Friday, October 27, 2006

Evidence Based Trading: Why Philosophy Matters

The late Ayn Rand emphasized that philosophy was the most practical of disciplines: it governs the ideas that lie behind all we do and think. The philosophical premises we assume affect how we approach trading.

A beautiful example of this is David Aronson's new book, "Evidence-Based Technical Analysis". It's a well-written, thought-provoking text, with many practical examples of how to conduct data analysis in an objective way.

Starting with the premise that knowledge consists of statements that are found to be true, Aronson, writing in the positivist tradition of philosophy, excludes subjectivity as knowledge. He explains:

"The most important consequence of TA adopting the scientific method would be the elimination of subjective approaches. Because they are not testable, subjective methods are shielded from empirical challenge. This makes them worse than wrong. They are meaningless propositions devoid of information. Their elimination would make TA an entirely objective practice." p. 148

This is bound to rub many traders the wrong way, but it's an important challenge. What is knowledge? How do we know what we know in the markets? How can we demonstrate that knowledge is such, and not illusion?

Once we start with the premise that all knowledge consists of explicit propositions that can be tested for truth, we necessarily are led toward trading that is rule-based and rigorously backtested.

Is there another, *valid* form of knowledge and trading? Can we prove that? I'll be interested in readers' comments before I offer my own alternative in my next post.


Krasimir said...

Every consistent trader strives to trade as objectievely as he could and this is one of the main challenges in trading. If one falls prey of his own subjectivity soon or later he will fail at the markets.

What is knowledge? What is knowledge for one may not be knowledge for other. Knowledge is what is working for you, what is your truth and how you percieve reality. Market knowledge is how you percieve market. If you are a successful trader your knowledge about the market is that the market is friendly place. If you are losing trader your knowledge about the market would be that the market is a battlefield and you are victim. Consequently, what is knowledge for you may be illusion for others and vice versa.


John Wheatcroft said...

As a former systems engineer and relatively new trader it occurred to me that the reason I was a successful engineer is that I invented processes that worked and then followed the process. The only thing I had to manage then was risk. I reasoned that the same approach would have to work in the trading arena and have begun doing just that - inventing processes that work and managing risk - and I am becomming a successful trader -

Thus I would have to agree with the premise - truth is testable but more importantly - repeatable.

Carl said...

"This is bound to rub many traders the wrong way, but it's an important challenge."

Wow. The two beasts you just 'rubbed wrong' are the two 4-ton elephants asleep on each of my shoulders. They usually don't wake up and pester me until around 9:30 AM (when the equity markets open). I call them fear and greed.

Without them (no matter how hard I try to control, monitor, and objectify them along with the data my brain takes in) I don't participate in the market, and no one else does either for that matter... in my opinion.

"What is knowledge? How do we know what we know in the markets? How can we demonstrate that knowledge is such, and not illusion?"

We might accurately characterize knowledge as a collection of sensory data that our brains perceive and store. And yes... absolutely... our state of mind (what we desire to receive from the world and what we fear to receive from the world) biases the data we perceive and retain as knowledge. Robert Anton Wilson's books, Prometheus Rising and Quantum Psychology, do an oustanding job of exploring these issues. These books have given me tools that help me more accurately evaluate what I perceive as a trading opportunity.

"Once we start with the premise that all knowledge consists of explicit propositions that can be tested for truth, we necessarily are led toward trading that is rule-based and rigorously backtested"

Sure... but the *rules* we build confidence in will begin to compete with one another won't they? For example, observe the current market. With a majority of technical indicators demonstrating a massively overbought market do I sell short at the first sign of weakness? Oh no, wait! "The trend is my friend". The markets are on a tear right now so I'll go all in for more up at the next sign of weakness.

Oh look! It's our friends fear and greed again! =;-)

"Is there another, *valid* form of knowledge and trading? Can we prove that?"

Great question. If the markets were governed and could be traded by rules *only* then I know of several talented computer programmers who would be retired by now. Instead the markets are ruled by humans. Some looking at tecnical analysis, some at fundies, some at chicken entrails or the phase of the moon. That mix makes a market.

Valid is a *loaded* word ;-) Please qualify and quantify it in you answer.

Does valid mean an approach is repeatable, consistent, profitable? How much work do I have to put in?

It's 1963, my pal Warren Buffet whom I like because he is folksy and enjoys cheesburgers asks that I invest $10,000 in his investment interests. I give him $10,000 and never look at the markets. Would that have been a valid investment?

You probably see where I am coming from. I'm not convinced valid is a safe word in the domain of the markets. There is ALWAYS a counterexample, an *invalidation* of any approach. In the end there is the right edge of the chart, and a decesion on whether or not to participate. The equity retained, and built, in our accounts might be the validation of success in this domain. Are you going to tell us that our attitude, our intentions, will manfifest our results? Perhaps so, I hope so, because I desire to build equity in my accounts!

Can't wait for your answer! Great post!

yinTrader said...

Hi Brett

'Atlas Shrugged' is often seen as Rand's most extensive statement of Objectivism :

"My philosophy, in essence, is the concept of man as a heroic being, with his own happiness as the moral purpose of his life, with productive achievement as his noblest activity, and reason as his only absolute."

How does one go about testing philosophical premises?

Fabio Santoni said...

