Some food for thought: Eckhardt

William Eckhardt; infamous for his bet with Richard Dennis and the original Turtles.

Read more about him here, and the turtle story here. To see his performance, check his amazing equity curve here.

Here's a piece that he presented back in '96. Many of the ideas put forth, especially those applying to his methods and systems, are worth exploring.


June 29, 1996
Ritz Carlton Hotel

Address by William Eckhardt
President, Eckhardt Trading Company

At Eckhardt Trading Company ("ETC") we try to take a scientific approach to trading. This is not an easy description to live up to. We try to earn it by paying a lot of attention to the foundations of the subject, to the soundness of our methodology, and to the correctness of our statistics.

In terms of the foundations of the subject, we rely heavily on Decision Theory and Utility Theory. To see the usefulness of this, first note that there are two respects in which profits and losses are not equivalent. One is objective and has to do with nonlinearity. For example, it requires a 100% profit to balance a 50% loss. The second is subjective and has to do with risk aversion, for many people even the prospect of a 150% profit does not compensate for the risk of a 50% loss. Through Utility Theory, such imbalances can be treated in a rigorous, quantitative manner and in this way uniform and unified procedures can be developed. Look at the question of risk management. Any trader who survives any length of time knows something about his subject, but in my experience, traders simply graft risk control on top of whatever else they are doing, often in an arbitrary way. For instance, many prospective clients have asked me what’s the most I’ll lose on one trade. I can look up these statistics, but this is not something I would ordinarily pay any attention to. It doesn’t matter how little you lose on an individual trade, but how much you might lose on your whole portfolio. You’re not going to keep a ship afloat just by making sure the leaks are small. The important thing is to limit portfolio risk, the trades will take care of themselves.

We have devised a portfolio theory quite different from the classical theory that permits factors such as risk aversion, the nonlinear imbalances between profits and drawdowns, and long-term utility growth to be built in at the ground floor. They are all part of the formulas that define what it means for a system to be good. In this way, on even the most preliminary test run of a new idea we are forced to take into consideration the subtle and complex relations between drawdowns and long-term growth. At ETC we are dedicated utility maximizers and pay particular attention to the rate of expected utility growth.

It has been shown again and again, that without proper controls, even the most honest researcher will unconsciously bias research usually in a favorable direction. Trading systems research is especially rife with possibilities for this kind of wish fulfillment. During more than 20 years, we have seen an amazing variety of ways in which research can mislead or falsify. In response to this we have developed a veritable gauntlet of tests that any system must pass to be taken seriously. We test for post-dictiveness, for computer glitches, and for statistical artifacts. We test for overfitting, for maldistribution of returns, and the degree to which a system takes advantage of unusual and possibly nonrepeatable circumstances. Theses are just a few of the potential sources of trouble that we routinely monitor. This battery of tests can bring runaway enthusiasms back down to earth.

An important feature of our approach is that we work almost exclusively with price, past and current. One reason for this is that to make any progress in the early stages of quantitative investigation you usually have to reduce the relevant factors to one or two crucial variables. Price is definitely the variable traders live and die by, so it is the obvious candidate for investigation. The other reason is that in a system that’s making good use of price information, it is very difficult to add other information without degradation. Pure price systems are close enough to the North Pole that any departure tends to bring you farther south.

Many systematic traders spend the majority of their time searching for good places to initiate. It just seems to be part of human nature to focus on the most hopeful point of the trading cycle. Our research indicated that liquidations are vastly more important than initiations. If you initiate purely randomly, you do surprisingly well with a good liquidation criterion. In contract, random liquidations will kill the best system. At ETC we expend a lot of our research effort on liquidations.

Most standard statistical techniques are inappropriate for analyzing trading. Statisticians have developed many delicate techniques that squeeze information from minimal data, but these give false results in this business. I tell traders that if the results don’t sock you in the eye, they’re probably not real. Accordingly, we use only the most robust and assumption free statistical tests, We have an aversion to summary statistics that obliterate important structural elements. For assessing systems, we use a technique called bootstrapping so that the complete distribution of past outcomes can make itself felt in decisions; the distribution is not simply viewed in terms of its mean and variance which can give a distorted picture.

