FX Positioning - The Technical Analyst

Transcription

sept/oct 2005
The publication for trading and investment professionals
www.technicalanalyst.co.uk
FX
Positioning
A new tool for gauging
trader sentiment
Outlook for EURUSD
Interview
Chart Pattern Dynamics
Back on
trend?
Trevor Neil:
After Bloomberg
Reversal or
continuation?
WELCOME
It is not often that new indicators emerge that are genuinely valuable. The information
contained in trading flows has allowed dealing platform provider FXCM to develop a
contrary indicator based on the FX positioning of their retail customers. In this issue
we look at how they have developed the indicator as a proprietary service for
institutional users.
We continue our "Technical Analysis in the Commodity and FX Markets" series of
seminars in London on October 25th. This follows recent visits to South Africa and
Dubai where attendance by local traders and asset managers shows that interest in
all areas of technical analysis is growing strongly in the less developed markets.
Further information can be found on pages 30-31.
Matthew Clements, Editor
CONTENTS 1 > FEATURES
EUR/USD
Outlook remains uncertain
Claude Mattern of BNP Paribas explains that a
resumption of the euro bull run still depends on
breaking a key resistance level
FX Positioning
Gauging trader sentiment
Kathy Lien presents FXCM's new FX sentiment
indicator for institutional clients
Trading the Percentage
Volume Oscillator
Highlight Investments Group discusses how
to develop a volume-based mechanical
trading strategy
© 2005 Clements Biss Economic Publications Limited. All rights reserved. Neither this publication nor any part
of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the prior permission of Clements Biss Economic
Publications Limited. While the publisher believes that all information contained in this publication was correct
at the time of going to press, they cannot accept liability for any errors or omissions that may appear or loss
suffered directly or indirectly by any reader as a result of any advertisement, editorial, photographs or other
material published in The Technical Analyst. No statement in this publication is to be considered as a
recommendation or solicitation to buy or sell securities or to provide investment, tax or legal advice. Readers
should be aware that this publication is not intended to replace the need to obtain professional advice in
relation to any topic discussed.
September/October 2005
SEPT/OCT
>07
>12
> 17
>>
THE TECHNICAL ANALYST
1
8&#XXXDRHDPN
&."*-DPOUBDU!DRHDPN
-0/%0/
1"3*4
3&"-5*.&
5&$)/*$"-"/"-:4*4
2605&4
/&84
015*.*;&
.0/&:."/"(&.&/5
#"$,5&45*/(
*/%*$"5034
$6450.
53"%&4:45&.4
'03.6-"4
/08*/4065)"'3*$""/%5)&.*%%-&&"45
Trevor Neil goes back to trading after
Bloomberg
Larry Williams on the COT report
33
43
CONTENTS 2 > REGULARS
04
INDUSTRY NEWS
Editor: Matthew Clements
Managing Editor: Jim Biss
Advertising & subscriptions:
Louiza Charalambous
Marketing: Vanessa Green
Events: Adam Coole
Design & Production: Paul Simpson
MARKET VIEWS
EUR/USD: 1.2750 is key level for re-establishing bull market 07
10
US dollar index: The up-wave 5 is next
The Technical Analyst is published by
Clements Biss Economic Publications 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]
TECHNIQUES
Gauging FX speculative positioning
Trading the Percentage Volume Oscillator
Peak outs and permissions
Chart pattern dynamics
12
17
22
27
INTERVIEW
Trevor Neil, T-Capital
33
SUBJECT MATTERS
The process of price discovery
35
SOFTWARE
Insightful S-Plus
39
BOOK REVIEW
Trade Stocks and Commodities with the Insiders
by Larry Williams
43
COMMITMENTS OF TRADERS REPORT
LONG-TERM TECHNICALS
EVENTS
44
46
48
SUBSCRIPTIONS
Subscription rates (6 issues)
UK: £150 per annum
Rest of world: £175 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)
3
Industry News
NEW CHARTING
PACKAGE FOR MOBILES
TraderMade has released their latest
product TM-Cell which allows users
to access real-time quotes and charts
on a BlackBerry or mobile phone
without the requirement of additional hardware. Alison Trace, head of
North American sales, says, "Giving
users the ability to view short-term
charts will provide valuable mobile
insight into recent price action
allowing for more informed
decision making and execution
away from the desk."
www.tradermade.com
BIS shows exchange
trading up 11%
The Bank for International
Settlements (BIS) Quarterly Review
for Q2 2005 reported an 11% q/q
rise in derivatives exchange trading
volume for the second quarter at
$372 trillion. However, this was
down from the 22% increase seen in
Q1. Total volume includes fixed
income, equity index and currency
contracts. The BIS says the contin-
ued growth in trading is due to the
market’s changing view on the outlook for interest rates. As such, the
increase was greatest in short-term
interest rate derivatives although
activity in long-term bond contracts
declined slightly. Turnover in
exchange-traded currency derivatives
was up 15% in Q2 to $3 trillion.
CQG connects
to NYMEX
Charting software provider, CQG,
has been certified by the New York
Mercantile Exchange (NYMEX) as
an independent software vendor
allowing CQG users to trade on the
NYMEX ClearPort trading system
via the exchange's Electronic Order
Network (NEON). CQG customers
can already access the CME, CBOT,
LIFFE, EUREX and MATIF
exchanges.
www.cqg.com
DOW JONES NEWSWIRES AND
ESIGNAL EXPAND AGREEMENT
Dow Jones Newswires and eSignal
have announced an expansion of
their distribution agreement which
increases Dow Jones news and information available on eSignal,
4
QuoTrek, MarketCenter LIVE and
FutureSource workstations. The
Dow Jones Capital Market Report
and Economic Report will now be
available on eSignal along with Dow
Jones Commodities Services,
AgriWire and MetalsWire.
www.esignal.com
Industry News
ON THE MOVE...
Heiner Weber has moved from
DeTraCo (Derivatives Trading Corp)
in Switzerland where he was
involved in technical research to
become an executive director at UBS
Wealth Management in Zurich. At
UBS he will be advising ultra high
net worth individuals and be
involved in the development of
quantitative trading strategies.
Heiner
Weber
Ron Williams has joined Stockcube
research, the London based independent market research firm headed by David Fuller. Williams was
previously a technical analyst at
research house, IDEAGlobal.
New pointand-figure
guide
Harriman House Publishing has just
released “The Definitive Guide to
Point and Figure”, by Updata analyst
and P&F expert, Jeremy du Plessis.
The 500 page book is an exhaustive compendium covering all aspects
of point and figure including chart
construction and interpretation;
price targets, optimisation and oscillator overlays.
www.books.global-investor.com
EXCHANGE NEWS:
Eurex say that it will expand its product portfolio to include
futures on selected individual equities from October 24, 2005.
The exchange will initially list individual futures on all components of the European benchmark indices DJ Euro STOXX 50
and DAX, as well as 10 components of the SMI Index.
The LSE has extended SETSmm, its electronic order book, to
cover the constituents of the FTSE AIM UK 50 index. The
migration is scheduled to take place on 5 December 2005. The
remaining 100 Small Cap Index securities traded on the
Exchange's Main Market will also be transferred to SETSmm at
the same time.
The LSE saw 6.8 million UK and international trades in
August, an increase of 37% on August 2004, and the third highest monthly number of trades on record. The average daily
number of trades across the exchange was 308,142.
The CME says that for the first time in the exchange's history,
trading volume exceeded 10 million contracts, with 10.7 million
contracts traded on Thursday, September 1st, a global record
for a futures exchange.
NYMEX Europe has reported that the Brent crude oil
futures contract reached an estimated volume of 16,591 contracts on its first day of open outcry trading in London on
September 12.
5
Widen your horizons
Choosing a dynamic and innovating bank provides you with new perspectives and widened
horizons, to help you achieve your ambitions today and tomorrow. With SG Corporate
& Investment Banking all our clients - from corporate clients and financial institutions to public
sector clients and investors - benefit from sound analysis, reliable advice and the best financial
solutions. With a growing worldwide leadership in our areas of excellence, we will do our best
to respond to your financing, capital management or investment requirements by combining our
areas of expertise, our innovation and our cross-product approach ■ SG CIB, your partner in Euro
Capital Markets, Derivatives & Structured Finance
■
www.sgcib.com
Societe Generale is authorised by Banque de France and the Financial Services Authority, and is regulated by the Financial Services Authority for conduct of UK business.
EURO
CAPITAL
MARKETS
DERIVATIVES
STRUCTURED
FINANCE
Market Views
EUR/USD:
1.2750 is key level for re-establishing primary bull market
by Claude Mattern
"It is far better to foresee even without
certainty than not to foresee at all."
This statement from Henri Poincaré (a
French mathematician) might easily
apply to technical analysis in general and EURUSD in particular.
Following the downside correction
from the 1.3670 high at the end of
2004, the four year bullish trend of
EURUSD is currently in doubt for the
first time and we are now in a period of
great discussion. Yet - in the spirit of
Poincaré - an outlook on EURUSD is
still helpful in preparing us for, and
guiding us through, the market.
Here several traditional technical
analysis methodologies are presented in
respect of EURUSD to see if an overall map emerges to describe possible
outcomes.
1. Trends & Cycles
Strong intermediate correction
within a bullish primary market
EURUSD's primary bullish move from
the October 2000 low is the third such
primary bullish move since 1970 (see
Figure 1). Over this period, major
troughs appeared in February 1985 and
October 2000, while the major peaks
appeared in January 1980 and in March
1995. Assuming the first bullish trend
started at the beginning of the 70's, a
15-year a cycle can be observed, with a
10-year rise and a 5-year fall. The current rise is the beginning of a third long
term cycle. If such behaviour is stable a big if considering the cycle is based
on only two observations - a new
upward trend should be seen in the
next few years.
Thus, the secular trend is bullish. The
velocity can be compared with the
Purchasing Power Parity (PPP) between
the euro and the US Dollar, which is
well correlated with a 15-year moving
average, hence smoothing the cyclical
pattern. (Note: The exchange rate is currently above its 15-year MA at 1.1490 and the
15-year MA is currently flat).
Since the beginning of the year,
EURUSD has undergone a third intermediate corrective move, from 1.3660
to 1.1870, slightly above last year's
trough at 1.1760 (see Figure 2). Such a
move retraces the rise of the fourth
quarter 2004, which was mainly the
result of EUR portfolio accumulation.
Corrections are often a way to gauge
the strength of a trend by testing resistance to a contrary opinion. However,
this third correction has been sharper
than the first two, which suggests the
bullish trend seen over the past three
years is decelerating. The extension of
that correction below 1.25 also implies
a reversal of the 200-day moving average.
EURUSD is now correcting upwards
from 1.1860, with potential towards
1.2750 - a key level. From an Elliott
Wave perspective, a rebound above
1.2750 would reverse the bearish scenario of the past few months.
2. Patterns
Decelerating uptrend
Ever since a double top formed at the
beginning of the year, technical discussions have been strongly biased
towards
EURUSD
downside.
EURUSD has now reached a critical
point in relation to the double top formation. During the downside correction, two trend lines were broken and,
according to the fan lines principle, a
bearish reversal might be expected after
the breach of a third bullish trend line.
However, it's worth noting that during
the prior downside reversal period
(1995/96), a choppier market was seen
with more than three trend lines being
broken before a clear reversal emerged.
For now, the currency pair remains
Figure 1. Monthly EURUSD
7
Market Views
the indicator is on a key resistance (on
its 2-year moving average and the 50%
retracement of the former fall). A
rebound above 1.2590 in the next few
weeks might suggest that the low at
1.1860 marks a major trough.
4. Activity
Slight increase during the last
rebound to 1.2590 (Figure 3)
The transaction volume on the futures
contract at the IMM has increased over
the last few weeks. However, this market is only a proxy of the spot market.
Indeed, net FX turnover on EURUSD
amounted to an average USD 94.8 billion a day in April 2004 (excluding
inter-dealers transactions), according to
the BIS survey. During that same period the average daily notional amount
exchanged on the IMM was USD 7.9
billion, representing only 8.3% of the
spot market. Currently, the average
amount of contracts exchanged is USD
17.0 billion.
The rebound above 1.25 seen at the
beginning of September did so in tandem with a rise in volume and an