I see a problem when you try to combine your propositions. Well, yesterday "the VIX was low" so, today, I prefer the long side of the trade, but yesterday was a day with "few stocks making 20 days high" and I would prefer the short side. I don't see how resolve this kind of problem, because I cannot apply a decent statistical method if the sample numerosity is small: if I can find a good sample of days with either "the VIX was low" OR with "few stocks making 20 days high", it's easy to find very few days with "the VIX was low" AND with "few stocks making 20 days high".

John Wheatcroft said...

Doc, I came up with another one - something I just read today - If you buy every correction in a bull market, you'll only be wrong once--the top. If you sell every rally in a bear market, you'll only be wrong once--the bottom.

As it happens those two statements are both testable and repeatable and are the absolute truth. They are also a set of "rules". They can (and are) programmed into machines for machine based trading. They form the basis of many trader's processes (sometimes unconsiously). They are one thing that is "known" about all markets - i.e. to be a successful trader buy low and sell high or buy high and sell higher - it is the same thing.

Jim0soYoung said...

You should get a sales commission from Mr. Aronson because I just bought a copy of his new book! Thanks for the tip - I'm a very strong believer in statistical methods and backtesting.

Brett Steenbarger, Ph.D. said...

Hello to all,

I want to thank everyone for their comments thus far. The questions of what is market/trading knowledge, how do we develop it, and how do we measure it are of immense importance and yet are rarely dealt with in a thoughtful manner. Aronson's book is unique in that it places these issues front and center.

I'll have much more to say about these matters in tomorrow's post. Suffice it to say at this point--echoing many of the comments--that there are many things we know subjectively that we can't necessarily place into words that form testable propositions.

In his book "Tacit Knowledge", Michael Polanyi used the example of facial recognition. I *know* what my wife's face looks like, but could not describe it to you in a series of sentences that would provide for anything except random identification.

Similarly, when Shakespeare (or any fine poet/writer) states something about the human condition in a metaphor or image, this is not a testable proposition, but it does represent a kind of understanding--a form of knowledge.

Intuition can be "into wishin'", as the saying goes, but a mother's subjective knowledge about her baby is often more than a random guess.

The big question is whether there can be an objective basis for subjective decision making. How we proceed to learn markets hinges on our answer to that question.


The Market Speculator said...

I am a firm believer in making use of intuition that is *based on experience*. For instance, I, like many traders, have experienced a few instances where I knew the market was at a turning point, but I couldn't quantify it. My subconscious was probably was going though some sort of pattern recognizing process that I could not quantify. Maybe it was organizing a large set of data that I had observed, such as the hundreds of charts I go through on a daily basis, the readings, televison shows, conversations, and matching it to a previous instances without my really knowing it. Is this something that could ever be tested objectively? I doubt it.

I keep a list of all of my intuitive (trades that aren't based on setups I use) versus more objective trades (patterns and setups that I have tested). The win rate and percentage gain is actually higher on the intuitive side, although the sample is very small. While I can't learn much from these subjective trades, my results tell me something positive is happening. Aronson is correct when he says "they are meaningless propositions devoid of information" in terms of understanding exactly what process caused the results. However, I do know that when I do get that feeling, I should go with it. Those meaningless propositions may be useless to the academic in me, but they can do wonders for the portfolio.

A little off topic, but probably of interest here, the rights to Atlas Shrugged were purchased by Lions Gate and it's rumored that Angelina Jolie and Brad Pitt will play Taggart and Galt. I hope they can do justice to the book. It will be a tough one to adapt to film.

Brett Steenbarger, Ph.D. said...

Hi Market Speculator,

Thanks for the note; I've enjoyed your blog. Thanks also for the update on the Atlas Shrugged movie, which has been long contemplated.

I think you're on the right track in trusting your intuition, but also keeping score on those intuitive trades to verify that your feel for the market is rooted in something real. There are many things we know without being able to verbalize them, including most performance skills. That doesn't mean we can't objectively measure their outcomes.


Brandon Wilhite said...

Dr. Steenbarger,

You've just made my day with this posting!! Epistemology was my particular interest in graduate school. Positivism has significant problems, as you seem to know (I like Polanyi btw). In my opinion Kurt Goedel's work pretty much destroys positivism, not to mention many other philosophic and also scientific discoveries.

A classical definition of knowledge is "justified true belief," and we could talk for hours about counter-examples, but I'll just give one. Trader Joe beleives that XYZ is going to go up, because of reason Q (which is valid reasoning), and so he goes long. Much to his delight XYZ takes off and he makes a ton of money. However, Trader Joe later finds out that reason Q did not exist during his win streak, but actually it was another reason, reason Z that caused XYZ to skyrocket (or maybe he can't figure out the reason). So did he have knowledge? Strictly speaking, it did not fit the definition, and so if that is your definition you must say no.

The biggest stumbling block I see before people when it comes to knowledge is the need for certainty. I believe that we can have knowledge without apodictic certainity and we could give many examples of such knowledge like the one you give in your post. In my view successful traders need to overcome this stumbling block.


Brandon Wilhite said...

Just wanted to add that I am very systematic and rigorous in my trading. Most people can't stomach my it's not rigor or scientific method I disagree with, but I think that the grounds from which we start should be solid, hence my previous post.

Brett Steenbarger, Ph.D. said...


Thanks for your insightful comments. I think you're onto something important: how the need for certainty is very different from the need for knowledge. Much of what we know in markets (and life) is probabilistic. Testing systems reduces, but does not eliminate, uncertainty. Ultimately, however, risk adjusted P/L tells us how successful all "systems" are, whether they're subjective traders or objective algorithms.