Our aversion to summary statistics that obliterate structure extends to the trading systems themselves. For instance, we avoid moving averages of price in making trades. Such moving averages are popular mostly because they’re mathematically tractable, but they smooth away all the structural information inherent in the price data.

Another popular tool, the price breakout, may be far better than the moving average, but it still eliminates most of the relevant structure. A breakout trader keeps two pieces of structural information, the high and the low for a given time period, but ignores all the price structure in between. For this and for other reasons we judiciously avoid breakout trading in all parts of all our systems.

It’s a lot easier to look scientific than to be scientific. We try to avoid the kind of delicate fine tuning that gives on the feeling of being very accurate, but that is in fact mostly arbitrary. We have taken to heart the research that shows that simple yes-no schemes, either fully accept or fully reject something, are more useful and more robust tan delicate weighting schemes. For instance, we do not favor trades according to how good they are supposed to be, instead we use the following rule: if a trade is good enough to make, it’s good enough to make at full size; if a trade isn’t good enough to make at full size, then don’t make it at all. We adhere to this kind of reasoning all the way down the line. All five systems we currently use are given equal weight. We also try to give equal weight to each of the fifty or so markets we trade.

I would characterize our overall approach as "conservative". This does not mean that we avoid market risk, for market risk is the raw material from which profit is fashioned, but we are conservative about what we know and about what can be done. My experience with Decision Theory indicates that knowing what it is you are ignorant of is in fact a powerful position to be in. The task of the trader is to locate those few areas where ignorance is not complete and to convert this information into profitability in an efficient way. False knowledge can be very detrimental to this process, but acknowledged ignorance can be quite beneficial.

EDIT:
Adding a link to another older thread about Eckhardt's discussion on randomness.
 
Quote from tireg:

William Eckhardt; [/url]. To see his performance, check his amazing equity curve here.

Interesting stats. Looks like the huge returns dissappeared after the bubble burst and his last 5 years average return was about 9% per year....
 
OK let's see what happens when I apply a moving average crossover method to Coca Cola stock, symbol KO, start 20 March 1967, end 18 July 2006, and follow these rules:

1) When the value of the 100 day exponential moving average value becomes greater than the value of the 200 day exponential moving average value then buy.

2) When the value of the 100 day exponential moving average value becomes less than the value of the 200 day exponential moving average value then sell.

3) Position size is 5 percent of account equity divided by the average true range.

Number of trades 8
Total profit $ 1751566
Profit after subtracting $ 10.00 commission & slippage per transaction: $ 1751406
Heat is 5 percent.
Drawdown is 0.1321 (13.21 per cent).
Cumulative Annual Growth Rate (CAGR) is 44.64 per cent.
CAGR / Drawdown is 3.38
Instanteously Compounding Annual Growth Rate (ICAGR) is 7.44 per cent.
Annually Compounding Annual Growth Rate (ACAGR) is 7.72 per cent.
Information Ratio is 0.32
Initial capital is $ 100000
Slippage and commission per transaction is $ 10.00
Uses 20 period average true range to calculate position size
Long trades only
Growth rates are calculated after subtracting commission & slippage.

===

This method shows about 44 percent growth per year and the greatest drawdown is about 13 percent.
 
I guess I'll begin this exploration into Eckhardt's approach.

Eckhardt's emphasis of having to be 'right' or 'wrong', and the idea of FULLY rejecting or accepting an idea

This possibly stems from his academic background, as this approach strikes me as the way they approach things in schools. Either you're right or you're wrong - there's no in between. The fallacy of this is, of course, in real life nothing is ever as simple as black and white; there are always 'grey' areas, especially in the area of finance and financial statistics.

