this equates to several terabytes per year.
This data has to be captured, stored,
indexed and queried to support efficient backtesting. One way of acquiring
the data is to buy it from trading venues
or data vendors. However, such data is
often delivered on CDs after-the-fact.
In-house data capture is required to test
with today's and yesterday's data.
Market data has a temporal dimension
July/August 2007
in which the sequence of data is material to understanding what happened.
This temporal dimension must be
stored and indexed as a first class property, so it can be replayed in order.
High Performance Backtesting
Dealing with data volume is not the
only issue in backtesting. In the quest
for ever-faster deployment of new
algorithms, there is a strong desire to
backtest algorithms quickly. Traditional
databases are not fast enough to capture or replay - in real-time - the volume of events needed to support trading usage and are not designed to handle time-series (temporally ordered)
data.
A new breed of tick database has
evolved to fulfil these requirements.
These databases support the temporal
ordering of events and can replay market data in the same order as the events
originally happened. Such databases
can also handle storage and replay of
thousands of events a second. In a high
performance backtesting framework, it
may be possible to run many thousands
of algorithmic permutations against
historic sequences at the same time.
This enables thousands of possibilities
to be evaluated concurrently and hugely accelerates time-to-market for algorithms.
Backtesting Multiple Asset
Classes
Algorithms for other asset classes may
introduce additional requirements to
the backtesting operation. For example,
foreign exchange traders may want to
backtest using data from multiple trading venues e.g. EBS, Reuters, Currenex,
Hotspot and a bank's own liquidity
pools. In futures and options it might
be CBOT, CME, Eurex and Liffe, while
in fixed income it might be Brokertec
and eSpeed.
In addition, some algorithms are
cross-asset in nature, trading multiple
asset classes in the same strategy. As a
result, multiple asset-class streams must
be replayed in order to backtest and it
may be necessary to replay a sequence
composed of several independent →
35
Software
Sample Backtesting Interface
deltas. Deltas describe cumulative
changes to the historic data. For example, if our algorithm hits a bid or offer,
although in the past it was still there,
the simulator should remove it from
the order book and remember this as a
delta to be applied henceforth to history.
In other circumstances, traders may
want to simplify things and assume perfect liquidity in order to test certain
rules in an algorithm. Alternatively, they
may want to use a totally simulated
world that uses synthetic market data
rather than historic market data.
Figure 2. Backtesting Control Panels: On the left screen a historical sequence and a strategy to stream it through has been selected for backtesting. On the right screen a wizard
guides the user through control options including speed of playback, pause and step
controls.
Strategy Tuning
A key aspect to backtesting is strategy
tuning; running strategies in various
different configurations, with the same
data, to see which permutation is the
most profitable. Data on each algorithmic permutation can be collected and
compared with the most profitable
configuration to be used for live trading. The effectiveness of a particular
permutation may change over time and
thus regular tuning is required to ensure
the algorithm is being run in its optimal
configuration.
asset-class specific sequences, retaining
the temporal ordering and relationship
of the different market streams.
Backtesting with News
A new backtesting requirement that has
emerged over the last year is the need
to support algorithms that trade on
news. Certain news moves the market,
particularly when it is economic news
or shock news (e.g. an unexpected war
or hurricane). Traders realise that if
they can respond before their competitors they can gain an advantage and are
using algorithms to monitor news
along with other market data.
As an example, Dow Jones has made
this process more quantitative by creating elementized news feeds, which use
tags to identify specific elements within
news events. In order to backtest algorithms that incorporate news, news
feed events have to either be recorded
or simulated. Since elementized news
feeds are a new phenomenon, it has not
been possible to acquire years of elementized news, but as algorithms that
trade on news become more commonplace more archives of unstructured
and elementized news will become
available for backtesting.
36
Simulating Market Impact
The most complex requirement in a
backtesting framework is the challenge
of simulating market impact. Replaying
historical data and feeding it into an
algorithm is one thing. But what about
simulating the circumstance where an
algorithm wants to take advantage of
opportunities in the market? Where
does it place its orders? In a comprehensive backtesting framework a market simulator is required. Algorithms
can route orders to the market simulator, which will respond as an external
market would (such as an equities
exchange, futures exchange, forex
venue or bond venue).
