CISCO Futures
1-303-306-1521 1-800 800 7227 Fax 1-303-368-9449
Market Profiles/Meta-Profiles Develop Market Structure
Foreword
Years before the concept of complex markets became
fashionable in economics
circles, Profiles were being used to analyze real world (complex)
markets.
Profiles use feedback from traders (agents) to discover auction market
demand (value). Value derived from price and time becomes the primary variable,
replacing
price itself. Value analysis frees market analysis from much of the noise and
misinformation of
price alone, making the Profile the fundamental element,
the starting point, in market
analysis. There are two disparate sources for profile data: cleared volume
in the original CBOT description (ca 1985) and the tick data method developed
by CISCO (1987). In balanced markets the two methods converge. If value
has changed within the timeframe covered, the methods may not agree.
Value Supplants Price: A question of reaction time
Market Profile, a History
The Tick - TPO (Meta) Profile
Internet http//www.cisco-futures.com
Email dljones@cisco-futures.com
November 8, 2004
Copyright 2004. All rights reserved.
The CBOT Market Profile (tm, CBOT)
A Market Profile is a graphic displaying price and volume on the vertical
axis, with cleared price activity on the horizontal broken down by trading
period. Market Profile is a subset of the Liquidity Data Bank (LDB) report.
LDB reports are released only by the Chicago Board of Trade (CBOT).
The horizontal activity is identified by letters for each time period (for
half-hour periods the letters are A = 08:00 to 08:30, B = 08:30 to 09:00 and
so on). These letters are called BRACKETS or TPOs (Time-Price-Opportunity).
Value is defined as those prices included within the central 70 percent
of the volume, beginning at the peak volume of the day. Market Profiles are
available only from exchanges that report volume at price. Market Profiles
come from cleared data and hence are not available in real time.
The CISCO Meta-Profile
A Meta-Profile replaces the volume of the Market Profile with tick data,
generating TPOs (That-Price-Occured or Ticked). These TPOs are used as a
surrogate for the volume of the CBOT Market Profile. Value is defined as those
prices included within the central 70 percent of the TPOs. Meta-Profiles are
generated in real time and are limited only to markets that produce tick
data. Meta-Profile methodology was created by CISCO in 1987, it was published
in 1987 and it has been in continuous use on the CISCO Bulletin Board and
website since then. In balanced markets values from Meta-Profiles and Market
Profiles agree quite well. In directional markets they may diverge, an effect
potentially useful to traders and market analysts.
A Typical Backyard Auction
Profiles come directly from auction market data. Price-over-time
and its relation to value can be
understood in all sorts of auctions, not just stock markets or futures.
Prior to going into detail on Profiles,
think of a standard auction, one for someone's household goods.
The time has come to auction three old-style Schwinn bicycles. All are
apparently identical. In selling the first one, the auctioneer suggests
a starting bid of $40. The bidding in dollars actually goes:
19, 23, 26, 28, 31, 33, 36, 37, 38, 40, 41, 42, 43, 44, 45, 47, 48,
50, 52 and 57, where the first bike is sold.
You are interested in buying one of the two remaining bicycles.
What information is available from the bidding? Can analysis
of the bids can
help discover the market structure of this auction?
A plot of the bidding price frequency distribution at $5 levels shows:
Dollars # Bids
56 - 60 X
51 - 55 X
46 - 50 XXX
41 - 45 XXXXX
36 - 40 XXXX
31 - 35 XX
26 - 30 XX
21 - 25 X
16 - 20 X
Fig MP-1: Bids on a Bicycle.
This price - volume (bids) display generates a profile of what independent
buyers were willing to pay for bike number one. Shape of the profile
(the X's) approximates the bell shape of a normal distribution.
A prospective buyer could look at figure MP-1 and surmise that value
likely lies somewhere between $36 and $50, in spite of the sale at $57.
When the second
bike went for $44 a potential buyer reasonably believes that one will
have to pay around $45 for the third bike. Anything under $36 would be a steal,
anything over $50 is too much. The buyer thus has a realistic,
market developed value range to work with and can devise a strategy.
Note that value is a range, not just one price.
This is because value is found from many votes (bids) and each person voting
(bidding) has a somewhat different opinion derived from their biases and
their interpretation of the feedback data (the bidding).