Eckhardt's application of this idea is the rejection of dynamic position sizing - he claims an 'all or nothing' approach, going full size on a trade or not at all, and gives equal weighting of the 5 systems he uses and even the markets within for purposes of portfolio allocation.

While on the surface this appears to be a good idea and is simple to impliment, many studies have shown that dynamic position sizing can be advantageous and is an effective way of money management. A simple approach to this would be either by reducing exposure to systems that are losing or markets that are unfavorable to the strategy, or to give more weighting to the systems that are doing well. Advanced concepts of dynamic position sizing include use of optimization of equity curve, optimal-F studies, and self-tuning or dynamic system weighting.

Decision Theory and Utility Theory: combining concepts of risk aversion and objectivity

This approach is quite interesting, and I find that much of it makes sense. The idea of integrating the concept of risk aversion into a utility curve means that he prefers the farther left side of that curve, opting for lower drawdowns at the expense of opportunity lost of potential higher returns.

His idea of objectivity in research is something all systems developers face at some point; how can we trust the results of our tests and simulations? His objective testing tests for: post dictiveness, computer glitches, statistical artifacts, overfitting, maldistribution of returns, and nonrepeatable circumstances. Along with his rejection of summary statistics, all of these are great ways to ward against making the wrong type of assumptions. Once you do enough testing you often start to see what market conditions the system does well in and where its weaknesses are.

Trading Systems and ideas
Once again, unfortunately, Eckhardt's idea that you have to completely accept or reject a concept forces him to make choices about his systems he chooses to use. He's decided on pure-price systems, rejecting other types of systems (possibly volume-based, sentiment-based, and fundamental-based), saying anything else tends to stray away. As an extension of the rejection of summary statistics, he rejects moving average type systems, as well as price-breakout systems. I find it interesting that these two are the most commonly used systems for his style, which is long-term trend following.

From his emphasis on 'price structure', two examples of which are 'high' and 'low', and the statements made about various trading systems, we can guess that his approach relies on chart patterns, possibly intra-day.

Not saying that one is better than another, but I think that fully rejecting whole classes of systems, particularly when many have been effective, may lead him to miss some potential good ideas. There has been some work done that shows interesting results on trading correlated systems with split dynamic position sizing.

Eckhardt also places much emphasis on 'liquidations', by which I assume he means exits. He claims that a random entry system can be profitable by good exits, but systems can not be made profitable by random exits. This has similarities to Van Tharp's study, and I would tend to agree that exits are often overlooked.
 
People who know me personally on ET will be quick to note why I've responded to this thread. Something to do with blood running thicker than water.

For starters, Bill Eckhardt has earned well over 9 figures trading. Most of that income was generated in his own personal trading account rather than as a fee whore. He was already VERY wealthy before he ever started trading a public fund.

Secondly, Bill's intelligence prohibits him from being fooled by randomness. It doesn't matter if a system is "profitable." It matters as to the why. Eckhardt is pointing out how the variables inherent in most systems produce results that are, well.....variable. In other words if you tell me an MA crossover of 8, 21 has had tremendous results I'll counter by saying at some point the market will go just high or low enough for a signal to be produced and that signal will then be eminently fadable. Likewise if you come back to me with a new set of variables it still won't matter. The cause/effect is tenuous and to build a logical statement advocating such a system is impossible. How can one argue with any statistical certainty that if the 8 day average exceeds the closing 21 day average the market should perform X? If one had an infinite number of samples for these types of systems (every conceivable market or asset class in every society throughout civilization ect.) you'd see the results are purely random. Some guy posted a KO summary on this thread. LMFAO! The Dow is ALSO up around 2000% during your sample period. So what's your point. ANY trend following system would have gotten you long KO. Did that make the system valid or merely the play valid? Hopefully you all understand the difference. The question Eckhardt poses is philosophical in nature. Is rightness quantifiable or rather expressed by a measurable input of price action?
 
Quote from Pabst:

People who know me personally on ET will be quick to note why I've responded to this thread. Something to do with blood running thicker than water.

