Thursday, August 04, 2011

An expensive lesson

I have said more than once on here that historically the success rate of a trend following system can be as low at 40%, but even then you can make significant sums of money by controlling your risk (and losses) and by letting your profits run. This is on the assumption that the system shows a positive expectancy where your average profit is bigger than your average loss. (This is clearly shown on the summary results of the trades log).

It is also a critical part of the overall trend following philosophy that you need to react to all signals your system gives, as by missing just one or two signals a year, you have missed out of the major trends (and profits) which can offest your losses from other trades, and generate the overall profits for the year.

Well, I've made a royal screw up by not following my own guidelines, and am now in a position of watching a potentially significant trend develop, while being stuck on the sidelines. This makes me feel a lot worse than entering a position (or bunch of positions) and being stopped out for a loss. At least if I had done that, I would have followed my rules. It's also worth remembering the saying that "You can't win the lottery if you don't buy a ticket". Well, I didn't buy my ticket and now all those profits are passing me by.

I could enter positions now, but in a lot of cases the ideal entry point has been left behind, adversely skewing the potential risk:reward ratio.

How did his happen? There are two points worthy of note here:

1) I had formed an opinion on what I believed would happen in the markets, rather than following the signals my system indicated. Specifically, I had expected the market to bounce following the resolution of the debate on the US debt ceiling. The plan was to wait for the bounce to finish after the agreement was signed, and, if the indices were still on a short signal, to start loading up on short positions then. As it happens, the bounce occured between the close last Friday and Sunday night when the futures markets opened (before the agreement was ratified), since when all the major indices have basically been trending downwards non-stop. 

2) I had also developed a recency bias - simply put, the poor relative performance the last few months (where no meaningful trend had developed, and led to a run of losses) had made me wary of entering the market too soon, rather than blindly follow the signals. This was not helped after I got hit with a run of losses in March, when I received a short signals on the indices (just prior to the Japanese Tsunami) and plunged into too many positions, too quickly. The sharp rebound resulted in a number of losses, as can be seen on the trades log.

William Eckhardt (Richard Dennis' right hand man) stated in his Market Wizards interview that he had run two parallel accounts following the same basic signals etc - the only difference was that, on his own account, he overrode some of the signals, exited some positions early etc. The other account, which he ran for an associate, simply followed the internal logic of the basic system. The result?

"Although the performance in my account was good, the account trading entirely on the mechanical system definitely did better. I had known that a good system would outperform me in a windfall year, but I thought I could outperform the system in a mediocre year...this experience indicated otherwise".

The moral of the story is clear - if you have taken the time to develop a historically profitable system, second guess the signals it generates at your peril!

No comments:

Post a Comment