increase in open interest, providing further weight to the bullish signal.
Figure 2. Weekly EURUSD
Figure 3. IMM EUR FX future contract
within a long term bullish channel, with
the support line at 1.2060/1.2150 over
the next few weeks (return line at
1.4510/1.4600). Within that channel,
the exchange rate is evolving inside a
short term bullish channel, with
1.2170/1.2200 as support.
Key support levels: the recent downside correction met support at 1.1860, a
little above last year's trough at 1.1758.
The 38.2% retracement of the rise seen
over the past four years is at 1.1590 and
the 50% retracement level of the rise is
at 1.0950.
8
3. Indicators
Corrective rebound expected
After a triple swing in the last upward
movement the monthly RSI has been
moving downwards, giving a bearish
divergence. Such a configuration can be
compared with the reversal seen in
85/87 which signalled a deceleration of
the bullish trend. A return of the RSI
towards 50 could lead to a deeper correction, albeit without reversing the primary bullish trend. The weekly indicator is already rebounding and is now
around 50, after a low at 25. Currently,
5. Outlook
1.2750 is a key technical point
If EURUSD stays above 1.2170
/1.2200 in the next few weeks,
EURUSD is likely to return towards
1.2750 - the major level from a technical point of view. Indeed, a rise above
1.2750 would reverse the medium term
outlook for the next few months, suggesting a rise towards 1.30 at first (this
year's bearish trend line) or even above,
targeting 1.36.
The outlook for the next few months,
however, is conditional on a breach of
1.2750. Below this, a new downside leg
can be expected, arguing for a retreat
below 1.2160, towards 1.1900. A fall
below the latter would be a strong
warning of a major reversal. A rebound
above 1.2750 is favoured for now.
Claude Mattern, MFTA, is technical
analyst at BNP Paribas
Market Views
US DOLLAR INDEX:
THE UP-WAVE 5 IS NEXT
by Stéphane Billioud
D
espite the recent downside breakout
of the bullish channel that has driven the US Dollar Index* since the
beginning of the year, the decline that started in early July at 90.75 is still viewed as a
correction.
Indeed, from an Elliott Wave standpoint, the overlapping
of the subwaves suggests that this decline is the 4th wave of
the 5-wave rise initiated at 80.40 in late December 2004.
Given that based on Elliott-Waves principles, the bottom
of wave 4 and the top of wave 1 must not overlap, the
85.45/86.00 support area** is expected to limit the downside until the US Dollar Index enters the up-wave 5 (if it
has not already done so). The minimum target of this upwave 5 is the 90.75 region (top of wave 3) that should be
reached by the end of the year. The lower end of the medium-term bullish channel, coming recently at 87.61
(+0.25/week), should be the main intermediate resistance.
From a longer-term perspective, a break above the 90.75
high would argue for an extension of this up-wave 5 to the
92.30 region (high of May 2004 and top of wave 4 of the
preceding decline of a larger degree cycle) or even to the
94.40/73 gap left open since September 2003, before the AB-C downside consolidation that is supposed to come after
any 5-wave rise begins.
* The US Dollar Index is a trade-weighted average of six
currencies: Euro, Japanese yen, British pound, Canadian
dollar, Swedish krona, Swiss franc.
** Top of wave 1 & 2nd Fibonacci retracement of the
81.28-90.75 rise.
Stéphane Billioud is global technical analyst, FX and
Fixed Income Research at Société Générale in Paris.
Figure 1.
10
Techniques
GAUGING FX SPECULATIVE
by Kathy Lien
A new alternative
to IMM FX Data
12
Techniques
POSITIONING
T
he popularity of currency
trading has increased significantly over the past
three years. According to the Bank
of International Settlements
Triennial FX Survey published in
September 2004, foreign exchange
volume has surged 57% since
April 2001. Over the past few
years, many futures and equities
traders have made the switch to
trading foreign exchange in the
spot or cash market. However one
of the major impediments to a
clear understanding of the foreign
exchange spot market has always
been the lack of volume and positioning data.
The Commodity Futures Trading
Commission publishes a weekly
Commitment of Traders report that
breaks down commercial and noncommercial positioning in the futures
market. Although the report is very
valuable, it is released with a 3-day delay
and tends to be a reflection of smart
money speculative positioning. The
biggest non-commercial players in the
futures markets tend to be hedge funds
and system traders. Since most of these
professional traders are trend followers,
positioning tends to be skewed in one
direction when the trend is strong,
reducing the reliability of the COT
report as a contrarian indicator.
Acknowledging the need for more
transparency through real time positioning data in the FX market, FXCM
now publishes a weekly Speculative
Sentiment Index (SSI). Based on proprietary customer flow information,
the SSI is an index designed to recog-
nize price trend breaks and reversals in
EURUSD, USDJPY, GBPUSD and
USDCHF. Every bank has this information on their own client positioning,
but because of its profitability in house,
they rarely disclose this data. FXCM on
the other hand remains neutral and
does not trade against its clients. It is
therefore able to make this data publicly available. Institutional clients
receive a real-time snapshot of the data
twice a day (5am and 1pm EST) while
the general public receives a near real
time snapshot of client positioning
once a week on Thursdays at 10:30am
EST, which can be found at www.dailyfx.com.
Conceptually similar to contrarian
analyses using IMM open position data,
the SSI provides an alternative
approach that is more timely and more
accurate in forecasting currency price
movement than the IMM report. The
publicly available SSI is a near real-time
snapshot (as of 5am EST on the day of
publication) of market sentiment and
the open interest of small non-commercial forex market participants. The
indicator is compiled using aggregate
order flow information from FXCM's
over 55,000 non-commercial clients.
FXCM specializes in online foreign
exchange trading, and is the largest
retail currency-trading broker. As a
contrarian indicator, the index is precisely the most effective when the customer base is positioned "incorrectly."
The SSI offers a distinct advantage
over similar products. The main source
of this advantage is the fundamental
difference in the quality of its data sample:
Large and growing sample size: In
contrast to other indicators that tend to
rely on a relatively small sample of tens
or even hundreds of traders, the SSI is
based on a sample size of over 55,000
non-commercial customers. This large
sample size contributes to the robustness and accuracy of the SSI; and with
a consistently growing sample (3,000
new clients every month), the indicator
is continually able to improve.
Diverse sample: Unlike IMM data
that is susceptible to distortion from a
handful of large funds included in its
sample, FXCM's data is derived from a
diverse customer base representing a
much wider spectrum of traders ranging from small retail customers to larger high net worth individuals to small
institutional and hedge fund players.
The vast majority of FXCM customers,
over 95%, choose to trade for themselves rather than opt for managed
accounts, and as a consequence the
data sample encompasses an extremely
diverse set of individual trading styles,
risk preferences, time horizons and liquidity constraints. The SSI leverages the
globally diverse client base of the company's speculative traders, enabling the
index to capture global market sentiment, without containing a bias
towards American trading activity.
Active traders: Rather than commercial concerns looking to hedge their FX
exposure, the bulk of FXCM customers are speculators in the truest
sense of the word. Unlike the typical
"buy-and-hold" retail customer found
within many equity brokerage houses,
most FXCM clients are near- or microterm traders averaging two trades per
day and employing strategies that are
entirely technically based. Over the past
six months on average, FXCM customers placed over a million trades per
month.
→
13
Techniques
Ratios
(Negative ratio indicates net short)
Currency
Last
Week
Present
% Change in Positions
Outstanding
EUR/USD
GBP/USD
USD/CHF
USD/JPY
+1.32
-1.56
+1.43
+1.12
-1.16
-1.16
+1.75
+1.02
+2.3%
+1.9%
-5.3%
-13.6%
Table 1. Ratios from August 25, 2005.
Thus the size, breadth and activity of
FXCM's customer base are able to provide a good representative sample of
overall speculative behavior, especially
evidenced by the way FXCM client
trading activity closely tracks the most
and least actively traded currencies at
the Chicago Mercantile Exchange
(CME).
Data is "scrubbed clean"
To enhance its predictive value, the data
used in reporting is adjusted for trading
anomalies and unusually large positions
that may skew the indicator. The end
result yields a data set that is perhaps
the most accurate picture of FX
investor sentiment and positioning on
the market, an ideal data set for constructing a sentiment indicator.
Interpreting the SSI Data
Each week, the SSI data is displayed in
table format like in Table 1. The
absolute number is a ratio of lots
bought to lots sold, with a positive
number indicating that speculators are
net long and a negative number indicating that speculators are net short. By
construction, index values never lie
between -1 and 1. When the aggregate
number of customer lots bought for a
particular currency pair exceeds the
aggregate number of customer lots
sold, the index will be greater than positive 1, and vice-versa. In addition, the
index will never equal -1 exactly, again
by construction.
Interpretation of the index is straightforward. Excessively positive index values (greater than +3) imply that the
speculative market is positioned long in
Figure 1. The ratio of longs to shorts in the EURUSD is -1.16, which is within the extreme +/3 range. The EURUSD ratio just flipped to net short from net long yesterday afternoon and
since then the EURUSD has rallied approximately 150 pips. Over the past week, positions
remained relatively unchanged with only a 2% increase in open interest. Short positions rose by
27% while long positions fell by 16%. The contrarian SSI indicator is currently signaling more
gains ahead for the EURUSD with a possible test of the 1.24 level. The likelihood of an extension rally is especially strong since the USDCHF ratio is providing a confirmation signal.
(Institutional service, i.e. two bars per day - one at 5am EST and one at 1pm EST).
14
a given currency pair, while excessively
negative index values (less than -3)
imply that it is short. Consistent with its
use as a flows-based contrarian gauge,
the SSI generally signals a reversal in
price action when the index measures
greater than / less than three in
absolute value.
Each week there is also a historical
chart of previous ratios against price
action with respective commentary.
Figure 1 is a snapshot of our commentary from August 25, 2005 for
EURUSD.
Examples of Using the SSI
Price reversals at extreme levels:
Figure 2 shows that on July 21, net
positioning in USDJPY shifted from
longs to shorts into extreme positioning. At the same time, the brief sell-off
ended and USDJPY went on to rally
over 300 pips while net positioning
remained short.
As a Contrarian Indicator: Figure 3
shows that on December 20, the ratio
flipped from net long (+1.20) to net
short (-1.09). The currency at the time
was trading at 1.3285. Between the
times when the ratio remained in net
short territory till it flipped back again
to net long on January 3rd, the
EURUSD rallied 234 pips.
Our empirical analysis shows that the
flipping of the ratio is a more accurate
signal of a turn in prices than extreme
positioning for EURUSD and this is
true even if the flipping occurs from
fairly neutral territory. As such the SSI
is a contrarian indicator where a negative ratio signals more gains ahead
while a positive ratio signals more losses.
In addition to shifts in positioning
being a contrarian indicator, the areas
highlighted in Figure 4 also demonstrate that continual contrarian positioning by clients only supports the
opposing trend. In the area highlighted
to the left, positioning in EURUSD
remained in net long territory for 2 ½
weeks as prices continued to move
lower. Only when positioning shifted
from net longs to net shorts did a rally
Techniques
increased 32% over the week. The following day, the US released an exceptionally strong non-farm payrolls
report. Yet despite the optimistic
report, the dollar sold off 200 points
against the euro and many traders were
left confused with the price action.
However, it makes perfect sense
according to the SSI. With positions
nearing extremely short levels, and
given the short attention span of speculators and bears who have probably
endured significant losses throughout
the rally, the bounce in the EURUSD
after the 100-pip drop was clearly a
reflection of short covering.
Figure 2.
When to Use The SSI
Unlike other similar tools, the SSI is
generally more immune to abuse in
