Auction Market Theory and Market Profile

EURUSD 5,10,20 day Overlay with bracketing on all 3 profiles. Are we heading to another breakout?
http://i.imgur.com/PGMoLgq.gif

EURUSD 5,10,20 month overlay, Bracketing, With weekly volume lower as it retracing up. Any signs of what the commercials are doing from your calculation, if any, mzvega? Or we already exhausted up here?
http://i.imgur.com/U857k0w.gif

A rough calculated "guesstimate" tells me yes, the data/numbers tells me yes, however as I make changes to my spreadsheet, I am working on the functions, its hard to keep an ongoing database, if your working on it most of the time, so I am unable atm to tell if its an "above avg".......Lets see if I can find you an example...... but for the moment great charts. Im no longer looking at weekly vol, relative to previous week (learning new things). what I'm doing is tracking the overlays vol on a day to day basis. Its a matter of every day knowing if you have increasing or decreasing vol of the ie. 5 day distribution of prices. Helps determine participation..... I got the same chart "view" as well. I also for intraday(minor auction) use the 2 day and the 3w/the current (0-3 merge).

I feel the up auction is finish in geppy.now its time for distribution and to see whether price going to pull the poc up or is it going to create a new poc.this is a quarterly chart.

the above is a wrong chart I have posted I will take it out as seen as possible as site is blocking from editing my own post.must be some bug.

Tks mzvega, interesting idea on the use of volume shift on the moving 5day overlay profile, I may have to revise my thoughts on that.

Price rotation today test the lower limit & quickly rejects & recover to where it started today. Very strong responsive move with high conviction. Sure looks like the upper limit will be tested tmr for a breakout perhaps.

http://i.imgur.com/A65oROc.gif

there's some reference material that helps explain the concept....... I just need to find it.
By tracking it daily, it also helps to isolate it, by octs & qts. The LTF participants operate in octs 1 & 4 (in a bracketing distr.) So in a bracket you can analyze the the LTF positions, That should give you the bias, once its b/o, no longer bracketing , the daily vol changes of the distr. allows you to monitor its "flow". Its an interesting concept......

I've placed most my trades today on that bias......

tku mzvega.

Break out during Asia session above 5day limit, UK open drove it back down towards the 5day-mid after it broke below the upper Oct. Then US news release drove it back up to the upper limit & breakout this time. Basic model yielded a 2/3 win.
http://i.imgur.com/YMt53vS.gif

We've gone from balance to to trending on the 3,5,& 10. Spite the crazy PA, was able to find a few entries today today......

UJ provided a good responsive opportunity during Japan session open, my entry was just nice & close to the lower octant. Target to Mid & close the trade. UJ is in 5 & 10 day balance, so either it extend its 5 day balance downwards to the lower limit of the 10day balance or rotate within the 5day which it did today.
http://i.imgur.com/2dB85vB.gif

EU continues upwards on like what mzvega mentioned above, trending mode. It may be gunning to the rn of 3700. Today POC is below 3675 though so value is not building higher for now. Monday maybe we see if any, acceptance within this 2 levels mentioned.
http://i.imgur.com/26a4V0I.gif

Here is Kiwi pair i mention a few post ago about it being in very good balance. It has since gone on responsive since from 0.8100 & breakout today above the 5 & 10 day resistance. The 20 day is also Balance. Very good market condition to operate in for this pair.
http://i.imgur.com/5xUSRo5.gif

Great chart! You may be able to refine and open a few more windows of opportunity....... a balanced market has one cohesive distrib., a trending market has more than one dist. to the overlay. I've been learning that 3tpo's define brackets in the 1,2,3,5 overlays. 4tpo's for 5 & above. But I did see in the same material mention 3tpo's for 1,2,3 & 5tpo's for the others. Its hard to tell, the chart doesn't have tpo's so I i cant tell if "technically" you may have been able to "measure off" other distb.'s in the overlay, giving either a refined entry or other possible window of opportunity.........

I'm so glad you're back.......