For general auction markets price in the form of transactions (ticks) is
the simplest output data.
In the bicycle auction, price changed as fast as new
bids could be elicited by the auctioneer. Market reaction time (feedback)
was quite short because the market was small. Larger auctions for equities,
derivatives, futures, etc.,
take longer to generate feedback (more traders equals more opinions and hence
more time required to come to a 'market' conclusion). Also, the wide variety
of agents sets the stage for a multiplicity of interpretations
of the observed market behavior.
These varied interpretations lead to the conflicts evident in price data
in the three cases below.
1) Demand differs at various prices.
A day's recorded summary prices (open, high, low and close)
can provide misleading signals: There is typically
little trading volume at the highs and lows, so these prices are often not
representative of market demand. Likewise, the open and close may well be
near the low volume highs and/or lows and themselves may not be representative.
Many analysts try to smooth out these discrepancies with moving averages and
other filter techniques. Filtering implicitly relies on a certain periodicity
in the market. The evidence for any such periodicity is weak. The structure
of the market with it's balance period followed by a directional phase (trend)
argues for no such general periodicity unless each balance and ensuing trend
occupy comparable timeframes, which is not observed. This structural question
is addressed in the section on market unit.
2) Tick streams give the earliest possible evidence for market change.
Most tick streams generate a good approximation to random noise much of the
time. Many traders
attempt to make sense of these data by collecting them into one or five minute
bars to integrate out some of the noise. This is helpful if the integration
period is long enough to be coherent. A five minute bar in a market that takes
thirty minutes to respond to feedback is usually far too short to be
meaningful (see item 3) below).
3) Coherent market response is driven by feedback.
Traders (agents) respond to changing conditions. This response time is
clearly longer than
a few seconds since ticks are so noisy. Just how long is a most important
practical question. If time to integrate a change is relatively
short, as in an efficient market, simple techniques like one or
a few minute aggregations of ticks becomes useful. If, instead, the minimum
time is relatively long, the short aggregations retain most of their noise
characteristics. One study confirming Steidlmayer's original 30 minute choice
sets the coherence time at twenty five to thirty minutes [MP-5]. We term this
the "natural" or relaxation time of the market. This natural time is consistent
with the market process: long compared to a 'tick time' but still short enough
to support multiple value changes within a trading day.
Natural time controls analysis in two ways. If a measure
e.g. one minute tick bars, is short compared to the natural time,
the measure will contain more noise than news. As measure time is extended
there will be increasingly less noise.
On the other end, if the
measure is too long, say a whole trading day, a great amount of significant
observable market phenomena will be averaged out. A proper timeframe permits
following market changes as closely as possible, within the time limitations
placed by the market itself.
Converting Price to Value
Since time is the catalyst in converting price to volume, the thirty minute
period is a good starting place. A word of caution:
the market is complex and future research may refine the measurement
of market response time. There are situations in which a market's natural
time varies. For this work, however, we accept thirty minute periods.
This can be understood as dividing a day into a series of thirty minute
parts, or mini-days, for analysis.
Typically, a market day will be six to eight hours (exchange hours).
In a seven hour day there are fourteen half hour periods. If one price, say
10132, trades in one period only, while 10250 traded in seven of the day's
periods, the price 10250 was traded in more timeframes and is thus more
popular than 10132. Restated, 10250 is most in demand and is most valued by
the market. Over the day a market will develop a hierarchy of values, often
forming a bell shaped curve of price over time.
In normal market situations a price-demand plot typically traces out a
somewhat bell shaped curve, narrowest at the top and bottom prices; fattest
in the middle. The shape of the curve translates price to value. Time is
the catalyst.
Price over time leads to relative value. Since the elevated demand periods
are associated with prices, price over time also leads to actual value. Value
found this way is interpreted within the overall time framework, e.g. a day.
Value found today may change tomorrow. Each day sets its own value and the
profile is the tool an analyst can use to track changing value.
Market Profile in futures is a trademark (1985, 1993) of CBOT that refers to
a specific database developed by the CBOT, the Liquidity Data Bank (LDB).