periods when it is least effective.
Traditional technical indicators actually
confirm the effectiveness of the indicator itself. Thus, when conventional
technicals perform well, practitioners
will know to avoid using the SSI.
Conversely, when other signs deteriorate, the SSI can be seen as an ideal
trading tool that identifies short-term
price reversals while generating comparatively fewer false signals. The SSI
performs best in relatively uncertain
choppy markets and/or periods of
information updating and forecast correction leading to potentially severe and
abrupt reversals in price. In these
instances, the technically-based speculative community tends to be positioned wrongly more often than not, a
consequence of unclear or conflicting
short-term indicators. The SSI works to
exploit this speculator uncertainty during periods of greater volatility and
two-way price action, giving index users
a great tool to gauge the direction and
magnitude of any possible reversal
movements.
Figure 3.
Figure 4.
occur. The same is true with the current price action. Overall, speculators
have been net short (net of one blip)
since September 8th, while the
EURUSD has done nothing but climb
higher.
Explanation of Price Action
There are many reasons to explain
shifts in positioning causing contratrend movements. For example, on
November 4th, the SSI report indicated
that EURUSD net shorts outpaced
longs by 2.73 to 1. Short positioning
Kathy Lien is chief strategist at
FXCM
FXCM, L.L.C.® assumes no responsibility for errors, inaccuracies or omissions in these materials. FXCM, L.L.C.® does
not warrant the accuracy or completeness of the information, text,
graphics, links or other items contained within these materials.
15
Techniques
TRADING THE
PERCENTAGE VOLUME OSCILLATOR
by Victor Kalitowski and Andrew von Stuermer
S
pikes in volume can be
used to anticipate index
trend reversals and can
be developed into a simple
low-risk, high-reward trading
strategy.
About Volume
Volume is the number of shares, traded
during a given period, for an individual
security, or for an entire index or stock
exchange.
Volume data simply indicates the
number of shares that were transferred
from one group of investors to another. Sometimes, this data is available broken down by market price; in those
instances, one can study how much
trading activity took place at specific
price intervals.
Two unique definitions are at the core
of our volume methodology: We call
any volume that occurs as an index or
share is moving lower "supportive volume"; conversely, any volume that
occurs during the time an index moves
higher is defined as "resistive volume".
As a general rule we have found the
following to be true: After a long run in
one direction, the appearance of a significant
volume surge signifies the transfer of a large
number of shares from one group of market
participants to another. It is at this point that
the market may become "overbought" or "oversold", whatever the case may be.
Based on an analysis of such volume
spikes (How extensive are they? How
far away from a previous reversal point
do they occur? How prolonged over
time are they?), investors can anticipate
when the market is likely to reverse -
over the short-, mid-, or long-term. On
a larger timeframe, a significant supportive volume spike can spark a bullish trend while a considerable resistive
volume surge can mark the onset of a
new bearish trend.
In our study of volume patterns on
the major indices (i.e., the determination and quantification of resistive and
supportive volume spikes), we use
index moving averages (IMA) as a key
tool. We also use the Percentage
Volume Oscillator (PVO). This is an
indicator that allows us to quantify the
magnitude and the duration of specific
volume spikes.
About the Percentage Volume
Oscillator (PVO)
The PVO measures the divergence
between two VMAs with different settings. The shorter-term VMA is known
as the "fast VMA" whereas the longerterm VMA is called the "slow VMA".
The formula for calculating the PVO is:
PVO = ([Fast VMA] - [Slow VMA]) /
[Slow VMA]
→
Figure 1. Example of an intraday PVO. S&P 500 index. April 22, 2005. Fast VMA = 15 minute
moving average; slow VMA = one-day moving average.
17
Techniques
By calculating and charting the difference between a slow and a fast VMA,
you can determine the extent to which
the fast VMA trades above or below
the slow VMA. This provides an indication of the intensity of short-term
trading activity (compared to the average trading activity over a longer time
span).
The PVO evaluates volume spikes
with unbiased, mathematical precision,
allowing you to gauge the impact a volume spike might have on the market
over different timeframes. For instance,
large positive oscillator values are
indicative of significant volume spikes.
Figure 1 shows an example of an intraday PVO. Two VMAs are compared: a
15 minute VMA as the "fast" VMA and
a 1-day VMA as the "slow" VMA.
Figure 1 clearly illustrates how the
peak value of the PVO corresponds to
the point at which the index reversed its
trend from down to up.
The PVO value provides an indication of the magnitude of a particular
volume spike. Because it represents a
percentage value, it has the added benefit of being amenable to comparisons.
The PVO permits you to analyze volume spikes that occurred in the past,
allowing you to establish associated
critical values at which an index trend
reversal is most likely to occur (over
various timeframes). You can use this
information to build a trading system
and tailor it to your personal trading
style.
Approach
We looked for critical PVO levels,
inflection points at which index reversals are most likely to occur. While the
following research was performed on
the S&P 500 and is most suited to a
medium-term trading approach based
on daily volume data, the basic principles outlined could also be applied to
other timeframes and other indices.
Based on tests over the last 250 trading days, we have established the most
appropriate critical PVO levels as follows:
18
Date
PVO
5-day change
in index*
Trend
VMA
Spike
Decision
04/21/2005
16.7%
-2.10
down
Supportive
Go Long
04/20/2005
17.6%
-36.29
down
Supportive
Go Long
04/19/2005
14.9%
-34.98
down
Supportive
Go Long
04/18/2005
14.7%
-35.23
down
Supportive
Go Long
02/01/2005
14.7%
21.00
up
Resistive
Sell Short
01/31/2005
16.4%
17.52
up
Resistive
Sell Short
01/28/2005
17.0%
3.49
up
Resistive
Sell Short
01/27/2005
15.2%
-0.86
down
Supportive
Go Long
01/26/2005
15.8%
-10.56
down
Supportive
Go Long
01/25/2005
14.9%
-27.57
down
Supportive
Go Long
01/24/2005
12.3%
-20.77
down
Supportive
Go Long
01/10/2005
11.7%
-11.83
down
Supportive
Go Long
01/07/2005
13.2%
-25.73
down
Supportive
Go Long
12/21/2004
15.9%
2.05
up
Resistive
Sell Short
12/20/2004
19.0%
-4.03
down
Supportive
Go Long
12/17/2004
20.5%
6.22
up
Resistive
Sell Short
12/16/2004
13.4%
13.96
up
Resistive
Sell Short
12/15/2004
13.3%
22.91
up
Resistive
Sell Short
10/25/2004
11.4%
-19.22
down
Supportive
Go Long
10/22/2004
11.6%
-12.46
down
Supportive
Go Long
10/21/2004
13.1%
3.20
up
Resistive
Sell Short
10/20/2004
11.3%
-9.99
down
Supportive
Go Long
10/19/2004
11.1%
-18.61
down
Supportive
Go Long
10/06/2004
14.0%
27.25
up
Resistive
Sell Short
10/05/2004
14.0%
24.42
up
Resistive
Sell Short
10/04/2004
15.5%
31.65
up
Resistive
Sell Short
09/24/2004
11.6%
-18.44
down
Supportive
Go Long
09/23/2004
14.3%
-15.14
down
Supportive
Go Long
09/15/2004
14.6%
4.10
up
Resistive
Sell Short
09/14/2004
13.7%
7.03
up
Resistive
Sell Short
09/13/2004
13.2%
12.19
up
Resistive
Sell Short
Table 1. Trading days with elevated PVO levels. S&P 500 index. August 2004 to July 2005. Fast
VMA = 5-day VMA. Slow VMA = 25-day VMA. * The 5-day change in the index is calculated
as the difference between the current close and the close 5 trading days ago.
Techniques
Date
VMA
Spike
Strategy One (using dollar cost averaging)
Decision
Open
Price
Close
Date
Close
Price
Returns
Average
Returns
04/21/2005
Supportive
Buy
1,159.95
8/3/2005*
1,237.92
6.72
04/20/2005
Supportive
Buy
1,137.50
8/3/2005*
1,237.92
8.83
04/19/2005
Supportive
Buy
1,152.78
8/3/2005*
1,237.92
7.39
04/18/2005
Supportive
Buy
1,145.98
8/3/2005*
1,237.92
8.02
02/01/2005
Resistive
Sell
1,189.41
4/18/2005
1,145.98
3.65
01/31/2005
Resistive
Sell
1,181.27
4/18/2005
1,145.98
2.99
01/28/2005
Resistive
Sell
1,171.36
4/18/2005
1,145.98
2.17
01/27/2005
Supportive
Buy
1,174.55
1/28/2005
1,171.36
-0.27
01/26/2005
Supportive
Buy
1,174.07
1/28/2005
1,171.36
-0.23
01/25/2005
Supportive
Buy
1,168.41
1/28/2005
1,171.36
0.25
01/24/2005
Supportive
Buy
1,163.75
1/28/2005
1,171.36
0.65
01/10/2005
Supportive
Buy
1,190.25
1/28/2005
1,171.36
-0.16
01/07/2005
Supportive
Buy
1,186.19
1/28/2005
1,171.36
-0.13
12/21/2004
Resistive
Sell
1,205.43
1/7/2005
1,186.19
1.60
1.60%
12/20/2004
Supportive
Buy
1,194.65 12/21/2004
1,205.43
0.90
0.90%
12/17/2004
Resistive
Sell
1,194.22 12/20/2004
1,194.65
-0.04
12/16/2004
Resistive
Sell
1,203.20 12/20/2004
1,194.65
0.71
12/15/2004
Resistive
Sell
1,205.72 12/20/2004
1,194.65
0.92
10/25/2004
Supportive
Buy
1,094.80 12/15/2004
1,205.72
10.13
Strategy Two (using 2nd volume spike)
Open
Price
Close
Date
Close
Price
Returns
Buy
1,152.78
current
1,237.92
7.39%
Sell
1,181.27
4/18/2005
1,145.98
2.99%
Buy
1,190.25
1/28/2005
1,171.36
-0.16%
Sell
1,203.20
12/20/2004
1,194.65
0.71%
Buy
1,094.80
12/15/2004
1,205.72
10.13%
Buy
1,103.66
10/21/2004
1,106.49
0.26%
Sell
1,134.48
10/19/2004
1,103.23
2.75%
Buy
1,110.11
10/4/2004
1,135.17
2.26%
Sell
1,128.33
9/23/2004
1,108.36
1.77%
Decision
7.74%
2.94%
0.02%
0.53%
10.09%
10/22/2004
Supportive
Buy
1,095.74 12/15/2004
1,205.72
10.04
10/21/2004
Resistive
Sell
1,106.49 10/22/2004
1,095.74
0.97
10/20/2004
Supportive
Buy
1,103.66 10/21/2004
1,106.49
0.26
0.97%
-0.02%
10/19/2004
Supportive
Buy
1,103.23 10/21/2004
1,106.49
-0.30
10/6/2004
Resistive
Sell
1,142.05 10/19/2004
1,103.23
3.40
10/5/2004
Resistive
Sell
1,134.48 10/19/2004
1,103.23
2.75
10/4/2004
Resistive
Sell
1,135.17 10/19/2004
1,103.23
2.81
09/24/2004
Supportive
Buy
1,110.11
1,135.17
2.26
10/4/2004
2.97%
2.34%
09/23/2004
Supportive
Buy
1,108.36
10/4/2004
1,135.17
2.42
09/15/2004
Resistive
Sell
1,120.37
9/23/2004
1,108.36
1.07
09/14/2004
Resistive
Sell
1,128.33
9/23/2004
1,108.36
1.77
09/13/2004
Resistive
Sell
1,125.82
9/23/2004
1,108.36
1.55
Total:
1.46%
31.54%
Total:
28.10%
Table 2. Using critically high PVO levels to generate simulated buy and sell signals. S&P 500 index.
Expected returns from August 2004 to July 2005 (250 trading days). * 3/8/2005: The date when the final data for this article was compiled. →
19
Techniques
• 13% as the critical level for resistive
PVO values;
• 11% as the critical level for supportive PVO values.
Our research indicates that these values will catch the majority of index
trend reversals and resulted in the
largest hypothetical returns over the
250 trading day period. Table 1 shows a
compilation of trading days with elevated PVO levels. It also provides the
difference between the current index
close and the close 5 trading days ago in
order to define supportive and resistive
volume spikes.
The most basic trading decision one
can make based on this information is
to sell on resistive volume spikes and to
buy on supportive volume spikes.
Based on the PVO values and trading
decisions presented in Table 1, we can
formulate two simple trading strategies:
20
Strategy one:
a) Close "Short" positions and open
"Long" positions on supportive
volume spikes;
b) Close "Long" positions and initiate
"Short" positions on resistive volume spikes;
c) Add to a "Long" position on each
additional supportive volume spike;
d) Add to a "Short" position on each
additional resistive volume spike.
Strategy two:
a) Open "Long" positions only on a
second or even a third supportive
volume spike;
b) Close "Long" positions on a first
resistive volume spike;
c) Open "Short" positions only on a
second or even a third resistive volume spike;
d) Close "Short" positions on a first
supportive volume spike.
The difference between the two
strategies is that the first applies a dollar cost averaging approach whereas the
second assumes a trader will only make
a single buy/sell decision. The dollar
cost averaging approach is based on the
premise that a trader will never remain
in cash - short positions are entered
immediately upon the closing of long
positions, and vise versa.
In the second strategy, traders are
assumed to limit their dealings to a single buy/sell decision. Until a good
opportunity arises, the trader will
remain in a cash position. A trade is
entered only on the appearance of a
second volume spike (which provides a
much better return than initiating a
trade on the first - or on the third - volume spike).
In Table 2, we have compiled the
hypothetical returns from these two
trading approaches (based on the last
Techniques
Figure 2. Depiction of the key trades outlined in Table 2. Dollar cost averaging approach.
250 trading days).
Table 2 shows that the use of a dollar cost averaging approach yields a
yearly return exceeding 31%. However,
only a portion of one's portfolio can be
invested this way, because funds need
to be held back in case a trader wishes
to add to an existing position should a
second, third, or subsequent volume
spike appear.