3tpo will be the limit for overlay from 1 to10days, 5tpo is 10days to 20,it was mentioned in the book. for larger overlay 5tpo is used due to the higher number of Total TPO but in most cases, they are more likely be align with the short TF overlays. I can get good estimate even when i using the histogram (instead of TPO) as i double check on the 30min bars itself just to make sure i have it pin with reasonable tolerance. So even with multiple distribution overlay, the choice would be to tread on the larger distribution, provided the risk reward is acceptable.

Tks for the feedback.

Starting with Thurs. EOD analysis. Thurs EOD analysis the distribution shows that we were still trending from the BO of the LTF distr. From last week. However, the internal data showed that TFF increased 9.49 to 11.40. That was the first clue that this trend may be slowing. Add an increase in volatility (39.53 to 11.40), tpo’s skewed to the downside, Attempted dir., IB range & % increase, etc…….. The data suggested that the trade plan for the following day (Friday) was to short the market until the chart or data says otherwise.

And Friday’s EOD analysis shows we in fact came to a pause in the short TF distribution. Volatility & TFF also decreased & tpo’s skewed to the upside. I’d like to think this is just a minor pause in the current trend due to the fact that we are still trending on the LTF distributions. I will be looking for signs of initiative or responsive behavior to the shorter TF distribution for the best opportunities……………………

I just wanted post an article for those who may find some time to read this Holiday. Being that we trade "Auction" markets, I thought that traders ought to know about the "type" of markets they participate in to better understand how best to analyze it......Many people under estimate the value of "tick data". The OHLC price bar charts are "2 dimensional" in nature, when you add tic data variables, you're able to view it in a more realistic "3 dimensional" way ........just another way of looking at the market..........

CISCO Futures

1-303-306-1521 1-800 800 7227 Fax 1-303-368-9449
Internet http//www.cisco-futures.com
Email [email protected]

Market Profiles/Meta-Profiles Develop Market Structure
November 8, 2004
Copyright 2004. All rights reserved.

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.


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.

Demand can be read from the frequency display (the X's). Bidders spent little time at the lower prices, recognizing they were not realistic. There was also little enthusiasm at prices above $50, although the buyer at $57 likely was proving the old Russian proverb "there are two fools in a market, one asks too little, one pays too much".

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).

Value Supplants Price: A question of reaction time
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.

The full article can be found here : http://www.cisco-futures.com/marketprofile_basic.html

TkS claudia

Appreciate your post & too on that weekend analysis of of the Cisco post which is very educational, do keep them post if u have time.

The point on "Converting Price to Value" mentioned a caution on that says "There are situations in which a market's natural time varies". Brings to mind the contrast of Pete Steidlmayer concept of Time variable profile which place time ahead of price instead by only considering a trade only if time is leading or when a profile measure across is less in tpo count then the previous. Here is the latest webinar from pete for anyone interested.

http://traderkingdom.com/market-prof...todays-markets

Yesterday EU was churning. With most reference point indicating increase activity & flow. With 5day overlay bracketing, pending trade was just done above the upper limit & close out near 3775. Now price is still going higher TESTING the 20day bracket upper limit or ard 3800. By US session the testing will be more meaningful.
http://i.imgur.com/3NeC1ye.gif

Hi,

Can anyone interpret this market profile indicator what does it mean as price is going into a lighter or darker area? Is the lighter area = most value (meaning most daily traded volume/transactions at that price point) and darker areas where price has been least lower value area?

Thanks.