This data structure is a display of TPOs from clearing (TPOs record cleared
price
at time) and a value area calculated from volume at price. This methodology
can be called LDB - Volume. An LDB - Volume Market Profile is not real time
and is currently limited to CBOT futures, the only futures for which there
is an LDB report. As advances in reporting continue, volume may be reported
in the ticks of some futures. In such case, the 70% volume limits can be
calculated. From the beginning the Market Profile was tied to the gaussian
(bell shaped) distribution. A part of the practice was to take the early,
nascent distribution and identify it as the "third standard deviation",
evidently assuming activity later in the day would fill out the bell shaped
distribution curve.
First published in 1985 and updated in 1991, in the CBOT Market Profile manual,
the phrase "Market Profile" has become, over time, a generic name to describe
auction market analyses on a variety of markets outside of CBOT (e.g.
Steidlmayer on Markets, Vol II, p 84, 93). Quote services now offer "Market
Profile" analyses on markets not cleared by the CBOT. Also, a number of trader
websites sell value area data
identified as coming from the Market Profile. None of the three are possibly
using CBOT Liquidity data for those analyses on markets other than CBOT because
there is no Liquidity Data available. The methodology used to create these
new "Market Profiles" is a technique that utilizes quoter data (ticks) in place
of the CBOT cleared data in the LDB. This technology can be called the
"Tick - TPO", an approach developed and published by Donald Jones (Stocks &
Commodities, Sep 1987).
The profiles developed from ticks (TPO's) are called Meta-Profiles to differentiate
them from the volume based CBOT Market Profiles.
The overwhelming limitations of the trademarked CBOT Market Profile methodology
(cleared data report required, only CBOT futures covered and non-real time)
was the impetus for the development of the Tick -TPO methodology. This method
is a quoter (tick) based analysis, generating TPOs from tick ranges
and using TPO counts to measure demand. "Tick - TPO" profile methodology
is real time and and can be applied to all markets that produce ticks.
Tick - TPO methodology was published (1987) as a method for estimation of
the Market Profile value area.
In most of the futures world, the words
"Market Profile" will almost always refer to the Tick - TPO process, not
the LDB - Volume process trademarked by the CBOT. At CISCO, those reports
following the CBOT methodology will be clearly identified (e.g. Liquidity
Data Bank reports, Market Profile).
Market Profile versus Meta-Profile
In the work of Jones (MP-4) it was shown that balanced markets produced
the same value area. Recent research at CISCO has found that in directional
markets the two value areas can differ substantially. One or the other may
be correct, while the other will be quite wrong. For example, take a market
that opens on it's highs. Sometime within the first hour or so the highest
volume - price of the day (point of control) occurs. For the rest of the day
price is a lot lower, tracing out a clear bell shaped curve for the TPOs (TPTs
actually, for "that price ticked"). The price of the maximum TPTs is the
correct Tick - TPO point - of - control and the correct point - of - control
for the day. In this case the 70% volume is misleading. Of course, just
the opposite can happen, where, say, the market drops late in the day and
the 70% volume POC is the correct one. It appears to be a 50 - 50 proposition
for which will be right in a particular case.
Market Profiles were devised by J. Peter Steidlmayer [MP-1a, 1b]
for day timeframe analysis of feedback driven markets. Source data was
the end of day Liquidity Data Bank volume report from the Chicago Board of
Trade. An original "Market Profile" is a display of each price traded within
a trading day , as well as the time brackets during which each price traded.
The total volume traded at each price for each trading day is also included.
(Quoted from the CBOT website.)
The price - time pairs are called TPOs (Time Price Opportunity).
The CBOT website definition is consistent with the CBOT
Market Profile manuals [MP-1a]. Market Profiles are an integral part of the
CBOT Liquidity Data; the only source of volume for each price.
The value area (central 70 percent of volume) is based on eod volume
from the LDB. The Market Profile (TPOs) accompanying the LDB data was built
from the times within the day that the trades were cleared. This profile used
30 minute timeframes. There are two serious limitations to profiles made
from LDB data: first Market Profiles are available only for CBOT futures,
there is nothing for all other exchanges. Secondly, Market Profiles cannot
be calculated in real time, since LDB reports are available only periodically
throughout the day.
Steidlmayer's Market Profile concept integrates market behavior on a 30
minute basis. The time of integration (30 minutes) converts price to value.
