Arbitrage

Why hair care is better than quantitative analysis.

When I first heard the term statistical arbitrage, I thought it was a joke. The kind you play on any non-suspecting, green behind the ears and out of his depth associate you found running around the banks in those days. Little did I know that in years to come, the very concept would entail even more complicated terminology. Pattern recognition, time series forecasting, and principal component analysis will mean different things to different people, but one thing is for sure: the holy grail, it is not.

But let us start at the beginning. Say that you observe someone going to the pharmacy every day for nine days in a row. Would you make a bet that they would do the same on the tenth day? If your answer is why not, then you have mastered the very essence of what trading purely on probabilities is all about. The only thing you have to bear in mind is that you have to shut out every notion that there is anything else going on. There is no fundamental analysis, no macroeconomic interpolations, and no other human input that would ever contribute to what a quant has modelled based upon a historical analysis of the numbers.

So, what are the nuances? Sticking to our example, the happy chap who made his way to get medicine on a daily basis, may have been cured. In other words, a pattern will repeat itself until it doesn’t. Whilst that would normally lead one to conclude that the trade is over, it could be that he just missed the appointment that day, only to resume the pattern the next day. This is where it gets complicated, and no pattern is perfect. There are pauses, breakouts, retrenchments, momentum, mean reversion and anything else you would associate with a random universe.

Some trades look better than others, but what makes things difficult is that we are not the only ones looking at these potential opportunities. There are many equally smart and superbly tooled up traders with all the computing power in the world playing the same numbers. In many cases, it isn’t so much the pattern one has to look at, but how many people are likely to have on the same trade. Once I learned that the way to make money is to “arb the arbs” and do the exact opposite of what the model is telling us, I lost interest. So, the next time someone tries to impress you with a bunch of technical jargon, consider doing something much more productive, like washing your hair.