Great trade!!!! Good to see you caught that. Unfortunately I woke up on the wrong side of that one. (Trading in the PST is a nightmare) Great chart & data analysis. If I wasn’t such a perfectionist I'd be finished with mine. Interesting video. I remember when we saw the earlier ones. I didn’t take much from it. As we already split the evolving days sessions, kind of served the same purpose.
Since it was the last Qtr. of the year at the beginning of the quarter I started to trade from a strictly “Analytical” approach I’m learning at Cisco. I’m experimenting with “run pause” patterns of tpo's, etc. So as my education evolves, so does my trading. I started a trade explorer at the beginning of the quarter using a trading account I rarely use, just to see if I can use some of this new info to improve on my knowledge of trading the Auction process. I pretty much take a bias for the day & stick with it. Anyway, it’s always great to see your post & analysis. Normally it just the sound of “crickets” around here. I’ll continue to post the weekly data. However, like this previous week if there’s not enough tick data the charts won’t generate. But please don’t be a stranger around here……

Thats a pretty old indicator, but from a basic sense traders are either going to be responsive or initiative above or below value. I think the lighter area is where the most volume traded, relative to whatever TF your using. If your trading the "indicator", then you have a 50/50 chance of knowing the probabilities of rejection or continuation. But if you're analyzing the "data" from the indicator, you can use the data to weigh the probabilities & "intent" of traders behavior. Unfortunately, that indicator imo doesnt provided the analytical data needed for "analysis". But the many traders who use it are just trading the "picture", they cant give you anymore info other than price may go up or it may go down. "Analysis" uses the data to give you the probabilities, and within your tolerance for risk, you trade according. But to make a long story short, you may need find the traders who trade the "graphic" to give you the best advice on how to trade it. Unfortunately, I trade the "data" from the graphic, NOT the the picture itself......sorry I couldn't be of more help

I just wanted to post a "snippet" of an article, for a little "food for thought"
The Current State of Financial Market Theory: Real Markets as it applies to AMT

CISCO Futures
1-303-306-1521 1-800 800 7227 Fax 1-303-306-1598
Internet http//www.cisco-futures.com
Email [email protected]

Auction Market Theory
Donald L. Jones
CISCO Futures
February 2009

The Current State of Financial Market Theory: Real Markets
The real markets people trade are far from the idealized CAPM. The non-stationarity and feedback aspects argue strongly for markets to be non-linear, at least at times of high risk. Markets are not efficient, rather they are effective. The standard deviation of the annual market measured on a market day basis not just risk. It also includes elements of opportunity. Curiously, little of this information was addressed in the foreword of the 2000 reissue of CAPM.

Theory as it applies to Auction Markets
Theories are often thought to be incomplete without the ability to make predictions. This is not so. At base, a theory is an explanation of a phenomenon. A thorough understanding of the phenomenon may well lead to the ability to predict at some level. However, the feedback nature of a complex market may preclude an ability to develop lasting predictions. For example, if the theory were found to predict tomorrow's price, everyone would soon learn and the feedback would adjust the market to make the prediction fail. This is obvious in futures markets which are essentially zero sum (in the absence of transaction costs).

Predictions that do not directly bear on profiting from trading are not disallowed. A study of the persistence of trends found that a mature market experiencing a new high price had an eighty percent probability of reaching a yet higher high within ten days (DJ1). This finding illustrated that the market had momentum under certain conditions, a feature not allowed in efficient markets.

The complex nature of the market is a strong argument for a phenomenological approach to market theory. A market has many facets, elements that are functional parts, but still with a large degree of independence in their behavior. As one of these elements changes, that changed behavior is returned to the market via the feedback mechanism. While elements such as volume, volatility, range, time behavior and the like are individually measureable, they all combine to form the market one studies or trades. Each element must be discovered and studied on its own before there is any thought of considering how it affects the whole. The vital question of extreme market behavior such as large, fast changes in value, must be understood. It is the thesis of Auction Market Value Theory that each of the market elements can be isolated and must be understood prior to forming a coherent description of a market. The experienced trader collects all the market information available, assesses each piece, ending with an understanding of the market situation. Not all pieces are necessarily needed to find the market's condition. Different sets will be used at different times.

As an analogy, think of a homicide detective. Experience has taught that there is a wide variety of clues. Arriving at a crime scene, the detective collects all clues available. Many clues are irrelevant and are later discarded. The pertinent clues lead to a solution of the crime. At the next crime scene, again all clues are collected. Again, the irrelevant clues are discarded, except this time it is a different set of clues that are irrelevant. If you were to suggest to the detective that to save time only the relative clues be collected, you would be laughed at: which clues are relative becomes known only through the discovery process of the crime investigation.