Profiles are not inherently defined as 30 minute phenomena. Steidlmayer used
30 minutes probably because he found it to be a good balance between times that
are too short (movement into the volatility regime) and too long (loss of
definition). Value may change from day to day, or even within a day, since
it comes directly from the market's feedback and is responsive to changing
market conditions. Being able to measure such change requires adequate
coherence.
Steidlmayer's CBOT manual [MP-1a] and book [MP-1b] created a completely new way
of reading market demand, based on understanding the dynamics of the traders
as reflected in their feedback to the market. It is noteworthy that Steidlmayer's
insights predated academic economic thinking by many years. This is not unusual in
science. A solution for a specific need is often generalized and reinterpreted
later.
Jones in 1987 [MP-4],
developed another Profile application using real time tick data to
generate TPOs and found the 70 percent value area from that data, moving
the Market Profile concept into real time. Thus, there are two different
ways of arriving at value from trading data. In most of what folllows
'profile' will refer to real time tick
generated analyses (Tick - TPO), since volume is not generally available
from a quoter.
In the future, continuous volume is expected to become more readily available,
accompanying ticks, and so then the Tick - TPO Profile will have two ways
to calculate the 70 percent value area.
Profiles (Market and Meta) Identify Value
The illustration of the bikes at auction carries over to double sided auctions
for stocks, bonds, futures, derivatives, etc. A Tick - TPO Profile or Market
Profile (for CBOT futures) graphic similar
to the bids for the bikes, combines price and time to track
market development throughout a day. It locates value regions
of price acceptance and rejection, and identifies the market's
structure. The trader who uses profiles can learn who in the market
is active at which prices (buyers or sellers) and those prices the market
accepts and those rejected.
Changing the focus from price alone to value permits analysis of
individual prices such as the high and low and the most popular, which are
in the middle. It avoids
filtering problems and makes it possible to study a market's response to change
in real time.
Purpose of the Profiles
Profiles were developed to aid in trading the futures markets within the day time frame. It
was clear to Steidlmayer that the same approach was valid for
all auction markets [MP-1b]. As more work on the complexity of auction
markets appeared, it became increasingly obvious that
Profiles are the basic building block for auction market analysis.
Profiles provide a continuous probe of market feedback. As noted above,
profiles show the relationship of extreme prices with those better accepted,
solve much of the problem of tick noise
and provide the basis for determining market response time [MP-5]. In addition
to these three major questions, profiles provide a wealth of information
to the analyst, some of which is detailed below.
1. Value for previous complete days and the current day developing value
2. Deviation from previous day's value, change in value
3. Point of Control finds center of demand
4. Deviation of the market from the day's and previous day control prices
5. Level of trading activity at all prices, showing price acceptance/rejection
6. Who is dominating trade, buyers or sellers
7. Directionality of price movement within the confines of value
8. A prompt reading of change in a market
9. Shape of demand, compressed or elongated, growing/shrinking range, etc.
10. Price range and behavior of the initial balance (day's first two periods)
11. Congestion as it develops
12. How well the market is responding, trade facilitation
This extensive list is not necessarily complete. Profiles are
the swiss army knife of intra-day market measurement. In a complex system, new uses
for a measuring tool may arise at any time. Applications of interest to
economics theoreticians include studies of market equilibrium, practical
measures for dealing with crashes, volatility and market noise. Active
agents (traders) will be more interested in measures of opportunity,
short timeframe volatility and market pulsing.
Constructing a Tick Based (Tick - TPO) Profile
Profiles are made from ordinary market data (TPO volume, ticks, one minute
bars, five minute bars, etc.). We follow the lead of Steidlmayer and use
half-hour time periods. Thus, to make a tick based profile one simply needs
to know the high - low range for each half-hour time period throughout the day.
(We go further with a 15 minute based profile and also make use of profile
concepts in the Day Trading Engine.)
Half-Hour Tick Bars
A convenient way to evaluate demand uses ticks and/or TPO's (available within
the day) as a substitute for volume (which often is not available). The time period
used in figure MP-2 below is one-half hour, the basic time frame of the Market
Profile. For each half-hour period, the recorded ticks are posted vertically,
opposite price. Trading does not always move monatonically from price to price.
Sometimes there are skips. If there is no tick for the price, a space is left.
