Wandering through the traders (not analytical, of course) open spaces of the Internet several times came across negative statements about my support for one of the points of building robust trading systems, which was described by Dean Hoffman in the article `` Trading Systems, who work & quot;.
http://www.traderstech.net/2011/02/16/trading-systems-that-work/
In particular, the following point caused the greatest rejection:
-using the same rules for different markets
One person says that he came up with a system, which for three years earned on RI every year and, wherein, didn't make money on Si — hence the conclusion that for different markets there cannot be the same rules. Logic, certainly, iron. But I think, which is likely, that in the fourth year of work at RI, the system will merge, and on Si it will work. Fit certain systems, as for RI, and for Si, so they make money on the left side of the graph, не проблема. The problem is different — so they make money on the right side. And no one said that creating a robust system is easy.. And it is doubly difficult for her to make money with the same conditions and parameters on different instruments..
As I have mentioned more than once, my opinion on testing systems for single tools is very simple — this is a useless occupation. We can never guess — was it playing out some kind of market inefficiency or just a curve fitting. And no IS and OOS will help in this matter.. Even quite likely, which was inefficiency, but there is no guarantee that it will continue further. But a positive test of the system with the same conditions and parameters for a large portfolio of instruments will definitely give an idea that the market inefficiency has been found.. On condition, certainly, that the system is programmed correctly and does not look to the future.
A. Gorchakov introduced a very correct systemic concept of `` fight against zero ''. The point was that the system usually earns two or three short periods a year., and the rest of the time there is a struggle with zero, so that if u drain, then a little. This is also the case in systemic portfolio trading. — while some tools are earning, others are fighting zero. Then those who have earned start to fight with zero, but earn new tools. Then the third, fourth, etc.. The most important thing is that the total earnings are greater than the total losses from the fight against zero. But in the event that only one instrument was played, then maybe the first time he would have earned, and then would start fighting zero for many years.
That is the meaning of the paragraph
-using the same rules for different markets
not that it's an end in itself — use the same rules and parameters for different markets, but that it simply cannot be otherwise — we don't recognize by positive tests, robust this system or just a random fit. And if in doubt, then what's the point of tempt fate on real trading. We will not be sure in such a system and at the first drawdown we will not stand it and stop the game..