A crime scene is a complex situation in real life. The perpetrator seeks concealment, hiding or leaving misleading information. It is the detective's job to uncover the truth. The real world of the auction market is merely another a complex process with many clues of varying utility at various times.

Auction theory has the goal of understanding the current market situation. Is the market balancing, trending or behaving in some other way? Obviously if one can understand the current market condition it was possible to understand the condition at an earlier time. Comparing the current with the earlier condition allows one to see if the market has changed.

A combination of the relevent market elements will lead to the conclusion that the market is behaving essentially as it was at the prior determination or it has changed. No change means continuation. If the market was trending the trend continues. Or if the market was congesting, it still is.

Change or continuation provides the basis for market decision. For instance, a trader can find opportunity in change. A conclusion that the market is changing out of a balance condition may lead to a decision to trade. A change from moving (trend) into balance is an argument to exit from a trade. In the case of a portfolio update the old equities should be eliminated in a balance or change to downtrend situation. The new equities should be bought into the portfolio when they are in balance or trending up. If they are in a downtrend, it clearly makes sense to wait until the downward move is over before purchasing them.

To the analyst/trader it all boils down to a determination of change or continuation. Another way to state this is to couch it as a search for the unexpected. If a market has been in equilibrium, it is expected that it will continue in equilibrium. If a market is trending, it is expected to continue. Change is the unexpected. One way or the other, an analyst/trader must search for the unexpected. The expected, the market condition (e.g. balancing or in equilibrium) must be known. The only way to find a market's condition is through empirical measurements. Sharpe, in Ref 1, pg viii, decries empirical studies, stating they have a "short half-life". In the absence of an over-arching theory like the normal distribution of CAPM, complex market analysis must rely on empirical market measurements.

Continuation at Market Extremes, A Special Case
In a 'Letter to the Editor' of Commodities Magazine (later renamed Futures), February 1973, we reported that on the occasion of a new contract high the probability of a subsequent higher high was 82%, of two subsequent higher highs was 62% and three higher highs was 46%. The same pattern was true for new contract lows.

We view these findings as a contradiction of the Efficient Market Theory long held by economists. The practical value is that continuation concepts and practices may be useful to those interested in market analysis and/or trading.

What Drives Markets
Fundamentally, markets are the place where buyers and sellers meet to exchange goods. Early on the commodity grain markets were initiated as a place to lay off risk. In December, a baker who needed 5000 bushels of wheat in July could buy a contract at the exchange for delivery at that time. The seller contracted to deliver the wheat at the appointed price at the appointed time. The baker had price protection for the bread he had to deliver seven months hence. But market conditions may change over seven months. Enter the trader, the speculator, who attempted to take advantage of the changed situation by buying or selling. Value of the wheat would most likely change over time and so the market offered opportunity to the trader.

Just as the Behavioral Economists are reality driven, even more so is the market. Market prices vary. This offers an apparent opportunity to the trader who is not a user of the commodity, rather one who seeks to profit on price change. The precepts of Auction Market Value Theory provide ways to read and understand the market and so the intelligent speculator is armed in the way that any other professional is armed by knowledge of his/her field. An external element, not a part of the theory is the background business climate in which the specualtor acts.

From the institutional (brokerage, exchange member) standpoint auction markets are fee driven, not profit driven. Win or lose, the broker gets a fee. Investors buy and hold, minimizing fees. The speculator makes many trades and generates large fees. A discussion of what drives short timeframe trading in markets might be illuminated by stating that brokers' success in advertising for new speculators is a major driving force. A fund of new money continuously keeps markets lubricated. Other trading, e.g. by hedgers and, in equities, portfolio roll-over is a source of liquidity exclusive of speculators.