An Elementary Tick-Time Display
TRADING DATE: 01 APR 04
CONTRACT: JUN 04 DJIA (CBOT) DAY (DJM)
TRADING BEGINS 0720 (CST);CLOSES 1515;TPO SYMBOLS ARE yzABCDEFGHIJKLMNP
FIRST PERIOD IS 10MINS; SUBSEQUENT PERIODS ARE ALL 30 MINS
Half-Hour Tick Bars
y z A B C D E F G H I J K L M N P
103950 1
103940 1 1
103920 3
103900 1 2
103880 2
103800 1 3
103780 2 2
103760 3 1
103740 1 2
103720 3 1
103700 8 1 2 3 2
103680 1 1 1 2 1 2
103660 1 1 1
103640 2 2 1 1
103620 3 2 1 1 1 4
103600 6 3 1 1 5 4 3 4
103580 5 1 2 3 1 2 2
103560 1
103520 1 2 1
103500 2 3 6 3 1 2 1 1
103480 1 1 2 2 4 2
103460 2 1 2
103440 1 1
103420 1 1 1 1 2
103400 3 1 1 1 2 4 4 Cl
103380 1 2 1 4
103360 1
103320 Op 1 1 1 1
103300 4 5 2 2
103280 2 3
103260 1
103220 2
103200 4 2
Fig. MP-2. Half-hour bars: y = 07 - 07:30, z = 07:30 - 8, A = 8 - 08:30, etc.
Odd prices deleted to shorten the display. Op = open, Cl = close.
The tick counts show that after the open at 103330 the market
traded down and then recovered in C period 9:30 to 10. Price
peaked in D period (10 to 10:30) and then tailed off the rest of
the day. Periods G through J saw accumulation identifying value.
Gaps (missing ticks) indicate no trading at that price.
The occurrence of trading at a price in a half-hour period has a name in
Market Profile lingo, TPO (Time-Price-Opportunity); for Meta-Profiles it
would be TPT (for that price ticked). Each traded price in the period will
be identified by a letter
unique to that period, e.g. the period 8 AM to 8:30 is designated 'A'. In
figure MP-2 the time period letters appear in the first line above the the
price 103950. For period A, the first trade is at 103480.
These letters are the TPO's ('time price opportunity'; or 'that price tick
occurred'). Several TPOs at a price indicate the market returned to that
price over a period of time. The more re-visits the more popular that price.
For instance the price 103480 traded in periods z, A, C, H, I, M and N, or
time periods: 7:30 to 8 (z), 8 to 8:30 (A), 9 to 9:30 (C), 11:30 to 12 (H),
12 to 12:30 (I), 14 to 14:30 (M) and 14:30 to 15 (N).
Figure MP-3 is just figure MP-2 with letters substituted for the tick bars
and a column is added for each of the tick bars at each price. Price continuity
is assumed. TPOs fill the missing tick spaces. The price
column with the consolidated TPOs (entitled 'Brackets') is a TPO Profile.
The gaps with in a tick column are filled; e.g. in D period the gap at
103880 has assumed trading since prices on either side were traded. Center of
value (Point-of-Control) is designated by the character '>'. The first four
chapters of the book Mind Over Markets (see references below), is a valuable
source for a description of Market Profiles.