The complexity of simulating market
impact begins when you consider that
when backtesting with data from the
past, your algorithm wasn't actually
there, and thus its actions will not have
an impact on the historical data. If you
hit a bid in the historic data, by default
that bid will still be there. This issue can
never be fully addressed without creating a time machine that could go back
to the day in question and run the algorithm live. However, there are certain
techniques that enable more realistic
impact simulation. One such technique
is to use a simulator that can remember
July/August 2007
Conclusion
To better ensure that algorithms are
ready for any eventuality and will actually work as expected requires high performance time-series capture and
replay, large data storage, realistic market simulation and continuous algorithm tuning. The latest backtesting
approaches, including high performance tick databases for capture and
replay of time-series data can help
ensure there are no surprises. Those
that use backtesting appropriately will
be well prepared to earn their 'Boy
Scout' algorithmic trading badge.
John Bates is Founder and Vice
President,
Apama
Products,
Progress Software
Books
TRADING TIME
New Methods in Technical Analysis
S
Published by Oasis Research Ltd
246 pages
ISBN: 0955466806
£60
Trading Time is available to order
from www.trading-time.com
haun Downey is well known on City trading floors as a technical analyst at
CQG and for writing regular FX research for electronic brokers EBS. His
long awaited first book, "Trading Time" brings together his own technical
trading ideas using CQG indicators, many of which he has developed himself.
As Downey makes clear in the preface, the book is really about how best to time
your entry and exit and ensure that you trade in the 'correct' time frame. Whilst
there is more of a focus on day trading, he also emphasises the importance of
looking at longer term charts as an essential guide to getting the big picture.
The first chapter covers the author's introduction on trading time and highlights
the importance of using indicators correctly depending on the time period being
traded. For example, the traditional moving average treats each time period equally
despite some periods having much greater activity than others. This problem can
be cured by using a Volatility Time Average which adjusts to allow for changing
range and volatility over time. Downey has also created Volatility Time Bands,
which like Bollinger Bands, place standard deviations around the average and
which provide a better picture of market expectation and risk.
An entire chapter is also devoted to Market Profile, an indicator that emerged
from the trading pits at the CBOT that combines price and volume data and
remains an extremely effective but underutilised technique.
As the book repeatedly highlights, volume remains a neglected source of market
information for the trader. Afternoon trading volume is often a pre-curser to price
action the following morning, so for short-term traders Market Profile has a lot to
offer; but it can also be used for longer time periods. Downey provides many practical ideas for using Market Profile effectively because, like many TA techniques,
Market Profile is often misused. He corrects many misconceptions regarding anticipating support and resistance levels, interpreting economic data releases and identifying short-term trends.
Throughout the book Downey also makes the point about how market activity
varies throughout the day for different asset classes and how prices react differently to economic data releases. For example, bond prices are especially volatile on
days when inflation figures are released. Volatility Time Bands are therefore especially useful in capturing intra-day trend reversals. Downey also draws attention to
the proliferation of trading arcades and their impact on volume data. Because very
few arcades allow positions to be held overnight, greater attention should be
placed on late afternoon and early morning activity with regards to timing market
entries and exits.
The book contains countless examples (CQG screenshots) that are well annotated in explaining the various techniques. However, this does mean the book is really
best suited to CQG users as many of the indicators included are only available on
that platform. However, for traders using different systems, the various techniques
should be programmable (with the exception of Bloomberg).
Trading Time is a unique publication within the realm of technical analysis.
Many new books on TA are at once too simplistic and theoretical for the user to
apply its ideas effectively in day to day trading but this is a practical trading book
written by someone with a proper understanding of the global markets.
Recommended.
July/August 2007
37
Research Update
Trading Changes to the
S&P Index
When a stock moves into the S&P 500
index, demand from index fund managers will likely force the stock price to
rocket up, but how long are the effects of
the inclusion felt and for how long do
profitable
opportunities
exist?
CONFIDENT FORECASTING
Just how good are you at predicting
trends? Three researchers from the
University of Mannheim have tested
the trend recognition and forecasting
ability of two groups: i) financial professionals who work in the trading
room of a large bank and ii) novices
(i.e. students), based on the probability
they attached to a trend and the confidence they had in their own forecasts.
They found evidence of simultaneous
overconfidence and underconfidence.
Subjects tended to underestimate the
mathematically correct probability
(indicating underconfidence?), but
assumed confidence levels that were
too narrow or, in other words, underestimated the variance. They found no
evidence to suggest professionals were
less prone to these biases than novices.