An institutional fee driven mindset is not unreasonable in view ot the statistics. An estimated ninety percent of speculators will disappear within the first year they are active. A solid broker-trader relationship is hard to develop when the customers keep leaving. If indeed the trading business is not speculator friendly, then one element for success and a defacto part of auction theory is for the speculator to be prepared. In other professions a sometimes extended training period is normal. Doctors, lawyers, engineers, architects all undergo years of study. It is an adjunct of auction theory that anyone trading will better their chances by understanding their career field.

Full Article can be found here: http://www.cisco-futures.com/Auction_Market_Theory.html

Its apparent as to which "school of thought" I belong to, but to be fair Ill post the the "Brief Review of Current Market Theories", You don't need to "agree" with the different theories, just "Understand" them.......its part of the full article posted in the previous post

CISCO Futures
1-303-306-1521 1-800 800 7227 Fax 1-303-306-1598
Internet http//www.cisco-futures.com
Email [email protected]

Auction Market Theory
Donald L. Jones
CISCO Futures
February 2009

Foreword

Financial Market Theory (Financial Economics)
Modern financial market theory began with the studies of Markowitz in the 1950's. Sharpe completed the work in 1970 with the hallmark book, Portfolio Theory and Capital Markets (PTCM). Capital Market Theory, CAPM, developed in PTCM, assumes equilibrium and a gaussian distribution of price range, with decisions being made by a 'rational investor'. CAPM defines market risk as the standard deviation of the distribution. A basic assumption is that investors consider only expected return and risk.

Behavioral Economics (BE), the psychology of decision making, shows that humans deviate from the standard economic model. 'Bounded Rationality' considers the limited cognitive abilities of investors; 'Bounded Willpower' shows that people may make choices not in their long-run interest and 'Bounded Self-Interest' involves investors sacrificing their own interests to help others.

Econophysics (EP) entered the scene in the mid 1990's. EP shows that markets are often not in equilibrium and that the gaussian distribution is unable to describe the rich texture of auction markets. EP views markets as complex systems, systems that are in continual change from the feedback of the market's agents (traders). Like CAPM, EP relies on the observables price and price change. EP shows that the standard deviation does not correctly identify risk. EP proposes to find a valid mathematical basis for describing markets.

A separate track in market understanding was initiated in 1985 when J. Peter Steidlmayer presented the Market Profile (MP) as a market informational tool (CBOT1). The aim of MP is to convey information on what is actually occurring as the market is trading. That knowledge gives the trader an opportunity to achieve superior results. MP differs from CAPM and the later EP concepts in that the primary variable is value, although the primary observables remain price and price change. A companion text (JPS/KK) explored the inner workings of auction markets.

Auction Market Theory (AMT) follows the value track of analysis. The market is accepted as a complex system. Value is the primary variable for describing the market. In place of an overarching distribution function, AMT examines the many component parts of a market (reference points). Each component describes an aspect of the market. The sum of the descriptors, collectively, describe the market as a whole. Thus, the goal of AMT is a clarification of the auction process, search and evaluation of the component parts of market. A premise of Auction Market Theory is that an overarching mathematical formulation of a market distribution is not possible with current market understanding and mathematical capabilities . In view of the complex nature of auction markets, a closed form solution may never be possible. The utility of AMT is in its potential for explaining how the market works. At bottom, a theory should be an aid or pathway for decision making. One would like to predict the future to some degree. In markets, the complexity precludes true forecasting. However, AMT provides information reliable enough that market decisions can be based on the AMT data.

A Brief Review of Current Market Theories

CAPM: An Equilibrium Theory of Financial Markets
Many commentators have criticized CAPM's assumption of equilibrium. The observable, daily high-low ranges, often appear to not be in equilibrium. Volatility is defined as the standard deviation of the price - time distribution and it is also defined as risk. This too, drew comment since opportunity is present and must somehow be related to price behavior. Actually, the very existence of a standard deviation requires a known distribution. The distribution assumed, gaussian, is demonstrably not describing the market much of the time, hence the standard deviation is devoid of meaning at those times when the market is not in equilibrium. Most certainly, if the standard deviation is not defined, the volatility cannot be either.