LENGTH OF FIRST PERIOD = 10 MINS
Tick - TPO PROFILE* REPORT FOR 04 01 04
AND SEGMENTED AUCTION
COMMODITY -- DJIA (CBOT) DAY JUN 04
Price Brackets Segmented Auction
103950 D D
103940 DE D E
103920 DE D E
103900 DE D E
103880 DE D E
103800 DE D E
103780 DE D E
103760 DEF D E F
103740 DEFKP D E F K P
103720 DFKP D F K P
103700 DFJKLP D F J |K |L P
103680 CDGJKLP C D | G J |K |L | P
103660 CGJKLP C | |G | |J |K |L | | |P
103640 CGJKLP C | |G | | |J |K |L | | |P
103620 CGHJKLP C | | |G |H | |J |K |L | | |P
103600 CGHJKLMP C | | |G |H | |J |K |L |M | |P
103580 CGHJKLMP C | | |G |H | |J |K |L |M | |P
103560 CHKLMP C | | | |H | | |K |L |M | |P
103520 zCHILMNP z C | | | | |H |I | | |L |M |N |P
103500 zCHILMNP z |C | | | | |H |I | | |L |M |N |P
103480 zACHIMN z A |C | | | | |H >I > | | |M |N |
103460 zACHIMN |z |A |C | | | | |H |I | | | |M |N |
103440 zACHIMN |z |A | |C | | | | |H |I | | | |M |N |
103420 zABCHMN |z |A |B |C | | | | >H | | | | |M |N |
103400 zABCHMN |z |A |B |C | | | | |H | | | | |M |N |
103380 yzABCMN |y >z >A >B >C > > > > | | | | | |M |N |
103360 yzABCMN >y |z |A |B |C | | | | | | | | | |M |N |
103320 yABCMN y | |A |B |C | | | | | | | | M |N |
103300 ABCN A |B |C | | | | | | | N
103280 ABC A B |C | | | | | | |
103260 BC B C | | | | | |
103220 BC B C | | | |
103200 BC B C | | |
Fig. MP-3. Tick - TPO Profile developed from figure MP-2.
Price bar is on the
left, next are the Profile TPOs, followed by all the half-hour bars,
stated in TPOs. The half-hour bars are not normally included on a
regular Profile (Tick - TPO or Market).
The vertical bars on the TPO columns are the value areas (central 70 percent
of trading volume)
as they develop throughout the day. At end of day, the
value area lies between 103670 and 103320. The center
of trading for the day is at 103550, with 9 TPOs (zCHIKLMNP).
Odd prices are deleted to conserve space.
TPO's are the smallest data unit in Profile theory. Profiles clarify and
explain a day's market activity. The multiple TPO region above and below the
value area (central 70 percent of TPOs) is called "Range Extension" to
identify possibly meaningful trading outside the value area. Single TPOs
at the extremes of the day's range are named "Tails". Tails are markers
for where the price went too high or low, usually on low volume. Figure
MP-4, a more compressed version of figure MP-3, posts the parts of the profile.
Profile Structure
Profiles are a graphic of market structure, locating value
and demand. They show accepted prices, rejected prices and the most
popular prices.
Value is found among the accepted prices. The first two periods of the
day, the Initial Balance, may be among either accepted or rejected prices;
which it is becomes clear later in the day. Jockeying between traders is
defined by Range Extension (trading outside value) and tails (extremes).
All these data can be rightly called "feedback information". So a profile
is a continuous feedback monitor.
Tick - TPO Profile
Tick - TPO PROFILE* REPORT FOR 04 01 04
AND SEGMENTED AUCTION
COMMODITY -- DJIA (CBOT) DAY JUN 04
Price Brackets
103950 D | Upper Tail
103900 DE |
103850 DE | Upper Range
103800 DE | Extension
103750 DEFK |
103700 DFJKLP |
103650 CGJKLP |
103600 CGHJKLMP |
103550 zCHIKLMNP | |
103500 zCHILMNP | Value Area | Initial
103450 zACHIMN | 70% of Trade | Balance (first two periods)
103400 zABCHMN | |
103350 yzABCMN | |
103300 ABCN |
103250 BC | Lower Range
103200 BC | Extension
Fig. MP-4. A compressed version of of the half-hour bars of ticks and/or
periods traded converted to TPOs is to the right of the prices. Farther
right are the half-hour TPO bars. The profile clearly illustrates the
clustering of activity around the central price 103550. Value, defined as
the central seventy percent of trading, is 103670 to 103320. Value is not
a price, it is a range of prices, illustrating the market's choices. Prices
near the top and bottom, the extremes, are traded relatively lightly,
indicating rejection by the market.
Legend
See Dalton, Jones and Dalton [MP-2] for a thorough discussion of the
details and development of Market Profiles.