Glaser, Markus, Langer, Thomas and Weber,
Martin, "On the Trend Recognition and
Forecasting Ability of Professional Traders"
(June 12, 2007).
POOR REWARDS FOR EXTRA RISK
Equity investors are overpaying for risky
stocks. This is the conclusion of David
Blitz of Robeco Asset Management and
Pim van Vliet of Erasmus University
Rotterdam, who have found further evidence that stocks with low volatility earn
higher risk-adjusted returns than high
volatility stocks. In order to exploit the
volatility effect in practice the authors
argue that investors should include low
risk stocks as a separate asset class in the
strategic asset allocation phase of their
investment process.
Blitz, David and van Vliet, Pim, "The
Volatility Effect: Lower Risk without Lower
Return" (April 2007).
38
Researchers from the University of
Reading have examined the impact of a
compositional change in the S&P 500
index on the stocks newly included in the
index. Perhaps not surprisingly, they
found evidence of a significant overnight
price change that diminishes the profits
available to speculators. They point out,
however, that there is still profit available
from the first day after announcement
until a few days after the actual event.
They also find evidence of consistent
trading patterns during trading hours
over the inclusion event.
Kappou, Konstantina, Brooks, Chris and Ward
, Charles W.R., "The S&P 500 Index Effect
in Continuous Time: Evidence from Overnight,
Intraday and Tick-By-Tick Stock Price
Performance" (May 2007).
Driven to Distraction
Psychological evidence indicates that it is
hard to process multiple stimuli and perform multiple tasks at the same time.
Three US-based researchers have tested
the investor distraction hypothesis, which
holds that the arrival of extraneous news
causes trading and market prices to react
sluggishly to relevant news about a firm.
Their test focuses on the competition for
investor attention between a firm's earnings announcements and the earnings
announcements of other firms. The
authors find that the immediate stock
price and volume reaction to a firm's
earnings surprise is weaker, and postearnings announcement drift is stronger,
when a greater number of earnings
announcements by other firms are made
on the same day. Distracting news has a
stronger effect on firms that receive positive than negative earnings surprises.
Industry-unrelated news has a stronger
distracting effect than related news. As
such, a trading strategy that exploits postearnings announcement drift is unprofitable for announcements made on days
with little competing news.
Hirshleifer, David A., Lim, Sonya S. and
Teoh, Siew Hong, "Driven to Distraction:
Extraneous Events and Underreaction to
Earnings News" (April 16, 2007).
YIELD CURVE REACTIONS TO ECONOMIC NEWS
How do US interest rates react to macroeconomic announcements? Two French
researchers have investigated the shape
of the term structure reaction of swap
rates to announcements. The results yield
several stylized facts about the bond
market, including the observation of at
least four types of patterns in the term
structure reaction. The first type seems
to be the better known hump-shape and
is likely driven my monetary policy; a second type affects mainly the short term
rate positively; a third type affects nega-
tively maturities between 2 and 7 years;
and a fourth one negatively affects maturities between 6 and 9 years. They also
found that the existence of some outliers
in the one-day changes in interest rates
usually leads to a strong underestimation
of the reaction of interest rates to
announcements.
Guegan, Dominique and Ielpo, Florian,
"Further Evidence on the Impact of Economic
News on Interest Rates" (June 1, 2007).
All papers are available from the Social Science Research
Network, SSRN, www.ssrn.com
July/August 2007
CONTENTS:
PAGE 41
PROGRAMMING AND
INTEROPERABILITY
PAGE 44
STRATEGY SPOTLIGHT
July/August 2007
39
PROGRAMMING AND INTEROPERABILITY
Benjamin Van Vliet consults extensively on
building automated trading systems with professional fund managers and traders and is the
author of "Building Automated Trading
Systems" (Academic Press, 2007).
He is the associate director of the MSc in
Financial Markets at the Illinois Institute of
Technology's Stuart Graduate School of
Business, and is responsible for their courses on
computer programming for automated trading.
Van Vliet is also vice chairman of the Institute
for Market Technology, where he sits on the
advisory board for their Certified Trading
System Developer (CTSD) program.
TA: Why is Visual C++.NET your
language of choice in your book on
building automated trading systems?