Ultimately, CAPM came to be understood as pertaining to well behaved (equilibrium) markets. But it could not explain drawdowns, crashes and other extreme behavior emanating from non-equilibrium activity. Furthermore, as the Behavioral Economists have discovered, the supposed 'rational investor' may often act in irrational ways.

The Normal distribution of the daily trading range is well understood, where the first moment is the mean and the second moment is the standard deviation. Analytical advances can be made with this distribution that would be impossible if the distribution were more complicated or not known at all. An example of such an offshoot is the Black Scholes model for option valuation. A contributing factor in accepting the normal distribution for markets was the then prevalent Efficient Market Theory. If the market is efficient, discounting all new information instantaneously, no one could find an edge for trading profitably.

Although there has been much negative comment on the equilibrium market proposition, the econophysicists have done the best job of showing why the stock market is not stochastic (JJH 34, 63-64).

Aside from the stochastic problem, PTCM contains a theory with valuable insights for portfolio management and the afore mentioned options valuation. By and large CAPM does not describe real markets (they are not stochastic) and option valuation founders on the definition of volatility. Although CAPM is useful in balancing portfolios, it offers little help on the mechanics of buying and selling. If the market were truly stochastic and efficient, it should make no difference when the new stocks are bought and the old sold. But the market has been shown to not be efficient; so, theoretically at least, timing, denied by efficient market theory, can possibly be profitable. In the absence of a stochastic process, i.e. when the actual market is not in equilibrium, proofs based on stochasticity are invalid.

The fundamental problem of defining annualized volatility as the standard deviation of the annualized market has led options traders to seek a more current volatility. Implied volatility is backed out from the actual market pricing of options. A problem with implied volatility is that the iterative mathematical procedure involved is captive to market noise, particularly since price is the primary variable. Further, if the market's distribution is not gaussian, as is shown by the Econophysists, higher moments are significant in evaluating risk, i.e. the simple case cannot find dangers present in the real world. Implied volatility is reputedly a significant culprit in the spectacular failure of the trading firm Long Term Capital Markets ($1.3 trillion loss) (RL).

While it is now accepted that the stock market is demonstrably not a stochastic process and volatility is demonstrably not just risk, the real difficulty with CAPM is that markets are a complex process and CAPM is based on the exceedingly simple gaussian distribution and rational investors. At times the market is well behaved and the gaussian model is adequate. At other times, when a gaussian is not appropriate, the dangers may be extreme. One has a devils choice: use a simplified picture (equilibrium) and be able to analyze the market (knowing there will be times when the analysis is wrong) or admit the market is complex and be able to analyze nothing within the distribution format. The decision is not quite so draconian today. The field of Econophysics, addressing markets as complex systems, is making advances toward a market analysis that includes elements of reality, such as crashes. Further, the reality approach of Auction Market Theory offers a manner of dealing with complex markets based on an understanding of the principal elements of markets and how their reaction to feedback can be understood.

In spite of the substantial and known drawbacks with CAPM, the theory continues to be taught in financial MBA programs. This is probably a result of there being no logical successor theory and CAPM is teachable. By "teachable" we mean:
1. It is coherent and complete
2. Mathematics are sophisticated enough to make MBA level students
work on the mechanics.
3. Tests can be made exact and thus easily gradeable.
4. The general ideas can give students a grasp of capital markets.
5. MBA students who are not planning to be traders (the majority).
thus have enough information to get by.
6. Portfolio balancing/efficient frontier and other portfolio
concepts are impressive and useful. 7. Mathematically oriented students can manipulate CAPM to gain
new uses, e.g. Black Scholes option pricing

Econophysics is attempting to understand risk, to explain extreme events like crashes and basically to explain why markets behave as they do. Like CAPM, EP takes as observables, price and return (price change).

Modern Market Analysis, Complex Systems and Econophysics
Recent work by Johnson, Jeffries and Hui (JJH) in the emerging field of Econophysics treats the financial markets as complex systems. They consider the problems of:
1) Markets as complicated dynamical systems that are continually
generating high-frequency data series.
2) How the stochastic assumption gives misleading answers to practical
problems such as minimizing risk, explaining extreme events such as
drawdowns and crashes and pricing derivatives.
3) Why financial markets behave as they do.
4) What can be done to minimize risk.

