Building Better Models – Framing the Hypothesis

As a price follower when you analyse your trade statistics and performance metrics you quickly realise that your success is chiefly determined by the ‘outlier’. In fact, you can strip out the vast majority of your trades across an extended time-frame and classify them as simply random noise arising from a chance outcome in a large universe of meaningless data (aka an efficient market). Provided that you cut losses short and let profits run, this vast component of trades will simply contribute to a big fat net zero result to your overall performance.

We cannot underestimate the importance however of this vast array of ‘nothingness’. These are the vast number of trade events that dictate whether or not you are an above average risk manager or not. To the sloppy price follower, the small losing trades will result in death by a thousand cuts with an inevitable building drawdown with potentially no end in sight. Provided you can address this sea of ant bites through your systematic approach to risk management and the real market is Non-Gaussian in nature with Fat Tails……and provided your profit potential is open ended….then you can just sit back and relax under wide diversification and smell the roses.

Unlike predictors….price followers work on the same principle as the casino with a house edge. In the short term we just don’t care whether we are right or wrong…..we just don’t know..but we believe in the edge of the long term and are prepared to accept the losses along the way provided they are small. We believe in the outlier and our ability to reap the rewards that pay for that grueling grind we need to endure to catch them. The participants that fail over the long term are those who sit on the other side of the equation and think they do know the future and commit leverage to their ‘rightness’ only to end up wrong in the long term. This degree of certainty in the edge came from our backtesting not our gut instinct. We trust the scientific method as our preferred road map to step into an uncertain future.

If you cannot exploit the trends that are self-evident in the monthly chart where the impact of all participants are felt, then it is your system that is the problem….not the market. Reliance on a single trend following system to make your bread and butter is not recommended. Those variables that exist in your single system maybe the snags that prevent you from catching the outlier if and when it arrives. However reliance on many different trend following systems that are fairly uncorrelated makes the game easier. One of them is likely to bust through the noise and take a long surf with the outlier.

If we adopt the perspective of a scientist seeking that equation that best describes the reality then as a scientist we understand that our systems (models) are just approximations of the possible reality. As markets are moving feasts built upon principles of adaptive emergence we need many models to increase our chances of collectively capturing the essence of this dynamic reality.

In developing scientific theories it is all about framing the narrative of the hypothesis correctly. This is a lesson we could all take to the trading floor. As trend followers on this Blog we all know the narrative….these are the factoids that when put together in a particular order make the hypothesis that is our trading model. Statements such as ‘cut losses short and let profits run, manage risk at all times, follow the trend until it ends, only risk a small amount per trade, diversification is the only free lunch etc. etc. etc. etc. We have all heard them a million times before.

Now having this array of factoids in your head does not make a good model. It is how you frame them that is important. This means, how do you use them and apply them, how do you prioritize them and why? The hard lessons of experience tells you this story…but to understand this by simply smashing these factoids together will never get you anywhere.

Framing a hypothesis correctly is what helps a hypothesis turn into a great theory which can stand the test of time. What are our core assumptions and how do we validate them in a quantifiable manner? What are ancillary assumptions that are symptomatic as opposed to causal and how do we identify them?……etc. etc. etc. This is exactly how you should go about your trading journey…..discovering an edge and determining how to apply it….though in trading terms we do not use the scientific lexicon to do this.

In dealing with complex systems such as the markets there are no laws apart from those dictated by the physical laws of the universe in which these markets reside. What are these physical laws? What leads to non-stationery emergent features in our financial markets? How do we determine the difference between randomness and non-randomness? How do we identify non-random features of endurance? These are questions we need to ask in framing our hypothesis. Once that is done we then need to test our assumptions and then finally….the trading is the easy part. We just systematically apply our models that have been framed through the process we have undertaken in developing our hypothesis and validated empirically.

Trade well and prosper

The ATS mob

You must be logged in to post a comment.