BVV: There are all kinds of trade offs
when it comes to technologies and programming languages. The UNIX system is widely considered to be the best
platform for implementing automated
systems. However, it's extremely expensive and it takes a long time to develop.
The great thing about Microsoft is that
you can literally develop a system in an
afternoon on Microsoft Visual.NET as
opposed to building it from the ground
up. Visual C++.NET in particular is a
very fast programming execution environment and you can take advantage of
the MS tools that you need for development, so it's an excellent way to go.
Programming language and operating
system, however, are not the only decisions to be made, and may be not even
the most important ones. You have to
consider many other aspects of technology, including network architecture,
who is your ISP, do you have a direct
connection to the exchange, how far
will your server be from that exchange
and so on. If you have a trading system
where every microsecond is of the
essence and the success of the system
depends on speed of execution then
you'll be much more interested in
spending hundreds of thousands or
even millions of dollars in creating the
fastest infrastructure you possibly can.
Whereas if you're going to hold on to
your position for 30/40 minutes or 3 or
4 days, then every micro second isn't as
important and maybe what you're more
concerned with is speed of development so you can get something up-and-
July/August 2007
running in a couple of weeks instead of
a couple of months.
TA: Sticking to language for a
moment, as this might be viewed by
many as a barrier to automation,
would you advise those starting out
in automated trading to learn
Excel/VBA first?
BVV: There are a lot of traders who
use Excel and VBA. We're trying to
move away from Excel based systems
to be able to do more robust calculations in a more robust client server
application environment. C++ is best
for interoperability with other hardware
systems but it would be difficult to
move directly to any of the C++ or C#
languages without any programming
experience in VBA or VB.NET. I
would therefore probably recom- →
41
mend learning SQL databases and
starting out in VB.NET. There are
enough similarities to VBA that you
will be able to understand Excel/VBA,
but it will also allow you to deal more
efficiently with time series data than
Excel/VBA, which is an important part
of any automated trading system.
TA: Would you recommend that a
trader or fund manager hires a programmer or financial engineer to
program their strategy, or is it something they can learn to do themselves?
BVV: Learning to program is not easy
and it can take years to become proficient in it. Very often a trader or firm
will try and gain enough understanding
to a level where they can speak the language and understand the technological
issues involved, but then partner with a
programmer who can take their ideas
and implement them in a programming
language. One of the points of the
Certified Trading System Developer
program is to say if you're going to be
involved in a project to build an automated trading system, you need to
understand enough about technology
and the language of programmers so
that you can communicate effectively.
Each of the three functional areas trading, technology, maths/quants - has
their own language and skill set, and in
order to work well together we have to
learn enough about the other two functional areas to make the process work.
TA: In what circumstances would
you recommend using off-the-shelf
systems like those available from
Patsystems,
TradeStation
or
Trading Technologies?
BVV: Really the only proprietary thing
about an automated trading system is
the trade selection and position management logic. All of the other processes could be substituted with commercial off-the-shelf software. For example, I can connect to the exchange
myself but that's very time consuming
and very expensive. However, I can
42
“IN THE GAMES WHERE PROBABILITY OF
SUCCESS IS GREATEST BUT TECHNOLOGICAL
ADVANTAGE IS THE KEY, THE SMALLER PLAYER
IS GOING TO BE PUSHED OUT.”
probably license third party software
like Trading Technologies and I can run
the execution through their API. There
are many other external pieces (such as
quant libraries, accounting systems etc)
that I may or may not be building inhouse or licensing as third party software. Nevertheless, if I'm going to
automate my trading process my trade
selection package then becomes a kind
of middleware where I'm connecting to
a real-time data feed, I'm sending
orders down an API, and I'm pulling in
historical data from another source. So
it becomes a big problem of getting all
of these various packages and technologies to work together, especially
for smaller firms who don't have the
resources to spend millions of dollars
and years developing entire platforms
from the ground up. One of the simplest solutions is to licence from a company like TT.
As I see it, there's always an evolution.