The second trading approach yielded
somewhat lower returns; however, here
the entire portfolio may be invested.
There is no need to hold back part of
the funds, as no further money is
invested should a second, third, or subsequent volume spike appear.
Both trading approaches are simple to
implement and both have yielded a
good return over the last 250 trading
days. Figure 2 points to the main short
and long trades established, as noted in
Table 2, using the dollar cost averaging
approach. As you can see, the system
catches the main mid-term trend reversals and thus provides stable, high
returns.
Discussion
The appearance of large index volume
spikes indicates a significant number of
shares are changing hands, prompting
shifts to the prevailing support/
demand balance. A wholesale transferring of shares during a mature up-trend
(i.e., the building of a resistive volume
spike) can ultimately lead to the exhaustion of the bullish group of investors the market may become oversold, making it prone to downside reversals. The
opposite is true when large supportive
volume surges appear during mature
down-trends.
Our PVO analysis uses a mathematical model to characterize price movements based on critical volume spike
levels. But the strategy of using elevated PVO levels to anticipate index reverSeptember/October 2005
sals remains highly sensitive to the status of the prevailing long-term trend.
PVO levels should therefore be re-analyzed on a regular basis so that critical
PVO values can be re-calibrated for the
current market situation.
The attractiveness of this system is
that is generates trading signals in a
mechanical, objective way, excluding
the emotional factor from trading decisions. Our research shows that this
approach can be highly rewarding for
medium-term players.
Victor Kalitowski and Andrew von
Stuermer, Highlight Investments
Group.
We wish to acknowledge MarketVolume™
for their analytical support and for the proprietary data and chart material presented
in this article. www.MarketVolume.com
21
Techniques
PEAK OUTS AND PERMISSIONS
by Cynthia A. Kase
Cynthia Kase explains how certain Kase Statware® indicators can
anticipate a greater percentage of market turns and can provide
speedier confirmation of trading signals.
A
fter becoming a self-taught
market technician, in January
1990, I left a major oil company I had been with for 10 years and
joined a major bank to manage their
commodity derivatives book. Back
then, energy derivatives were in their
infancy and so, in the face of the looming Gulf crisis and the war that followed, I was faced with trading using
only a screen and my wits, having exited the active ebb and flow of information one gets through constant telephone contact with counterparties. In
the absence of support from immediate fundamental information and the
weaknesses of traditional technical
indicators, the areas in which innovation was called for became obvious.
Back in those days I was using the
Stochastic, RSI and/or MACD to identify market turns and to exit when such
a turn was imminent. What I found was
that if one of these indicators generated a divergence the market turned most
of the time. However, there were a lot
of market turns that took place in the
absence of divergence - our recent
research shows around 55%. The problem is that all the traditional momentum indicators use very simple underlying measures of trend, like moving
averages or rates of change of closes.
So,
I
developed
the
Kase
PeakOscillator and KaseCD, which
work as well as the traditional indica22
tors in predicting turns, about 80% of
them just like the traditional indicators,
but also catch about 80% of the turns.
Around the same time I learned about
serial dependency in time series. A random event can set off a series of predictable, or serially dependent, events in
turn. For example, in late April 2005
after a protracted relaxation in prices in
the energy market there was a reversal
back to the upside following forecasts
that a series of freak snowstorms
would hit the US Midwest. The storms
were random, but even in the early
stages of the rally, the upside run
became predictable and perpetuated for
over a week. So, in thinking about
improving momentum indicators, the
idea of using serial dependency seemed
like the way to go.
Thus I came up with the Kase Serial
Dependency Index, or KSDI. What the
index does, when looking at "up" markets, is divide the natural logarithm of
the high n days ago to the low today, by
the volatility, which is also logarithmically based. The opposite calculation is
made for the "down". Since volatility is
a one standard deviation logarithmic
rate of change, we can think of it as a
standard of measurement for the market. The higher the ratio of the actual
market movement to this measure, the
more it is exhibiting serial dependency or "trendiness". Should prices move to
about two standard deviations, there is
less than a 3% chance that the "trend"
will continue and prices can then be
expected to revert.
Volatility = stddev (ln(P/P[1])n)
KSDI (up) = (ln(H/Ln))/volatilityn
KSDI (down) = (ln(L/Hn))/volatilityn
Kase PeakOscillator
There are two indicators that are
derived from the KSDI, the
PeakOscillator and the KaseCD. The
Kase PeakOscillator in a sense parallels
a simple oscillator which takes the difference between two moving averages.
But in the case of the PeakOscillator,
the indicator takes the difference
between the up and down indices. The
actual index value is based on an "n"
where the number of periods in question, over a range of lookback lengths,
gives the highest returned value.
Figure 1 compares the PeakOscillator
with the Stochastic and RSI. The
PeakOscillator is divergent generating a
reversal signal, while the traditional
indicators are not. Some fine points on
the chart are that the first peak on the
Stochastic cannot be used for comparison purposes as there was no matching
price peak to go with it. We usually recommend a tolerance of about three
bars between the price and momentum
peak for a comparison to be valid. Also,
instead of discrete peaks, small plateaus
Techniques
“KASE STATWARE
IS A LIBRARY OF
MATHEMATICALLY
SOUND,
STATISTICALLY
BASED TECHNICAL
INDICATORS.”
- CYNTHIA KASE
Figure 1. CAC-40 Index with Kase PeakOscillator, Stochastic and RSI
of up to three bars are allowable, as
shown by the red box. The fact that the
PeakOscillator "worked" and the other
indicators did not is illustrative of this
indicator’s vastly improved performance.
Also notice the "PeakOuts" shown.
These are discrete one-bar overbought
signals. Traditional indicator overbought and oversold signals often perpetuate for some time in trending mar-
kets. The PeakOscillator's lookback
length optimization focuses the signal
to usually one-bar, making the use of
overbought or oversold indications
practical.
KaseCD
The KaseCD is to the PeakOscillator as
the MACD histogram is to a moving
average oscillator. Each is an oscillator
of an oscillator, with the:
KCD= PeakOscillator - average (Peak
Oscillator, n)
As illustrated in Figure 2, the KaseCD
has similar advantages to the
PeakOscillator, in that it often catches
turns that traditional indicators miss,
and develops much rounder and less
choppy formations than the MACD.
Also unlike the MACD it generates discrete overbought and oversold signals
like the PeakOscillator, called
KCDpeaks.
The Permission Stochastic
It is often said that entering the market
is the easy part. All one has to do is
choose an indicator or group of indicators and get in on relatively simple signals. While that is true, one of the keys
to successful entries is using multiple
time frames to trade - any given chart
should contain the actual time frame
upon which the trade is focused and a
higher time frame as a filter. If the
longer term "trend" is up, then long
trades will be more successful and vice
versa.
When I first heard about the idea of
using a longer time frame filter, it
appealed to me, but I never had the
patience to wait for the filter. For →
Figure 2. Dow Jones Euro Stoxx, March 2004
23
Techniques
Figure 3. IPE Gasoil, June 2005
example, if on the open there was a signal generated after the first 15 minute
bar, I didn't want to wait another 45
minutes for an hourly bar confirmation. So I developed the Kase
Permission Stochastic, and in turn the
Kase Permission Screen.
The Permission Stochastic updates
every bar and has a variable time frame.
The way this works is to use the normal
Stochastic mathematics, but instead of
filtering say a 15 minute bar with a 60
minute bar, it filters a 15 minute bar
with an n (four in the case of a 60
minute bar) ending with the current 15
minutes. So a 15-minute bar generating
a signal at 0915 would be filtered by an
hourly bar from 0815 to 0915 and so
forth every 15 minutes. If you prefer
using Fibonacci numbers you could
change "n" from four to three, and use
a 45 minute bar, or five and use a 75
minute bar. Either way, any bar you use
to filter would generate a higher time
frame bar and thus a signal at the end
of the smaller time frame bar. That way
any signal gets accelerated as much as
possible.
For those who like to interpret the
fine points of indicators, the accelerated Kase Permission Stochastic may be
sufficient as a higher time-frame filter.
However, I have added two additional
24
layers of automation. First is to place a
rule set against the Permission
Stochastic and simply display it as blue
(permission to go long) and pink (permission to go short). A further refinement is to color-code bars themselves.
This is done in the Kase Easy Entry
System (KEES). It works like this. Blue
shaded dots are buy bars, red or pink
are sell bars. If a bar has a number on
it, such as 1, 2 or 3, an entry may be
taken. A "1" denotes that the short and
longer term time frames agree, a "2"
means that the longer term time frame
is ambiguous, but that there have been
two rising or falling bars as confirmation of the entry, and a "3" means that
the longer term time frame is opposed,
but that there have been three rising or
falling bars as confirmation of the
entry. The underlying signals themselves that generate the blue or red dots
are simple Stochastic, MACD and RSI
crossovers.
Final notes
Kase PeakOscillator, KaseCD, Kase
Permission Screen & Stochastic, and
Kase DevStops (see Jul/Aug05 issue of
the Technical Analyst) all form part of
the Kase StatWare, a library of mathematically sound, statistically based technical indicators.
More information can be found on
www.kasestatware.com, where you can
also sign up for a demo of the indicators. The best way to understand the
indicators is to use them. They are
available on CQG, TradeStation,
eSignal, Aspen Graphics, and .DTN
ProphetX.
Cynthia Kase is president of Kase
and Company, Inc.
Unfair advantage
Professional trader?
IQ-Trader
A front-end with an intellect.
Automated strategies, backtesting
and optimisation, charting and
analytics, automated spreading.
Looking for an unbeatable front-end?
Don't build your business on chance.
Get an edge.
Read about IQ-Trader and Pro-Mark in our
magazine at www.inside-trade.com.
Pro-Mark
The ideal tool for high-volume
professionals. Unique functionality,
unrivalled power.
London
+44 (0)20 7940 0470
Tokyo
+81 (0) 3 3668 9372
Chicago
+1 312 922 7600
Singapore
+65 6216 0454
New York
+1 212 425 2680
Sydney
+61 (2) 9293 2523
Asia and Japan Hedge Fund Directory
2005
Qsjodjqbm!Tqpotps;!
Tfdpoebsz!Tqpotps;!
Techniques
CHART PATTERN DYNAMICS
by Claude Mattern
What is the most effective method to trade reversal and
continuation patterns?
O
ne of the most
common and useful
tools in technical
analysis is a price pattern at
the top or bottom of a trend
or during a consolidation
period. Tradition has led to a
well-accepted pattern classification among technical analysts. However, the distribution of chart formations
between these categories is
not so clear.
Two main effects may be observed:
1. Prices will either oscillate around an
equilibrium price during a pattern formation, or
2. A price move will be induced by a
shift in the equilibrium price during the
exit phase of the period.
A reversal pattern needs, according
to John Murphy, a prior trend, a break
of an important trend line and a large
base or large volume on the break. A
continuation pattern is a pause in the
prevailing trend. Among the twelve
most important patterns there is a clear
consensus on the classification of five
patterns as reversal formations (doubles; triples; head-and-shoulders;
roundings and spikes) and on two patterns as continuation formations (flags
and pennants). But contradictions exist
for five patterns: triangles, broadening
formations, rectangles, diamonds and
wedges.
Although the distinction between
reversal patterns and continuation patterns is of limited use because some of
the patterns are indeterminate, there
appears to be unanimity on one point;
the exit from a pattern will tell the
whole story.
From accepted definitions, two categories of patterns emerge:
1. Those that are defined by their peaks
(or troughs) and a line break and
2. Those that are featured by a range
and two border lines.
Patterns defined by a series of oscillations, suggests a sideways pattern.
The definition is mainly linked to the