For starters, Bill Eckhardt has earned well over 9 figures trading. Most of that income was generated in his own personal trading account rather than as a fee whore. He was already VERY wealthy before he ever started trading a public fund.

Secondly, Bill's intelligence prohibits him from being fooled by randomness. It doesn't matter if a system is "profitable." It matters as to the why. Eckhardt is pointing out how the variables inherent in most systems produce results that are, well.....variable. In other words if you tell me an MA crossover of 8, 21 has had tremendous results I'll counter by saying at some point the market will go just high or low enough for a signal to be produced and that signal will then be eminently fadable. Likewise if you come back to me with a new set of variables it still won't matter. The cause/effect is tenuous and to build a logical statement advocating such a system is impossible. How can one argue with any statistical certainty that if the 8 day average exceeds the closing 21 day average the market should perform X? If one had an infinite number of samples for these types of systems (every conceivable market or asset class in every society throughout civilization ect.) you'd see the results are purely random. Some guy posted a KO summary on this thread. LMFAO! The Dow is ALSO up around 2000% during your sample period. So what's your point. ANY trend following system would have gotten you long KO. Did that make the system valid or merely the play valid? Hopefully you all understand the difference. The question Eckhardt poses is philosophical in nature. Is rightness quantifiable or rather expressed by a measurable input of price action?

Thanks for the input, Pabst.

I did not mean to criticize Eckhardt as a person or his accomplishments; on the contrary, he is one of my rolemodels and much of my current work follows along similar lines as his previous work. At the same time, I find differences in perception and opinion about the markets. This is natural, and I wanted to explore some of these differences.

The example you give illustrates the point but in an indirect manner. Moving-average crossover systems are trend-following in nature, and their edge lies in being able to take advantage of persistency in price action; namely, that the trend has a higher probability of continuing in its direction once established than not. As you said, you argue that this system could be faded, which I will take a step further and even go as far as to say that the win rate of a counter-trend system using these rules would be higher than the win rate of such trend following system. Yet the system as you described it contains only entry rules, and not exit. As Eckhardt wrote, liquidation plays a key role. We could both take trades, with me on the side of the trend and you counter-trend, and we could both make money, given that our exit strategies limit our losses and achieve gains enough to overcome 0 expectancy. While I might be stopped out more, and you make money each time I am stopped out, at some point a trend will start (price persistency) and that is when the trend following system will make money.

To get to your question,
How can one argue with any statistical certainty that if the 8 day average exceeds the closing 21 day average the market should perform X?

Note that these systems as we've described them contain no inherent Edge. They are a product of market conditions. Statistical certainty comes when I apply either a bootstrapping or monte-carlo technique to the results of the system, and over, say, 10,000 trade-scramble and equity curve scramble runs, I get possible alternative outcomes for the system. From these outcomes, we look at the distribution of returns and make conclusions from it. Currently, I choose to account for the fat tails by using Chebyshev's inequality, so any statements about certainty would take into account larger than normal sigma events.

I think the point of your question, though, was 'How do we know that this system would perform better than random entry (since you only gave entry rules, we'll assume exit rules are the same for a random test and the system test)?'

In such a case, we test the system against, say, 10,000 runs of a random entry system. From this we get a new distribution of possible returns in the same market condition, and then we can compare the two results to see how much better than random the initial system did. Note said test was a test of entry-strategy only, since exit strategy was the same. A more accurate test would also take into account holding times of trades.

Which brings me to the conclusion drawn from above poster about KO. As we all know, for the last few decades we have experienced one of the greatest and most persistent bull market conditions ever seen. Common sense would dictate that in a bull market, buying strategies are the natural and proper (dare I say, 'right') entry strategies, since they take advantage of the better-than-random probability that the trend will persist. Can a counter-trend or short-only strategy make money in a bull market? Of course. Prices don't go up steadily forever, and there will be the inevitable pullbacks. This is an example of the 'gray' area - I would be curious to what Eckhard would think is the 'RIGHT' strategy to use in a bull market.