Let's just say I'm a private trader who
wants to build an automated trading
system. Well, given the constraints of
time and expertise I can license much
of the technology, maybe TradeStation,
and start doing it in my basement and
hooking up TradeStation using their
EasyLanguage. An institutional trader
is probably not going to be using
TradeStation execution. They may be
using it for charting, but they're probably not using EasyLanguage. But nevertheless, it's a good way to start. And
let's just say you start developing systems that make money. The next step is
to probably move up to a more professionally-focused execution package, like
a Trading Technologies. Maybe then
you read a book like mine which actually describes what they call the X_Trader
API for real-time data feeds and execution, and you start programming your
July/August 2007
own trade selection algorithm and you
start learning about the kinds of issues
that are in my book, like multithreading, interoperability etc. Let's say you
start to hire people - programmers,
mathematicians, more traders - and
your firm is growing. At some point
down the road, you may even then
choose to leave TT behind and start to
develop all of that software in-house
because as a more mature firm you really prefer to have the control to optimise
your technology and customise it for
your own trading decisions. Any piece
of software that tries to be everything
to all its customers is not going to be as
fast as possible doing the one thing you
want it to do. So the larger firms get,
the more they develop their own systems in-house.
TA: In your book you talk about the
KV methodology for discovering
new trading systems. What is this
and why is it important?
BVV: Discovering trading opportunities and implementing them as quickly
as possible is going to be an important
issue for small hedge funds and larger
institutions going forward. Let's say
Company A buys out Company B.
Between the time the offer is made and
the deal is consummated there is going
to be a relationship between the two
stocks. That relationship may hold for
only a couple of months but if you can
get your system up and running fast
then you're going to have much more
opportunity to take advantage of that
stat arb situation than someone who
takes a long time to set up. It's a short
term opportunity and the quicker you
can jump on it the better.
One of the things we've tried to
design is a methodology that works
“ANY PIECE OF SOFTWARE THAT TRIES TO BE EVERYTHING TO ALL ITS
CUSTOMERS IS NOT GOING TO BE AS FAST AS POSSIBLE DOING THE ONE
THING YOU WANT IT TO DO.”
with everything from HFT systems to
regular value based mutual funds. But
really, the point of our methodology is
to say that given opportunities come
and go, the speed with which one can
manage a team a developers to get
something up and running is very
important and it pays to plan ahead
using a standardised process, rather
than developing things ad hoc with
programmers, mathematicians, and
traders reacting to all manner of inputs
and discoveries. In other words, how
can we sift through thousands of ideas
quickly and find the 10 or 12 good ones
that show the most promise, and spend
our limited resources on developing
those, so that when we do find a good
idea we can put it in the pipeline and
everybody knows what's supposed to
happen when. Business practices are
becoming a much bigger determinant
of success.
Realising when a system no longer
works is an equally important part of
the equation. It's important to recognise that every strategy has a limited
shelf life. Even value and growth based
trades tend to run in five to ten year
cycles but still nevertheless eventually
go out of favour. The big thing is to
optimise your business processes to
look for as many opportunities as possible so when the window opens you
can pounce on it but at the other end to
recognise when the window is closed.
TA: Do you find that a lot of trading
systems are built around the quantitative side of technical analysis?
BVV: Most automated trading systems
are built on some form of technical
analysis. Generally when people think
about technical analysis they think
about Bollinger Bands and moving
averages, which are still quantitative
methods. If one thinks about statistical
arbitrage generally people don't think
about that as technical analysis.
However, it's still based on past market
prices and trying to understand relationships through mathematics. So
where you draw the line between one
and the other is sort of muddy.
However, most of the systems I see
being built are multi-instrument systems where they're trading one futures
contract against another, a basket of
stocks against another basket of stocks,
the options against the futures - something like that - rather than being in one
instrument and trying to pick the trend
up or down or trying to pick a sideways
market. The markets are getting more
efficient all the time and the trading
strategies are getting more complex.
Trying to control risk and uncover
short term inefficiencies is key
TA: Will the windows of opportunity continue to diminish in size,
beyond even the millisecond, and
does that mean that only the larger
players will have the necessary
July/August 2007
infrastructure to
opportunities?
exploit
these
BVV: As I see it, many of the areas in
which one can automate systems are
what I would call a commodity system
(not in the futures sense but a system
where the mathematics is well known),
for example calendar spreading and the
carry trade. Everybody understands the
mathematics of the carry trade and
there's not really anything proprietary
that you can dream up about the carry
trade. Well, let's call that a sandbox.