formation of the pattern, but it does
not imply the direction of prices. We
might conjecture, at this stage that the
basis of a chart pattern is a series of
oscillations, while a pattern defined by
its peaks only suggests that it is incomplete (a double-top might be a failed triangle).
A chart pattern is a series of
oscillations
Only the exit of a chart pattern, which
is defined by a series of oscillations, will
inform us about the direction of the
market. However, there are two complementary effects that influence price:
• A market adjustment where the
price oscillates around the equilibrium level. This is mainly position
adjustment,
known
as
distribution/accumulation periods
by technical analysts.
• A change of the equilibrium level,
due to a shift in supply and demand.
In this pattern the external conditions, mainly fundamentals, have
modified the beliefs and the opinions of the traders.
Defining a pattern as an adjustment
around an equilibrium price means an
exit of the pattern implies a major shift
in demand and/or supply. So, the first
important feature of a pattern is the
number of peaks or troughs before an
exit. The 'benchmark patterns' that
involve price adjustment around an
equilibrium price are triangles, rectangles and broadening formations.
Variations of those three patterns produce the wedge, diamond, flags and
pennants. Those patterns are assumed
to last until the forces of supply and
demand shift the price away from its
equilibrium level. The spike, the double-top, the triple-top and the headand-shoulders are all patterns that can
be classified as truncated triangles or
rectangles due to an earlier change of
the equilibrium price. As such, I propose a pattern classifications based on
the number of price oscillations (See
Table 1).
Analysis of the various chart patterns
has revealed that until the build-up of
the formation is complete, it is impossible to anticipate the direction of the
market. It is thus clear that within the
pattern, it is highly speculative to forecast the type of pattern that will appear.
We have seen that after three reversals,
the configuration remains open, even if
the market has already undergone a lot
of "consolidation distance." We have
also seen that the exit is the most →
27
Techniques
One-oscillation pattern
Two-oscillation pattern
Spike
Double-bottom and Double-top
Three-oscillation pattern
Head and Shoulders; Triple Top and Triple Bottom
Four-oscillation pattern
Triangle; Broadening; Diamond; Rectangle; Wedge;
Pennant; Flag
Rounded Bottom and Rounded Top
Non-clear oscillation pattern
Table 1.
important information of a chart pattern as it tells us whether the formation
or the oscillations have finished and
tells us the direction of the next move.
The exit is the result of a major shift in
supply and demand and traders must
intervene in the market with that idea in
mind.
the market. This phase of the pattern is
the most important one as this is when
one's position is managed. Four different exits can be surveyed: straight exit,
pullback exit, confirmation exit and exit
Trading the patterns
The progressive development of a
price pattern requires an adaptive strategy for the trader. This strategy could
be named the BET process, which
implies a three-step progression with a
strict order:
VARIOUS CHART
1. The Build-up period (B)
“ANALYSIS OF THE
PATTERNS HAS
REVEALED THAT UNTIL
THE BUILD-UP OF THE
FORMATION IS
2. The Exit of the pattern (E)
COMPLETE, IT IS
3. The Target (T)
IMPOSSIBLE TO
Each phase must be complete before
managing the following step. That
means that it is impossible to project a
target during the building of the pattern.
ANTICIPATE THE DIREC-
Build-up period
During this phase, no long-term positions should be committed, but previous positions should be kept.
Otherwise, range trading can be implemented, according to different scenarios.
failure. Those exits are closely related to
the way the market is trading a change
of trend - either a reversal or a resumption of the previous trend, after a consolidation. Position adjustments are
thus adding pressure to the fundamental shift of the demand and the supply.
The Exit
The exit of a pattern requires a shift in
demand and supply which reflects a
fundamental change in the opinion of
The Target
Some patterns have explicit targets
after the exit but these are only guidelines.
28
TION OF THE MARKET.”
For example, if the price exits from a
triangle on the downside, it implies
that, at a certain time or at a certain
price level, buying pressure will appear,
induced by the short position opened
during the build-up of the triangle. On
the opposite side, selling pressure will
appear after the downside break, on
closing long positions, stopping the
losses. So the amount of new open
interest built during the pattern will
induce the extension of the downside.
But this influence is only partial. Other
beliefs may also influence the strength
of the downward trend, rather than the
strictly technical points of view. That is
probably why the target of a pattern
should only be qualified as potential.
Conclusion
The behaviour of price after an exit of
a pattern will largely depend on the
type of product that is traded (equities,
bonds, commodities or currency pairs).
This can only be set by an ad hoc study
of the pattern, according to the market.
However, the pattern during its buildup cannot tell us where the price will go
later. The trader must wait for the exit
of the pattern, as it is only from the exit
that we have information about
whether the behaviour of the price has
changed. During the build-up of the
pattern, we have no information about
this change. We can only trade within
the pattern, but not beyond, as only the
market will tell us how it will exit.
Claude Mattern is technical analyst
at BNP Paribas in Paris.
)RUPRUHLQIRUPDWLRQFRQWDFW
*HUDOG3HUH]WHOHSKRQHLQ/RQGRQ8QLWHG.LQJGRP
RUHPDLOJSHUH]#LQWHUDFWLYHEURNHUVFRP
ZZZLQWHUDFWLYHEURNHUVFRXN
$XWKRUL]HGDQG5HJXODWHGE\WKH)LQDQFLDO6HUYLFHV$XWKRULW\
)6$UHJLVWHUHQWU\QXPEHU
presents
Technical Analysis in the Commodity and FX Markets
Technical Analysis in the Commodity and FX Markets is a one day seminar designed
to equip traders, hedge funds and fund managers with the most effective charting
techniques to trade these markets. With contributions from leading market analysts,
the event will present a range of studies that will enhance both short and long term
trading returns.
One Day Seminar – Tuesday, October 25th 2005
Trinity House, Tower Hill, London, EC3N 4DH
Speakers
Robin Griffiths, Rathbones
Head of Asset Allocation
Robin Griffiths was previously with HSBC
Investment Bank and is one of the global
financial markets’ best known and most
respected technical analysts.
David Sneddon, CSFB
Technical Strategist
David Sneddon joined CSFB from Investment
Research of Cambridge Ltd. in 1994 and is
currently a Director of the technical analysis
team based in London.
Thomas Anthonj, ABN Amro, Head of
Technical Analysis & Behavioural Finance
Thomas Anthonj looks at the role human
psychology plays in market analysis whilst also
focusing on more conventional techniques such
as Elliott Wave theory and Fibonacci.
Shaun Downey, CQG
Head of Technical Analysis
Shaun Downey provides daily research to
charting software provider CQG’s clients
such as banks and hedge funds. He is
also a partner at Oasis Research.
John Noyce, Citigroup
FX Technical Analyst
John Noyce joined Citigroup in 2000 becoming
European Technical Analyst in 2002. He covers
all asset classes with specialisation in the foreign
exchange markets.
Jeremy Goldwyn, Sucden
Global Head of Industrial Commodities
Jeremy Goldwyn has over 20 years experience
of the commodities and metals industry and
joined Sucden in 2004. He is currently Global
Head of Industrial Commodities.
Register online at www.ta-conferences.com
Sponsors
Registration fee = £399 + VAT
£100 ‘Early Bird’ discount when you register before 11th October 2005 = £299 +VAT
How to Register
1. Go to www.ta-conferences.com
2. Call +44 (0)20 7833 1441
3. Email your full name and address
to: [email protected]
Trinity House
Programme
Register online at www.ta-conferences.com
Desktop Quality
in a mobile environment
Information at your fingertips
Real-Time Quotes
and Charts on your
Blackberry or
Mobile/Cell Phone
-Don’t just look at where the
market is trading now. Get
the whole picture.
-Access to high quality real-time
charts defined by us or
the user.
-No software installation
means it is quick and easy to
set up.
-Reduce the number of mobile
devices you carry.
-Accessible world-wide 24/7
8 Archers Court
48 Masons Hill
Bromley
Kent BR2 9JG
United Kingdom
Phone: +44 (0) 20 8313 0992
Fax:
+44 (0) 20 8313 0996
E-mail: [email protected]
The RIM and BlackBerry families of related marks, images and symbols are the exclusive
properties of and trademarks or registered trademarks of Research In Motion Limited used by permission.
Interview
THE TECHNICAL ANALYST TALKS TO...
Trevor Neil manages the Isivuno fund at T-Capital in Cape Town,
South Africa. He was previously head of technical analysis at
Bloomberg.
TA: What was your incentive for leaving Bloomberg to set
up the hedge fund in South Africa?
TN: After four years as head of technical analysis there, I
felt it was time to get back to trading. I have been a trader
since my 'A' levels, preferring to go to the coffee futures
floor than to university. I had basically been running a posi-
tion for the last 27 years until I took the Bloomberg job. I
quickly realised that I needed technical analysis to time
trades. Cape Town appealed to me most as a place I might
like to actually live. When the opportunity to start a fund
there came along, I jumped at it. Why trade under grey skies
if you do not have to?
TA: What is your role at the hedge fund? What markets do
you trade and what techniques do you use?
TN: There are three of us trading the Isivuno Hedge Fund.
We are all technical analysts. My colleague and co-founder is
Jean Marx a respected trader in Johannesburg who has
operated a successful incubator fund at a large house there.
With us also is Andy Patterson, a well know technical analyst in South Africa.
Our mandate only allows us to trade South African equities, index futures and government bonds. Our trading is
almost totally technically driven. We need to know corporate actions but the fundamentals themselves do not determine our buying and selling. We use a trading plan to ensure
we are with the trend but we add alpha by taking advantage
of market reactions to go short as well. Our investors
expect us to make money when the long bull run in South
African shares comes to an end. To make money these days
as a technical analyst, using the RSI and MACD is not
enough. Everyone can see overbought stocks or a trend
changes in the futures or a trend line break in bonds. More
precision is needed and for this advanced techniques such
Tom Demark's indicators or some of our in-house developed indicators are required.
TA: The profile of TA has undoubtedly increased because
of its use by hedge funds. Why do you think this is?
TN: I think that most hedge funds appreciate that you need
technical analysis to time trades. When I was at Bloomberg
I met many fundamental analysts who were closet technical
analysts and I was also often asked by institutional desk →
33
Interview
heads to train their traders in technical analysis techniques
because their clients were using it. There is natural selection
in market analysis. Hedge fund managers tend to gravitate
towards this technique, because it works.
TA: During your time at Bloomberg, how did you see the
use of technical analysis develop in the markets? Do you
think TA is applied to a greater extent in some markets (ie
FX) than others?
TN: I think that technical analysis as a technique did take a
path through the markets. When the T-Bond contract started in the 1970's, many of the floor traders crossed from the
grain pit to trade bonds taking their TA skill with them and
through to the cash market. Then came the IMM currency
contracts and S&P500 index futures. I think TA skills
passed back to the cash market via the futures market. The
acceptance of TA in the equity market was slower. But this
is changing now probably lead by the retail market which
tends to be very sold on technical analysis. I think the rule
determining the effectiveness of technical analysis is: the
better the liquidity, the less the security can be manipulated
by an entity or group, the better it works. I would add that
it does not work well on low priced securities.
TA: Are there any particular techniques in TA that have
received attention in recent years?
TN: For me the best 'new' technique in technical analysis is
Tom DeMark studies. It is interesting that the TA used by
most people comprise the same techniques they might have
been using 30 to 50 years ago. We like the RSI, Bollinger
Bands, MACD and stochastics because they work. It is
somewhat comforting to think that what worked then still
works. In the meantime much new technical analysis has
come along but we have stuck with the things that work.
Tom's work is genuinely original and lateral in its thinking.
TA: You have spoken a great deal on combining technicals
and fundamentals. Do you consider this is vital even for the