Price Range Periods Description
103550 - 103350 z,y Initial Balance, first two periods
(Floor member control)
103950 D Upper, Selling Tail, 2 or more single prints
(Public sellers control)
Note: the single print, D, is technically not a tail
103900 - 103700 D - P Upper Range Extension, above Initial Balance
(Public buying pressure)
103250 - 103200 B - C Lower Range Extension, below Initial Balance
(Public selling pressure)
None Lower, Buying Tail, 2 or more single prints
(Public buyers control or public sellers
control discussed below)
103650 - 103300 y thru P Value Area, central 70% of trading
Wider = better trade facilitation
Higher (from prev. day) buying
Lower (from prev. day) selling
103550 Point of Control, maximum # of TPO's
A list in the Purpose section enumerated a dozen pieces of market
information available from profiles. Tomorrow, today's profile will
be history. The recent historical data is a research set for
identifying change, general market behavior and all the many elements
that go into the development of markets. Just as in capital market
studies [MP-6], where a history of returns led to a measure of volatility
and risk, here a much richer set of information exists because of the
complexity of feedback in markets.
Profile Characteristics
By the nature of auction markets, Profiles have a quasi-bell shape
in balancing (quasi-equilibrium) markets. Market Profile views the shape
as a gaussian, Meta-Profile makes no such distinction, accepting the
clustering around the median price as adequate to identify value.
Prices tend to go too high to
show they have gone high enough and similarly for the lows. The Russian
proverb cited earlier of two fools in the market can be recalled.
In a large market with many active
agents it is not too hard to find the trader who sells too cheaply or pays too
much. It happens daily.
Directional, non-balancing markets display changing value. The form of the
profile is much more elongated, less time on average is spent at any single
price and the profile structure may show several maxima where price paused
during a run. Of course, the market variables can still be found as in a
balancing market, but understanding can be more complicated since balances
have a large volatility component.
Volatility, an Application
Profiles provide both a visual and mathematical way to measure
volatility [MP-3]. A discussion of volatility is beyond the scope of this
note. It is enough to point out that volatility found via the Market
Profile methodology depends on experimental data without relying on
any assumption of a governing distribution function.
The Market Unit
Auction markets are constantly seeking value. A perceived change in value will trigger
a change in price. If value is stable (price is clustering) the market condition is
'balance'. A balanced market lists the center of value and the upper and lower limits
of the balance. There is enough information there to make a trade setup.
If the market is not in balance, market condition is either 'run' or trend or so
choppy that there is no balance. Market condition goes from balance to run to balance
to run and so on. The time from one run to the next is a 'Market Unit'. Posting the
successive Market Units over a long period graphically displays the wide variation
of one Market Unit to the next. This is evidence of non-cycality in auction markets.
The Market Unit
Data Quality
In taking tick data and converting it to Meta-Profile the two prime elements
are the formula (rules) and the quality of the data. Data quality depends on
the error rate and the amount of data available.
In the tick data of today, error rate is rarely a factor. Quantity varies
widely over a 24 hour day. For US markets there are often periods of no ticks
at all (say around midnight). Calculations with these data are invalid. Often
trading starts picking up at 5 or 6 AM and is going full tilt by 8:30 AM.
Some deliveries never have adequate data for valid calculations. The CISCO
Tick Status Table (across from the flag) lists each delivery with
color coding (green and yellow are ok, red and gray do not have enough data
for trustworthy calculations).
Other Limitations on Profiles
The basic premise of profile methodology is the 'bell' shape of the profile
distribution. This assumes the profile fits a 'Normal' distribution, a form
well known in sophomore statistics classes. A Normal distribution peaks in
the middle (price at peak volume for a profile) and the first standard deviation
includes about 70 percent of the volume about the peak. Profiles are found
with all levels of fits to the 'Normal', from quite good to not at all (think
a double distribution, one with two peaks).
The standard calculation of profile elements (point of control, value area, etc.)
does not include some standard of required fit for validity. Indeed, such a test
would be very difficult to devise. Consequently, a user of profile methodology
must be on guard and not use profile elements in trading decisions blindly.
References
MP-1a CBOT Market Profile (c1984)
MP-1b Steidlmayer, J.P, and Koy,K, (1986). Markets and Market Logic, Porcupine
MP-2 Dalton, J., Jones, E., Dalton, R., (1991). Mind Over Markets, Probus
MP-3 Jones, (2003). Bringing Volatility to..., Futures Mag. June 2003
MP-4 Jones, (1987). Estimating the Market Profile Value Area for Intraday
Trading, S&C Sep.
MP-5 Jones, (2003). Volatility and Stops..., Internal report, CISCO Futures