Who gets to play in that sandbox? The
biggest bully in that sandbox is the firm
that can throw the most money at
developing the fastest technology. So
the others have to ask themselves
where the opportunities are that they
can pick off elsewhere. The faster
things go the more expensive it
becomes to play in those sandboxes,
and smaller players get pushed out and
have to look for strategies that take a
longer time where milliseconds aren't
of the essence. You start to look at
strategies that may take minutes or
hours to work out probabilistically
rather than milliseconds. In the games
where probability of success is greatest
but technological advantage is the key,
the smaller player is going to be pushed
out.
Markets are always changing and this
creates all kinds of new opportunities.
But to the extent that risk is reduced by
holding positions for a shorter and
shorter amount of time, those kinds of
trades are going to be dominated more
and more by larger institutions that can
throw 10 million dollars at a problem.
It's happening already.
For further information on the
Institute of Market Technology's
Certified Trading System Developer
program, visit www.i4mt.org.
43
STRATEGY SPOTLIGHT
STEIN INVESTMENT MANAGEMENT
T
he Technical Analyst takes a closer look at the strategies and systems
employed by Stein Investment Management LLC, a registered Commodity
Trading Advisor that runs the "Trading Edge" program - a combination of
more than 20 uncorrelated mechanical trading systems, all trading the E-mini S&P
500 futures.
Boris Stein graduated from Minsk University in the former Soviet republic of
Belarus with a Masters degree in physics and computer science. After working as
a chief information officer for a major commercial company, he became one of
the first foreign currency traders in the newly independent state of Belarus. In
1995 he emigrated to the US as a political refugee. He formed Stein Investment
Management LLC in April 2006 and registered it as a Commodity Trading Advisor
in May 2006.
The Trading Edge program made returns of 180% in its first seven months of
trading in 2006 (from 1 June 2006) and made it into the list of Top 5 CTA programs under USD 10m in April's issue of Futures magazine. Around half of the
USD 9m assets under management belong to institutional accounts and approximately 20% is Boris Stein's own capital.
Why do you trade only the S&P 500 Index
Futures contracts (E-m
minis)?
Every market has its own personality
and I do not have time to delve into the
details and nuances of each of them.
The inefficiencies that I explore in one
market generally do not work in other
markets. If a trading system works in
many markets it can make only marginal profits. I have several reasons for
choosing S&P 500 index futures over
other futures: first, it is extremely liquid; second, I can use several unique
indicators outside of regular price data,
applicable only to stock indices (such as
TICK, TRIN, etc.); third, I am much
more knowledgeable in stocks than, say,
in pork bellies or cocoa.
What is your trading strategy?
We implement a strategy called "The
Trading Edge". It is a 90% systematic
and 10% discretionary program. The
program is designed to be as profitable
in a rising stock market as in a falling
market, because it assumes both long
and short positions. The program
incorporates around 20 rigorously
44
designed and tested independent
mechanical trading systems, all of
which are proprietary. I use all my 20+
models in parallel and take the trading
signals as they come, provided their
probable outcome meets the strict criteria of the proprietary risk control system. I do not automate my execution. I
evaluate each signal generated by my
systems before actually entering the
trade.
More than a half of my systems are
contra-trend, but the others work in the
direction of the most recent trend or
use seasonal indicators, where I evaluate the typical behavior of S&P at certain times of the day, certain days of
the month, and certain days of the year.
Which packages and systems do you use for i)
data, ii) charting and analytics, iii) backtesting and optimization, iv) programming and v)
execution?
How do you select which model to trade?
The two main criteria are the percentage of winning trades and the return as
a percentage of maximum drawdown.
It's about choosing the 20 best systems
among hundreds of other systems I
have designed over the last 12 years.
The number is not fixed, I may remove
a system if I see its performance
degrade or add a system from the
"pool" of other available systems if
they start to outperform. I re-evaluate
the systems a couple of times a month.
Are your models based on contrarian or trend
following strategies?
July/August 2007
For data, charting and backtesting/optimizing, I use TradeStation.
For execution, I use Trading
Technologies X_Trader. I also have JTrader and RanOrder as a backup. As a
former computer programmer, I also
write some "plug-ins" on C#.
What is the typical timeframe for your trading?
The average holding period is 2 days
and I usually make just one or two
trades a day. I don't enter the market on
days when I don't get any strong trading signals.
→
TM
TM
Describe the logical steps in your trade decision
making?
1. We analyse current market conditions - trends, sentiments and behaviour
2. We analyse the technical signals generated by our proprietary mechanical
systems
3. We analyse the trade risk and reward
according to our risk management system
4. If all of the above meets our criteria,
we enter the trade.