most technical of traders?
TN: This must be the ultimate weapon. Even then I look
more towards a quantitative valuation for help. I want to
know if a company is already fully valued if I am long of it.
Just as I would want to know if it's RSI is at 98%. However,
that is not necessarily a reason to sell out. Many overvalued
and overbought stocks go on to double in price. I look for
fundamental analysis to answer the question "what?" and
technical analysis to answer the question "when?".
TA: Apart from running the hedge fund, you are also very
much involved in delivering training to financial institutions.
What kind of training do you do and to what kind of institutions?
34
TN: For many years I have been giving skill enhancement
guidance to traders. Mostly this is for people who sit on
trading desks or for fund managers. I have done this for
most of the largest banks, institutions and exchanges over
the last 15 years. I tend not to do basic indicator explanation. What people want from me is the practical application
of technical analysis in their job.
TA: Because so much TA experience is gained on the job
and passed from one trader to the next, does this mean
there is great scope for more formal and structured education in TA apart from what the STA is achieving?
TN: I do agree many traders I meet use technical analysis
heavily but learned it from the guy next to them or from a
book. They are aware that they are making trades on the
basis of two lines crossing but do not understand what this
actually means. What is a reading of 15% in the RSI actually telling you? Why is it sometimes not a buy but a sell?
Traders want this deeper understanding because large
money is at stake. I try to give a sound and relevant skill to
the trader.
TA: Rather than general TA education is there demand for
training in more specialized and advanced TA techniques
such as Elliott Wave, candlesticks and the use of indicators
and oscillators?
TN: There is without question. I was in Frankfurt recently
were everyone on the institutional sales desk was a TD
Sequential user. They saw it as their most important indicator for short term trades yet they wanted time with someone who understood it intimately and who actually used it. I
was also asked recently to do a seminar for an energy desk
where most of the traders used Elliott Wave. Basically they
had all taught each other over the years so wanted it reexplained properly as some bad habits had crept in.
Currently, there is a lot of interest in Ichimoku charting and
I expect to be asked to do that for some institution soon.
TA: Do you think banks and fund managers are well
catered for in this respect?
TN: I think there is a real skill shortage in this area. There
are very few people in the technical analysis world who have
a genuinely broad knowledge of the subject. Also, there are
few who trade actively. Traders really notice this and appreciate it. There are few people who have had this privilege.
TA: Do you still use Bloomberg for your charting?
TN: It might be I use Bloomberg because I am used to it.
Show me something better and I'll switch!
Subject Matters
THE PROCESS OF PRICE DISCOVERY
by Robert Schwartz
A simple game called Polya can help us understand how prices are determined.
C
an a study of market
data help us to time our
orders for best price and
full order execution? The classic
academic answer is no. They say
that markets follow a random
walk and that future prices are
entirely unpredictable - i.e. there
is no footprint left in the transaction record.
If this was the case, technical analysis
and algorithmic strategies shouldn't
work.
So what are these academics missing?
One only has to look at a graph of
almost any security price to see that
prices do not follow a random walk and
that trending tends to occur, most especially in bursts.
Most academics, and many participants for that matter, overlook the fact
that accurate price discovery and complete quantity discovery are difficult to
achieve. This is in spite of the fact that
price and quantity discovery are two of
the most important objectives of a
market.
Quantity discovery is difficult because
institutional participants do not readily
reveal their desires to buy or sell, and
big buyers and sellers have trouble finding each other. As a result, orders are
sliced and diced and worked over time,
and most of the demand to trade is
latent. Liquidity is difficult to find.
Price discovery is difficult for the simple reason that prices are not set by
analysts who come to unanimous
agreement over share value. Instead,
they are set in the market where orders
are turned into trades. As such, price
discovery results in a variety of symptomatic market characteristics that
include runs and reversals, and support
and resistance levels.
Such difficulties in price and quantity
discovery provide the foundation for
technical analysis.
A market network
In our research, we look at the marketplace as a network. Within that network, participants have different beliefs
about share value, but participant also
have the ability to adapt their beliefs.
Price discovery is a path dependent
process where multiple price equilibria
exist. In other words, there is more
than one potential equilibrium price
point and the equilibrium eventually
reached depends on the route taken.
Examples of path dependent, multiple
equilibria include - VCRs v Betamax,
Silicone Valley v elsewhere, share price
of 45$ or 55$.
Why 45$ or 55$?
The trouble begins with the investor's
inability to assess share prices with precision. This results in different valuations (45$ or 55$).
A highly simplified, stylized model of
stock price formation starts from this
difficulty. For example, can a share analyst say with precision whether the
expected growth rate for XYZ share is
7.00%, and not 7.545%?:
XYZ plc - Analyst evaluation:
Dividend one year from now = $1.35
Appropriate cost of equity capital =
10%
1) Investor A: Using growth rate =
7.00%
2) Investor B: Using growth rate =
7.545%
Share price if g = 7.00%, $45.00
Share price if g = 7.545%, $55.00
Deriving k
So if we accept the fact that investors
have different valuations, we can then
draw a simplified representation of the
market, where investors in XYZ share
fall into two groups:
• A bull will pay up to $55
(but may not pay $49 if thinks
$48.90 likely)
• A bear will sell down to $45
(but may not accept $51 if he
thinks 51.10 likely)
$55 - The Bulls (k percent of investors)
$45 - The Bears (1-k percent)
So, what will the market clearing price
be?
In a model we call the "HST Model"*,
price formation is entirely based on
participants placing orders with respect
to $55, $45 and k (the % of market participants who are bulls), and where:
• all three are common knowledge
• a plain vanilla limit order book market is used
• orders arrive in series
• bid and offer quotes can be determined
In figure 1, we can see that if we
know the high valuation (of the bulls)
and the low valuation (of the bears),
price discovery is equivalent to discovering the value of k.
In 2005, we went further in analysing
k and price discovery by incorporating
the concept of adaptive valuation, →
35
Subject Matters
[
$55 k percent
Starting Point
Ask
Bid-Ask Spread for k>0.5
Bid}
alls
B
d
Re
2, 1
1, 2
Figure 1.
Dynamic price discovery
In Polya, there is an urn of unlimited
capacity. The player starts with one red
ball and one blue ball. Then he or she
draws a ball at random, replaces it, and
adds one new ball with the same colour
as the ball selected. The player continues to pick another ball at random,
replaces it, and adds another new ball
with the same colour as the ball selected. And so on. As more balls are drawn,
the probability of picking a red ball
changes to match the percent drawn.
(See Figure 2)
Event 4
Combos & Probabilities
5 red, 1 blue, pr = .20
4 red, 2 blue, pr = .20
$45 1-k percent
i.e. where investors adjust their valuation according to what they think others believe. The model we used was
based on the following assumptions:
• Two valuations: $45 and $55
• Start process with one buyer and
one seller and common belief that
k=0.5
• Participants arrive one at a time and
reveal themselves to be buyers or
sellers
• Participants revise their share valuations and k expectations based on
the proportion of buyers they have
observed
• Participants can have a memory, that
is they can continue to give weight
to their initial belief that k = 0.5
The idea that investors disagree about
security valuation and that they change
their opinion based on what others
think leads to a process called Dynamic
Price Discovery. And the best way of
understanding this process of Dynamic
Price Discovery is to play a game called
Polya.
1 red, 1 blue
Blu
eB
alls
3 red, 3 blue, pr = .20
2 red, 4 blue, pr = .20
1 red, 5 blue, pr = .20
Figure 2. Shows what the proportion of red balls (k) will be after a large number of events.
In the stock market, price discovery
proceeds as follows:
• A trader pool of unlimited size
• Begin with one buyer and one seller
(at the start, k=0.5)
• Buyer sets bid, Seller sets ask,
PRICE is mid-point
• Each event is the arrival of a new
buyer or seller
• Of those who have already arrived,
k is the percentage of buyers
• PRICE is the mid-point of the Bid Ask spread
There are different ways to play the
game of Polya. These include using different starting values for k (where the
market is split unevenly between bull
Starting Point 1 buyer, 1 seller
s
yer
u
B
1B
1S
2, 1
1, 2
Sel
lers
and bear valuations) and by varying
how long initial beliefs about k continue to matter (where Memory of k0 (at
the beginning of the game) is zero,
moderate or strong).
For each play of the game:
• A path for k is traced out for 200
events
• Each value of k implies a price
• Each path for k converges on a
value k200
• Each path for price converges on a
value k200
• The k200 are distributed from 0 to 1
• The P200 are distributed from $45
to $55
Event 4
Combos & Probabilities
5 buyers, 1 seller, pr = .20, k=.83
4 buyers, 2 sellers, pr = .20, k=.67
1 buyer, 5 sellers, pr = .20, k=.17
Figure 3. Shows what k and price will be after a large number of events.
36
Subject Matters
10,000 Replications
1200
10 Replications
55
1000
54
53
800
52
Frequency
Price
51
50
49
600
48
400
47
46
200
45
1
6
11
16
21
26
31
36
41
46
51
56
61
66
71
76
81
86
91
96 101 106 111 116 121 126 131 136 141 146 151 156 161 166 171 176 181 186 191 196 201
Events
0
1
2
3
4
5
6
7
8
9
10
11
1,…,11 are eleven sub-divisions of the range
Figure 4. Typical Price Paths Starting k = 0.5, Moderate Memory
Figure 5. Distribution of k200 Starting k = 0.5, no memory
10,000 Replications
10,000 Replications
1600
3000
1400
2500
1200
2000
Frequency
Frequency
1000
800
1500
600
1000
400
500
200
0
1
0
1
2
3
4
5
6
7
8
9
10
Figure 6. Distribution of P200 Starting k = 0.5, No Memory
2
3
4
5
6
7
8
9
10
11
11
Figure 7. Distribution of k200 Starting k = 0.5, Moderate Memory
10,000 Replications
1800
Final Distribution of K
1600
2500
1400
2000
1000
1500
Frequency
Frequency
1200
800
1000
600
400
500
200
0
1
0
1
2
3
4
5
6
7
8
9
10
2
3
4
5
6
7
8
9
10
11
11
Figure 8. Distribution of P200 Starting k = 0.5, Moderate Memory
As can be seen in Figures 4 to 9, each
path is different and the distribution of
the final price is highly dependent on
k0 and memory. It’s noticeable that
such paths leave footprints or patterns
in the data, which – if picked up early
enough – can be profitably exploited.
Figure 9. Distribution of k200 Starting k = 0.6, Moderate Memory
Thus the game of Polya provides a
way of understanding price discovery
and, as a result, helps us to explain why
markets may not always follow a random walk.
Robert Schwartz, Zicklin School of
Business, Baruch College, CUNY.
*Puneet Handa, Robert Schwartz and Ashish Tiwari
(2003). Quote setting and price formation in an
order driven market. Journal of Financial Markets
** Jacob Paroush, Robert Schwartz and Avner Wolf
(2005). Dynamic Price Discovery in a Divergent
Expectations Environment. Baruch College working paper.
37
Looking for the key to your
quantitative financial solutions?
Attend the Applied Quantitative
Analytics in Finance Event and
Unlock the Future.
Please join us in London October 6th, for a series of guru-led
presentations, networking, and demonstrations by thought
leaders from the Basel II Committee, Swiss Union of Raiffeisen
Banks, Swiss Federal Institute of Technology (ETH) in Zurich, UBS
Warburg, Ingenious Media Plc. and Zurich Financial Services.
Contact us for more information:
Date: October 6, 2005
[email protected] or call +44 (0) 1256 339800
Time: 9:00AM - 5:00PM
www.insightful.com/thekey/
Place: Museum of London
Software
INSIGHTFUL SOFTWARE
Modeling Financial Time Series with S-PLUS® and S+FinMetrics™
by Eric Zivot
A
nalyzing, modeling and predicting financial time series
are necessary tasks for developing successful trading strategies.
However, these tasks can be daunting.
There are many steps to the process,
and having the right software tools
maximizes the potential for success.
S-PLUS, and the add-on module
S+FinMetrics, are ideal software
tools for statistical modelling of
financial time series, and rapid prototyping and backtesting of trading
strategies.