Giving ONE isolated example of a specifically chosen stock that went up in a bull market, along with claims that the system that used to buy it made money is ridiculous.

Finally we come to the philosophical questions. These are some excellent questions, Pabst, and I would be curious to hear your thoughts as well as others' thoughts on:

Did that make the system valid or merely the play valid? Hopefully you all understand the difference. The question Eckhardt poses is philosophical in nature. Is rightness quantifiable or rather expressed by a measurable input of price action?

In regards to the first question - is such a buying system in a bull market condition, or any system in any given market condition, that makes money, valid? If the system is making money, do we care that 80% of that is b/c of market conditions and 20% b/c of the system's edge? Or how about if it's 100% market condition, trading an edgeless system that's 'RIGHT' for the market? Before, I thought no, such a system couldn't be valid because it had no edge. But practical experience has shown that edgeless systems can still make money due to market conditions. That's what we're after, right?

The key point is knowing WHY a system is making money. By knowing WHY, we won't end up like those other millions of day traders that thought they could make a living trading in the late 90's. As such, when market conditions change, we know to employ the proper systems, or to put the proper weighting on the systems that are appropriate for such condition.

As far as the second question, I don't think I follow. How would one quantify 'rightness'? Well, if you put a gun to my head, I would say that rightness is quantified by the performance of the trading system. I was right in this market, therefore I made money. But such statement and idea of bringing 'rightness' in reeks of hubris and is something I would like to avoid. I am not sure what you mean by the phrase 'Measurable input of price action'.
 
over time, i have given some thought to the random entry / profitable exit argument and i don't buy it - for exit to have an edge, there must be a process somewhere between entry and exit that produces that edge - in which case, there must be an entry more optimal than random entry - if a more optimal entry can not be executed due to some condition, such as lack of liquidity, then the exit edge is not a true edge, since the same condition can go against your entry.... - in fact, if there is something like an edge from exit, then the way to take advantage of it is not by random entry, but by countering it, i.e. fade the exit point by entering at it and exiting on the other side.
 
Quote from fader:

over time, i have given some thought to the random entry / profitable exit argument and i don't buy it - for exit to have an edge, there must be a process somewhere between entry and exit that produces that edge - in which case, there must be an entry more optimal than random entry - if a more optimal entry can not be executed due to some condition, such as lack of liquidity, then the exit edge is not a true edge, since the same condition can go against your entry.... - in fact, if there is something like an edge from exit, then the way to take advantage of it is not by random entry, but by countering it, i.e. fade the exit point by entering at it and exiting on the other side.

The way I see it, each trade has a loop:
Entry Setup, Entry, Exit Setup, Exit. They work in conjunction with each other to produce an outcome. Putting more weight to one side, either entry OR exit, makes little sense as PnL = (exit price - entry price)*shares. That's where position sizing comes into play. I think the point here Eckhardt was trying to make was that he puts a lot of emphasis on his exit strategies, as one should; many people focus too much on entries and not enough on exits. I am of the belief that for a given entry strategy, there exists an optimal complimentary exit strategy.

The random entry argument I don't entirely buy as well, because the trades that ARE profitable in random entry might be a result of sheer random outcome.
 
don't know if this fit's in the discussion of entry vs exit but I've worked with some serious perma bulls in the past. These guys are convinced US equity mkts only go up and trade accordingly. I'd say 95% of all opens they were long a basket of Naz, NYSE stocks along with the indices. If the mkts opened lower they were oversold and they bought, if they opened higher then it was the start of something big and they bought..all trades were opened and closed in the same day.

Funny thing, 75% percent of the time they made money and there good days far outsripped the bad. They knew to take profits after a huge spu move occured in their favor, and would add to their positions after a quick downturn, always taking advantage of any retracement to exit.

By my definition simply blindly going long every open is pretty close to random, yet these guys position mgmt after the open seemed to make all the difference and they had the numbers to prove it. And trust me it was maddening to watch. :)
 
Back
Top