What kinds of data do you use?
I use intra-day price data, market sentiment and volume data. For gauging
sentiment, I use Put/Calls ratio, VIX,
have also used candlesticks for many
years. The main thing here, again, is not
to believe in the common rules and to
do your own research.
Do you use chart patterns in your models?
I like it when a popular chart pattern
fails. Usually, it's a good time to enter a
trade. I write some rather sophisticated
programs in EasyLanguage for
TradeStation to recognize patterns, as
well as just using my eyes.
Why do you think your models are making
money? What is it about the market they are
exploiting?
My systems work well in all market
modes - bull, bear, trending, or oscillat-
“I LIKE IT WHEN A POPULAR CHART PATTERN
FAILS. USUALLY, IT'S A GOOD TIME TO
ENTER A TRADE.”
and TRIN (Arms Index). With regard
to volume, it's noticeable that higher
volume coincides with a wider daily
range so there is not much help from
volume analysis on those days. So I
look at volume to find anomalies, i.e.
those occasions when it does not
mimic the daily range.
What kind of technical signals do you use?
Most of the indicators we use are in a
proprietary form. They are based on
moving averages, chart patterns, probability models, overbought/oversold
indicators, cycles & seasonal analysis,
and reversal indicators. For overbought/oversold indicators, I like RSI,
but I also use stochastics, and %Rs.
I use Elliott Wave, Fibonacci,
DeMark Indicators and Gann in order
to determine expected reversal levels,
but in modified form. No popular indicators work in the exact form they are
described in books, but any of them
can become useful after modifying
them and defining specific market conditions when they become applicable. I
46
ing. I do not know exactly why they
work, but I think it's because they are
designed using rigorous back-testing,
and because they exploit the inertia in
human behavior and habits. I believe
that in the short term markets are emotionally driven. I re-evaluate all my systems from time to time, because markets do change. Actually, I've had a difficult period for my systems lately, but I
have learnt the lessons, made the
changes, and feel more confident than
before.
How does your trading system adapt to
changes in volatility?
The trading system incorporates an
algorithm to track market volatility and
is capable of auto-adjusting and selftuning. Statistical volatility (standard
deviation) and average daily range are
my measures of volatility, and these are
used mostly to make stop-loss and target calculations.
How do you measure and manage risk?
July/August 2007
I use the Compromise Stochastic
Dominance method to measure risk,
because I believe it solves the major
shortcomings of the Mean-Variance
approach.
I manage risk by using a sophisticated computer based risk management
system for adjusting trade size according to the equity in the accounts, most
recent performance results of the
employed mechanical systems, and
market volatility. I also widely employ
stop-loss orders and time stops. It
should be noted that risk is also
reduced because of the diversification
between uncorrelated mechanical systems that comprise our trading
approach.
I enter only trades with a very high
probability of winning, so it's typical to
risk 10% of equity. The average cash
position is 85% of equity (minimum
5%, maximum 100%).
What is your performance objective?
I target monthly returns of 12%. I
would be surprised to see monthly
profits in excess of 25%. The theoretical maximum monthly decline is 25%.
Given the fairly limited time the Trading Edge
program has been operational, how confident
are you that your results are not down to luck?
I will not argue that luck is not needed
when trading, but I am confident it is
not the main part in trading success;
otherwise I would not have worked 16
hours a day for the last 12 years doing
my research. I made a living by trading
my own modest account for several
years, having no other income except
trading profits, and I won second prize
in the futures division of the World
Cup Trading Championship in 2006,
sponsored by Robbins Trading
Company. This I hope suggests luck is
not the main determinant.
Boris Stein is the managing member and president of Stein
Investment Management LLC
(www.steininvestment.com).
Training Courses
Training with The Technical Analyst
The Technical Analyst offers a range of
exciting training courses for traders and
investment managers. We also offer
specialist in-house training on request.
Course Details
AUTOMATED TRADING WORKSHOP
A 2-day workshop that instructs delegates in all aspects of
building an automated trading system. The course begins from
an introductory level and each delegate will be equipped with
a PC along with the appropriate charting, data analysis and
programming software.
24 Sep
22 Oct
08 Nov
26 Nov
03 Dec
Stockholm, Sweden
London, UK
Amsterdam, Netherlands
New York, USA
Frankfurt, Germany
Duration:
Courses are run from 9am to 5pm and
include lunch and refreshments.