The S-PLUS environment is the premier solution for exploratory data
analysis and statistical modelling. It has
point-and-click functionality to import
various data formats, selected statistical
functions and graphics. It also has over
4,200 data analysis functions, including
the most comprehensive set of robust
and modern methods.
However, the real power of S-PLUS
for analyzing financial data and building trading strategies comes from its
flexible and powerful award-winning
object oriented matrix programming
language and graphics facilities.
S+FinMetrics is an add-on module to
S-PLUS with over 500 functions for the
econometric modeling and prediction
of economic and financial time series
that exposes the power of S-PLUS for
modeling univariate and multivariate
financial time series.
I have been an S-PLUS user for over
five years. I use it in my research and
my teaching. Many of my Ph.D students who go into the finance industry
use S-PLUS as their primary tool for
developing, testing and deploying trading strategies. In this article, I will illustrate how one of my students and I
recently
used
S-PLUS
and
S+FinMetrics in a research project
related to modeling and forecasting
high frequency intra-day foreign
exchange data*.
The main goal of the project was to
statistically model the dynamic intraday response of an asset to new information in arbitrage linked markets.
Typical arbitrage linked markets are
spot and futures markets, spot and forwards markets, and stock exchanges
trading identical assets. Assets traded in
these markets are tied together by certain arbitrage restrictions (e.g., cost of
carry relationships) which makes them
well suited for use in trading strategies.
In my project, I considered the behavior of the direct Euro/Yen spot rate
and the implied spot rate derived from
the Dollar/Yen and Euro/Dollar
exchange rates.
To illustrate the dynamic response of
new information with exchange rate
data, Figure 1 (created with S-PLUS)
presents one episode of exchange rate
movements around 23:50 GMT on
August 11, 2003, at which time Japan
released the first GDP estimates for the
second quarter of 2003. The line with
squares depicts the movement of the
dollar implied Euro/Yen price and the
line with triangles traces the direct
Euro/Yen price. The economic recovery data in Japan's GDP release →
Figure 1.
39
Software
caused appreciation of the Yen from
134.63 Euro/Yen to 134.50 Euro/Yen.
Both the direct and implied Euro/Yen
rates moved toward the new price level
following the announcement, with the
direct rate lagging behind the dollar
implied rate by about three minutes.
The figure clearly indicates an incidence
of how new macroeconomic information about the Euro/Yen price is first
transmitted through the dollar rates and
then through the direct rates.
The data preparation and statistical
modeling of the intra-day exchange
rate data to describe the dynamic transmission of new information involved
the following steps: (1) data cleaning
and variable construction; (2) estimation and evaluation of custom statistical models; (3) creation of specialized
graphics for displaying results.
Most statistical software packages are
not well suited for the above analysis.
In particular, they lack: flexible time
and date handling facilities and graphics
for irregularly spaced data; functionality for data cleaning; proper statistical
models; custom programming ability;
and capability for handling very large
data sets. S-PLUS and S+FinMetrics,
however, contain all of the tools
required for the project.
Data cleaning and variable construction
The raw exchange rate data (bid and
ask quotes) were captured from an electronic data feed and were recorded with
time-stamps up to the second from
midnight GMT. The data feed is not
always reliable and so the raw data
often contain errors and must be
cleaned prior to analysis. The time
stamps for different exchange rates are
generally not the same, and for the statistical analysis, all exchange rates must
be aligned to the same time clock.
Additionally, the foreign exchange
market is a global twenty-four hour
market with three major trading centers
(Tokyo, London and New York), which
exhibits complicated intra-day seasonal
patterns as trading passes from one
market to another. Furthermore, the
40
market is highly active except for the
weekend period Friday 22:00 GMT to
Sunday 22:00 GMT. Observations over
the weekend period behave differently
and should be modelled separately.
S-PLUS contains a powerful and flexible set of tools for representing,
manipulating and graphing irregularly
spaced time series data. As S-PLUS is
an object-oriented language, it has a
special class for representing general
time series data. For example, the first
five quotes for the Euro/Dollar
exchange rate on March 11, 2001 are
represented as:
> eurusd.ts[1:5,]
Positions
3/11/01 22:01:35
3/11/01 22:01:37
3/11/01 22:09:34
3/11/01 22:09:36
3/11/01 22:11:08
Bid
0.9326
0.9326
0.9326
0.9327
0.9322
The time series class allows for flexible date arithmetic and manipulation.
Important information such as time
zone and daylight savings time conventions are also easily incorporated into
the data.
Once data are represented as a time
series object, then built-in S-PLUS
functions can be used to perform a
variety of tasks such as aligning the
irregularly spaced data to a regularly
spaced time clock (e.g. every 5 seconds), interpolating missing values,
aggregating the data over fixed or moving windows, and computing technical
analysis indicators (see Box 1).
In preparing the data for the project,
I found that it was convenient to create
a library of new S-PLUS functions to
perform specialized tasks. This was easily done using the S-PLUS programming language. The resulting library of
functions can be shared with other SPLUS users.
Estimation and evaluation of
custom statistical models
The statistical modeling of the system
of exchange rate data involved a numSeptember/October 2005
ber of sophisticated econometric models including vector autoregressive
models, cointegrated vector error correction models and linear state space
models. Fitting, forecasting and simulation functions for all of these models
are available in S+FinMetrics (see Box
2).
Creation of specialized graphics
The dynamic response of exchange
rate prices to new information was captured by simulating the arrival of news
and then graphically tracing out the
future response of the exchange rates
Ask
0.9330
0.9331
0.9331
0.9332
0.9327
Institution
ONEC
AREX
NWHK
AREX
NWHK
from the estimated statistical model.
The model is constructed such that that
the arrival of news that leads to a one
basis point increase in the fundamental
value of the Euro/Yen exchange rate
should eventually increase both the
direct and Dollar implied Euro/Yen
rates by one basis point.
The construction of the dynamic
response path involved some custom
programming using the state space
modeling and Kalman filtering tools in
S+FinMetrics as well as the S-PLUS
Trellis graphics functions. Figure 3
illustrates the end result for the
Euro/Yen exchange rate during the
Tokyo trading hours. The solid blue
line represents the estimated response
of a one basis point increase in fundamental value, and the dotted red lines
represent 95 per cent confidence band
constructed from a bootstrapping procedure. The upper panel shows the
dynamic response of the Euro/Yen
exchange rate implied by the US Dollar
rate to new information, and the lower
panel shows the dynamic response of
the direct Euro/Yen rate to the same
new information.
The figure clearly shows that new
Software
Examples of S-PLUS functions
Aligning the irregularly spaced tick-by-tick data to a regularly spaced 5 second time clock can be done with the S-PLUS
align function:
> eurusd.5sec.ts = align(eurusd.ts, by="seconds", k.by=5,
how="nearest")
> eurusd.5sec.ts[1:5, ]
Positions
Bid
Ask
Institution
3/11/01 22:01:35
0.9326
0.9330
ONEC
3/11/01 22:01:40
0.9326
0.9331
AREX
3/11/01 22:01:45
0.9326
0.9331
AREX
3/11/01 22:01:50
0.9326
0.9331
AREX
3/11/01 22:01:55
0.9326
0.9331
AREX
The argument "how=nearest" is used to specify the tick nearest to the 5 second clock as the observation at that time.
Intra-day seasonal patterns may be captured using built-in spline and non-parametric smoothing functions. Outliers and
other data errors may be effectively identified using the unique collection of robust statistical methods. Custom time series
graphics, like the one shown in Figure 1, are easily produced. For example, the S-PLUS code to produce Figure 1 is:
>plot(exchdata.ts[timeEvent("8/12/2003 23:30:00",
+
"8/12/2003 23:59:59"), ],
+
reference.grid=F,
+
plot.args=list(type="b", pch=c(0, 5)))
> legend(0.2, 134.56, legend=c("USD/EUR * JPY/USD",
+
"JPY/USD"))
The plot function is a generic function which invokes the appropriate plotting method for the time series object exchdata.ts.
Box 1.
Examples of S+FinMetrics commands
Fitting a single lag vector error correction model with a prespecified cointegrating vector and a restricted trend to a bivariate system of exchange rates may be done with the single command
vecm.fit = VECM(exchdata.ts,
coint.vec=as.matrix(c(1,-1,0),
lags=1, trend="rc")
Figure 2 shows the h-step ahead predictions of log-returns,
with 95 per cent confidence bands, from the fitted vector error
correction model computed using the commands
vecm.fcst = predict(vecm.fit,
n.prdict=12)
plot(vecm.fcst, xold=diff(exchdata.ts),
n.old=12)
Figure 2.
Additional statistical and graphical diagnostics are readily produced to evaluate the fit of these models using the generic summary and plot methods
→
Box 2.
41
Software
“THE [SOFTWARE] CLEARLY SHOWS THAT NEW INFORMATION IS
INCORPORATED MORE QUICKLY AND EFFICIENTLY INTO THE
DOLLAR IMPLIED RATE THAN INTO THE DIRECT RATE.”
- ERIC ZIVOT
Figure 3.
information is incorporated more
quickly and efficiently into the Dollar
implied rate than into the direct rate.
The Dollar implied rate fully incorporates the information in about two and
one half minutes whereas the direct
rate takes more than seven minutes.
The figure clearly shows that new
information is incorporated more
quickly and efficiently into the Dollar
implied rate than into the direct rate.
The Dollar implied rate fully incorporates the information in about two and
one half minutes whereas the direct
rate takes more than seven minutes.
42
Summary
S-PLUS and S+FinMetrics are ideal
tools for modeling financial time series
and for developing, evaluating and
deploying quantitative trading strategies.
Eric Zivot is Associate Professor
and Gary Waterman Distinguished
Scholar in the Economics
Department at the University of
Washington. He is co-author of the
book Modeling Financial Time
Series Using S-PLUS.
* See "The Dynamics of Price Discovery" by
Bincheng Yan and Eric Zivot, available at
http://faculty.washington.edu/ezivot/ezresearch.htm
For more information about
Insightful please contact:
Richard Griffiths
Finance Account Manager
telephone:
+44 (0) 1256 339 823
e-mail:
[email protected]
website:
www.insightful.com.
Book Review
SECRETS OF THE COT REPORT
Trade Stocks & Commodities with
the Insiders: Secrets of the COT
Report
By Larry Williams
John Wiley & Sons, Inc.
209 pages, £31.99
ISBN 0-471-74125-6
How do you use the Commitment of Traders Report? This was a question that
the Technical Analyst asked several traders for an article in its Jan/Feb 2005
issue. We received several different answers, all of them plausible. We were
therefore very happy to come across a book that might provide a definitive
answer to this question.
The book does not disappoint. Larry William's experience and confidence is
clearly on show, and he speaks authoritatively on the meaning of the COT
report and presents some fairly detailed strategies on how to make use of this
potentially powerful dataset.
The COT report is released every Friday (www.cftc.gov) and provides a breakdown of each Tuesday's open interest for markets in which 20 or more traders
hold positions equal to or above the reporting levels established by the CFTC.
What is important is the change in the net trader position - the long contracts
less the short contracts - for each category of trader
The CFTC classifies traders into three groups: commercial traders (commonly thought of as hedgers), non-commercial hedgers (large speculators), and
non-reportables (small speculators). The big question for traders is: who should
we follow - the commercials or the non-commercials?"
According to Larry Williams, the commercials are without doubt the 'smart
money'. These are the hedgers who adjust their hedge book on a value basis, i.e.
they sell more as prices rise and buy more (or sell less) as prices fall. The strategy he puts forwards relies on placing the commercials net position in the context of their net position over the previous three years (by creating an index).
Next, he looks for extreme positioning by the commercials and uses this as an
indicator of imminent reversal. In other words, when the commercials are
extremely short relative to recent history, then a bull trend is likely to end.
But he doesn't get bogged down in a debate regarding commercials vs. noncommercials. To him, the other categories also tell an interesting story. As he