JJH on page 2, lists as goals, these practical questions:
1) When to buy.
2) When to sell.
3) Risk.
4) Predictability.
5) Crowd behavior.
6) Forecasting on basis of crowd behavior.
7) Forecasting time evolution of markets.

Complex systems concepts have only recently begun to be applied to financial market analysis. Complex systems are generally nonlinear with feedback acting to continually adjust the system. There appears to be no hard and fast definition of a complex system.

Peters in 1999 (EEP) listed these characteristics of a complex financial system:
1) The system has a purpose (e.g. to facilitate trading of say, soybean futures, stock market indexes, etc.)
2) The system is decentralized (many independent agents/traders)
3) Feedback occurs within the system (all agents observe the system and make changes in their behavior)
4) The system adapts to information from feedback (losers get out, winners increase holdings)
5) Adaptations are decentralized leading to innovation (each agent makes its own decisions)
6) Rules govern the system, rules can change or be changed (adapt) e.g. the movement from exchange floor to computer trading

Observations 1 - 6 describe auction markets such as stocks, interest rates, futures, derivatives and actually even markets as diverse as food and department stores (JPS/KK). The financial markets are double sided auctions where a buyer at one moment may become a seller the next moment.

Items 2 - 5 show that feedback is diverse, affecting each agent or trader in a unique, personal way. The net effect of the feedback manifests itself in macrosocpic market parameters such as price movement and volume. But it is impossible and not even desirable to isolate a particular feedback element from any one of the agents; the useful observable is the net market change related to that element.

JJH defines complexity in financial markets somewhat differently from Peters.
1) Feedback: change contains an element of remembering.
2) Non-stationarity: the statistical distribution changes.
3) Many agents: traders, institutions interact in time-dependent ways.
4) Adaptation: agents adapt their behavior to improve their chances.
5) Evolution: agents behavior evolves thru feedback and adaptation and the system may not be in equilibrium. It can exhibit extreme behavior such as crashes.
6) Single realization: i.e. averages over time are not equal to averages over ensembles.
7) The market is an open system coupled to the environment: one cannot discriminate between exogenous (outside) and endogenous (inside) influences.

JJH's analysis turns on measurement of price and return (JJH 16).
Briefly, they assume:
1) Price as a function of time is the primary observable.
2) The assumptions of CAPM that price changes are independent and identically distributed are not borne out by observation.
3) CAPM volatility is not sufficient to classify risk.
4) A market price series' expected value depends in part on previous movement, the market has some memory. This helps explain drawdowns and crashes.
5) Crowd action plays a role in volatility.
6) The 'zero risk' in writing options under CAPM assumptions is not true in real markets.
7) Real price series differ from the random-walk model.
8) Trader's beliefs/actions can create patterns sometimes leading to crashes.
9) Markets are non-linear.
10) Internally and at all times, price moves much more and faster than rational expectations (of return) would predict.

The point is that both econophysists and Peters see the markets as complex systems with feedback causing change dynamically. The econophysicists are more mathematically oriented and propose a stricter analysis. They also propose a mode of attack, one using distribution functions. At this point it is not clear how a general market distribution function might be developed. Looking at the complex market from a differential equation standpoint, a descriptive equation would likely be nonlinear (feedback probably does not behave linearly), non-homogenous (variables most likely could not be separated, even if they could be defined) and the coefficients are non constant (an increase of ten percent in the soybean crop would rarely translate into a change in price of ten percent). It would be most unlikely that a complicated differential equation could be translated into a distribution function. One might think of a complex market in terms of the parable of the blind men and the elephant. An analyst explores all the various parts, ending up with enough pieces to make a slow moving elephant. However, with a complex market, the feedback can drive a change so that the next time the various parts end up making a fast moving tiger.

engaging and riveting contribution as usual from you going towards this line of work and as always thank you MzVega