Who Should Attend:
Traders, fund managers, hedge funds,
risk managers, brokers
INTRODUCTION TO TECHNICAL ANALYSIS
The essential technical analysis course providing a thorough
grounding in TA techniques for traders and investment
managers new to the subject.
25 Sep
Principal Trainer
Trevor Neil
Trevor Neil became a
commodities trader at Merrill
Lynch in the mid 1970’s
before going on to work
at LIFFE giving technical
analysis support to floor traders.
In 2000 he became head of technical
analysis at Bloomberg where he was
responsible for training and technical
analysis software development.
Other Training Courses
04/05 October
TA for the Portfolio Manager & Analyst
London, UK
16 October
Trading the Yield Curve
London, UK
18 October
Introduction to Market Profile
London, UK
London, UK
SHORT TERM TRADING WORKSHOP (2 DAYS)
Our very popular course for all market professionals looks at a
variety of trading techniques for developing an effective short
term trading strategy.
13/14 Nov
London, UK
ADVANCED TECHNICAL ANALYSIS (2 DAYS)
This highly regarded 2-day course provides in-depth training
in the most effective technical trading strategies for more
experienced market professionals.
26/27 Sep
01/02 Oct
08/09 Oct
15/16 Oct
London, UK
Amsterdam, Netherlands
Frankfurt, Germany
Copenhagen, Denmark
DEMARK INDICATORS
The only course of its kind available on this unique and
fascinating technique: Tom DeMark’s indicators centred on the
famous, ‘Sequential Indicator’.
23 Oct
London, UK
Full course details can be found at: www.technicalanalyst.co.uk/training
For further information email: [email protected]
Register online at: www.technicalanalyst.co.uk
or call: +44 (0) 207 833 1441
48
GET QUALIFIED IN TECHNICAL ANALYSIS
The Society of Technical Analysts (STA) represents and accredits
professional and private Technical Analysts operating in the UK
Originally established in the 1960s, the STA provides its members:
• Education
Monthly lectures and regular teaching courses in technical analysis
• Research
The STA Journal publishes research papers on TA techniques and approaches
• Meetings
Provide members the opportunity to discuss technical approaches and markets
• Representation The STA lobbies on behalf of analysts with data vendors, exchanges and regulators.
The STA represents the UK at the International Federation of Technical Analysts (IFTA)
• Accreditation
The STA Diploma Exam is internationally recognised as a professional level qualification
in Technical Analysis
For more information on how to join and what is involved in passing
the STA Diploma exam, visit our website at: www.sta-uk.org or call
us on +44 7000 710207
The ultimate ‘add on’ for
your Bloomberg
Market Professionals in over 20 countries rely on Updata
for the power it adds to their Bloomberg Terminal
Updata Technical Analyst is a
•
Unlimited charts and analysis
on multiple virtual screens
•
groundbreaking application
that unleashes the power of
your Bloomberg.
For a low monthly fee and
•
no additional exchange fees
you have state of the art
•
•
Charts using familiar
Bloomberg codes for fast ticker
entry including your CIXs
A wealth of additional technical
analysis tools and custom
indicator libraries
•
Hi speed scanning of large
data universes on any technical
criteria with real time alerting
Program and backtest any
technical strategy, optimise
trade entry and exit
•
Report Writer and highly
configurable charts with your
logos and notes
•
Price targets and risk reward
ratios on any chart
analysis, scanning and alerting.
Chart your own 3rd party data
including Excel sheets on the fly
To find out more type:
UPTA<GO>
Download your Free Trial now at:
www.updata.co.uk
For a CD, Trial or Demonstration:
+44 (0)20 8874 4747
Or just Contact Updata by email:
[email protected]
Updata is a registered Bloomberg 3rd Party Application type REG<GO> All trademarks are the property of their respective owners

Issue 23 - The Technical Analyst

Transcription

Similar documents

Read full article - MultiCharts Trading Platform

Small Business Development Fund

Marketplace - Traders.com

“SMC AUTO

Illegal Street Trading

PPT - Service Leadership Education

Upredictable Markets a Certainty

BOY SCOUTS NIGHT OUT AT BLUE MAN GROUP!

NEVER do these 10 things + 3 Trading Stocks

The Final Reckoning