explains, the non-commercials tend to be trend-following and while they may
not do as well as you might expect, their strategies can still work. The big losers are the poor non-reportables - the small speculators who haemorrhage
money and whose side you should definitely avoid being on.
Larry Williams at times may appear over zealous. There are occasions when
you can't quite tie up his forceful and confident narrative with the slightly less
convincing graph in front of you on the page. He also spends much of the second half of the book talking about other trading issues, including the use of
technical analysis, and while his comments may often be good ones, this large
section seems more about getting things off his chest than an integral part of
his views on the COT report.
But it's of no consequence. This is an excellent and much needed book that
will probably be the benchmark text for traders in understanding the COT
report.
43
Commitments of Traders Report
COMMITMENTS OF TRADERS REPORT
14 September 2004 - 9 September 2005
Futures only (open interest) commercial and non-commercial net positions
10-year US Treasury
Source: CBOT
200000
700000
5-year US Treasury
Source: CBOT
300000
500000
Non commercial (LHS)
Commercial
Commercial
150000
600000
400000
200000
100000
300000
500000
50000
100000
400000
200000
0
300000
0
100000
-50000
200000
-100000
0
-100000
100000
-100000
-150000
-200000
-250000
14/09/2004
-200000
0
-200000
-100000
07/12/2004
01/03/2005
24/05/2005
Dow Jones Industrial Average
16000
16/08/2005
Source: CBOT
10000
-300000
14/09/2004
-300000
07/12/2004
01/03/2005
24/05/2005
Swiss franc
16/08/2005
Source: CME
50000
80000
Commercial
14000
Commercial
40000
60000
5000
12000
30000
40000
10000
0
20000
8000
20000
10000
-5000
6000
0
0
4000
-10000
-10000
-20000
2000
-15000
0
-20000
-40000
-30000
-2000
-20000
-6000
14/09/2004
-60000
-40000
-4000
-25000
07/12/2004
01/03/2005
24/05/2005
Pound sterling
50000
16/08/2005
Source: CME
40000
-50000
28/09/2004
21/12/2004
15/03/2005
07/06/2005
Yen
-80000
30/08/2005
Source: CME
60000
80000
Commercial
Commercial
30000
40000
60000
40000
20000
30000
10000
40000
20000
0
20000
20000
0
-10000
10000
0
-20000
-20000
0
-20000
-30000
-40000
-10000
-40000
-40000
-50000
-60000
-20000
-60000
-60000
-30000
14/09/2004
44
-70000
30/11/2004
22/02/2005
17/05/2005
09/08/2005
-80000
14/09/2004
-80000
07/12/2004
01/03/2005
24/05/2005
16/08/2005
Commitments of Traders Report
Euro
Source: CME
70000
40000
Commercial
60000
Review of the Commitments of Traders
Report released on September 9, 2005
20000
50000
0
40000
30000
-20000
20000
-40000
10000
0
-60000
-10000
-80000
-20000
-30000
14/09/2004
-100000
07/12/2004
01/03/2005
24/05/2005
Nasdaq
16/08/2005
Source: CME
30000
15000
Commercial
25000
10000
20000
5000
15000
10000
0
5000
-5000
0
-5000
-10000
-10000
-15000
-15000
-20000
14/09/2004
-20000
07/12/2004
01/03/2005
24/05/2005
Gold
16/08/2005
Source: CEI
0
180000
Commercial
160000
-50000
140000
120000
In the futures markets trends tend to continue
until the underlying market is over priced or
under valued relative to short-term supply and
demand. A characteristic of a trending market is
a building net position by the large non-commercial speculative traders (generally the commodity funds). The commercials (hedgers) net positions usually move in the opposite direction to
the non-commercials. If commercial net positions are near their largest or their smallest levels of the past one, three or five years, one can
imply that the trend of a given market has
pushed valuations too high or too low.
In the metals markets, gold and platinum have
recently had the largest net commercial short
hedge positions in the last five years suggesting
an over valued market. Silver has the smallest
net commercial hedge of the last year. The
interest rate futures net commercial hedge positions are neutral. The energy market net commercial hedges are just past the mid range of
the last five years and stock index net portfolio
hedges are near the smallest net hedge of the
last year suggesting a fairly valued to under valued equity market.
Currency market net commercial hedges are
generally not near over valued or under valued
extremes. Except for wheat, grains are also
generally neutral. CBOT Wheat has the largest
net commercial net long in five years suggesting
a grossly undervalued market. The net commercials in sugar have their largest net short position in five years suggesting the sugar market is
overpriced.
-100000
100000
80000
-150000
George Slezak
www.commitmentsoftraders.com
60000
40000
-200000
20000
0
14/09/2004
-250000
07/12/2004
01/03/2005
24/05/2005
16/08/2005
45
Long-Term Technicals
LONG-TERM TECHNICALS
Provided by Thomas Anthonj, ABN Amro, Amsterdam
EUR-USD: Broad range-trading between 1.1760 and
1.2930
USD-JPY: Nice bounce but worthless if 114.45 can't
be cleared
Trading in the middle of a broad range between 1.1760 (former bottom) and 1.2930 (potential left shoulder) it is difficult to pinpoint the
next move. But as long as the market continues producing higher
lows (last at 1.2127) we will stick to the view that we are in the
process of forming the right shoulder of a bigger H+S reversal pattern up to about 1.2930. A break below 1.2127 though would call for
a straight extension down towards 1.1760 from where the start window for a right shoulder up would be open again. Only a break
below the decisive support zone between 1.1760 and 1.1590 (38.2
%) would call for a much bigger setback towards 1.0950 and 1.0305.
The USD has shown a decent recovery in the last 8 months but
unless key-resistance at 114.45 (neckline/38.2 %) is broken decisively it looks like a re-test of the former neckline support which is
very common in an intact bear-trend. A break and weekly close
below trendline support at 107.35 would in this context be oil in the
fire of the bears who are still aiming at old bottoms at 101.70/25, and
potentially even at the H+S target around 95.80. A break above
114.45 would on the other hand reverse the earlier 3-years long
bear-trend.
GBP-USD: Below 1.8840/1.9140 further downside
expected
Gold: Triangle target at ~ 490 still in focus
Bouncing back up from key-support at 1.7310 (38.2 %), the market
could potentially have resumed its old up-trend. But as long as
key-resistance between 1.8840 and 1.9140 (trend line/76.4
%/potential left shoulder) is not taken out, the odds are in favor of
the bears in form of a big H+S reversal pattern. In this case we'd
have to expect a much deeper setback towards 1.5925 and potentially even 1.5065 (61.8/76.54 %). A break above key-resistance
would on the other hand call for a straight resumption of the uptrend targeting 2.0115 and 2.0165 (old top/Fib.-projections).
The breakout of the triangle consolidation observed lately is still
leaving room up to around 490. We stick to this view as long as the
market keeps on producing higher lows so that 429.50 should be
defended to keep the up-bias. A break below the latter would definitely call for a stronger setback to 379/372 (38.2 %/last major bottom), where the market would find support in case the whole uptrend was not completed yet. Breaking below the latter though would
only leave 354.40 (50 %) to prevent an even deeper setback to
330.25 or 300.30 (61.8/76.4 %).
46
Long-Term Technicals
DJ EURO STOXX 50: Intact trend channel calls for
3685/3865 next
S&P: 1253/73 could cause profit-taking - above,
1368 and 1553 next targets
The so-called accumulation phase seems far from being over and
remains valid as long as trend channel support at 3093 is not broken on a weekly close. Having said that we have to expect a minimum extension up to 3685/3720 (50 %/upper trend channel) or
3865 (old double top) before the market could be due for a
stronger setback. If trend channel support would give way we'd
have to be on alert as the market would have minimum downpotential to 2789 (38.2 %), which is a T-junction to distinguish
between a straight resumption of the up-trend and an even deeper
setback to 2429 or 2207 (61.8/76.4 %).
Although still in a bull-trend by definition we are on alert for signs of
weakness as the market is still trading within a Fibonacci-target cluster (1226 - 1246) for a potential top. This corresponds with the 61.8
% retracement at 1253 and trendline resistance at 1273. A close
below trendline support at 1204 would in this context deliver a first
hint that a bigger correction is due, but it would require a break below
the last major low at 1136 to confirm it. This would signal the end of
this accumulation phase and would give room for a minimum setback
to 1063/61 (38.2 %/old bottom). A break of the latter would even target 951 and 881 (61.8/76.4 %). To delay the risk of a bigger setback
the market would have to clear strong resistance between 1246 and
1273.
Brent Crude Oil future: Setbacks limited - next targets at 77.50 and 84.60.
Nikkei: Inverted H+S breakout projects 16800, but
14556 could cause a setback
The strong acceleration up is a clear sign that we are in the socalled maturing phase of a major bull-trend. Using Fibonacci-correlations we calculate targets for a potential intermediate top at 77.50
and at 84.60 for the Brent crude future. In case this contract would
break below the last top at 60.70 we would look for signs that we
are at least due for a minor setback that would most likely focus on
the light blue support zone between 51.37 (38.2 %) and 47.46 (last
major bottom). But this is the best we can foresee before the general up-trend is expected to resume.
Having spent more than a year in a sideways consolidation pattern
the market managed to not only to break and close above long-term
trend line resistance (now support at 11323) but also above neckline
resistance at 12282 and a former pivot point at 12788. Together this
is providing strong evidence that we have indeed started a new longterm up-trend. The inverted H+S formation projects a target around
16800, whereas a resistance cluster at 14556/15003 (old top/23.6
%) could cap the market for a while. Only a break and close below
the upward sloping trendline at 11286 would neutralize this bullish
momentum again.
47
Events
EVENTS 2005
Middle East Technical Analyst Forum
Technical Analysis in the Commodity and FX Markets
Dubai, United Arab Emirates, Tuesday September 27th 2005
With contributions from leading market analysts, the event will present
a range of studies and market views to enhance both short and long term
trading returns. Topics covered: Technical indicators, candlestick charting, intermarket analysis, & trading base metals.
Contact: [email protected] Tel: +44 (0) 207 833 1441
Organiser Date
48
Event
Venue
Contact
October
25
Technical Analysis in the
Commodity & FX Markets
London,
UK
[email protected]
+44 (0) 207 833 1441
October
12
Monthly
Meeting
London,
UK
[email protected]
November
09
Monthly
Meeting
London,
UK
[email protected]
December
07
Monthly
Meeting
London,
UK
[email protected]
December
06
'Technical Analysis in
the Dealing Room'
December
08
'Technical Analysis in
the Dealing Room'
Johannesburg, [email protected]
South Africa
Cape Town,
South Africa
[email protected]
GET QUALIFIED IN TECHNICAL ANALYSIS
The Society of Technical Analysts (STA) represents and accredits
professional and private Technical Analysts operating in the UK
The next STA Foundation
Course in Technical Analysis
starts 1st November at the
London School of Economics
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
Come and meet us at the Technical Analysis in the Commodity & FX Markets seminar on October 25th
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
In foreign exchange,
the most important
exchange starts here.
You and us.
At UBS, foreign exchange starts with the exchange between you and us. Understanding your needs.
Tailoring FX solutions to your requirements. And transacting them through execution systems that
are unequalled worldwide1. It's an approach that has won our FX business global recognition2. And
helped our clients to feel more confident about the FX decisions they make. Global FX strength,
bespoke FX solutions. You and us. www.ubs.com/fx
1
Best Bank Overall in Electronic Trading and #1 in Electronic Market Share, Euromoney FX Poll (2004); Best Bank in e-FX, The Banker (2004); Best Single Bank Trading Platform, FX
Week Poll (2004); 2 #1 Foreign Exchange House, Euromoney FX Poll (2004); FX House of the Year, The Banker (2004); Best Overall Bank in FX, FX Week Poll (2004).Issued in the UK by
UBS Limited, a wholly owned subsidiary of UBS AG, to persons who are not private customers. In the U.S., securities underwriting, trading and brokerage activities and M&A advisory
activities are conducted by UBS Securities LLC, a wholly owned subsidiary of UBS AG that is a registered broker-dealer and a member of the New York Stock Exchange and other
principal exchanges and SIPC.©UBS 2005. All rights reserved.

FX Positioning - The Technical Analyst

Transcription

Similar documents

Trading floor

Marketplace - Traders.com

Cavalia inc. 5100, rue Hutchison, suite 300 Montréal (Québec) H2V

Read full article - MultiCharts Trading Platform

PPT - Service Leadership Education

Small Business Development Fund

Illegal Street Trading

Brochure - QuoteStorm

“SMC AUTO

The Final Reckoning