DT-WAW: We just optimize losing system into a winning one! It is possible in some rare situations. +20, +20, -20, -20, -6 The above is not a loosing system, it's just a string of five numbers. These numbers could be the result of either a positive or a negative expectancy system. Just thought I'd point that out. Now, the way you ordered them, loosing and winning bets are consecutive. But is there anything about your betting strategy that would give you the expectation that your winning and loosing bets will be consecutive? If you are playing a game of chance, where the next outcome is in no way dependant on the previous one, then there is obviously no way to predict the order of your betting results, therefore any strategy of optimizing your return that depends on the order of those results will fail. Take those results: +20, +20, -20, -20, -6 and put them in a random order. Your strategy will no longer work. If, on the other hand, you are playing a game where the previous outcomes do affect the probability of the next outcome (trends, as in the stock market) then you use that information to attempt to predict the next outcome. If you use all the available information about the probability of what the next outcome will be, then there should be no way to predict the order of betting results. In general, if you can daytrade your portfolio (notice short term trends in it that you can use to predict the continuation of such trends) then your system is not optimal. To put it simply... +20, +20, -20.... ... if you thought that there is an above 50% probability that your next betting result will be negative, because your previous result was negative, then instead of reducing your size, why don't you simply pass on the bet (or take the opposite side.) voodoo ps. outcome here means uptick/downtick, betting result means profit/loss
... if you thought that there is an above 50% probability that your next betting result will be negative, because your previous result was negative, then instead of reducing your size, why don't you simply pass on the bet (or take the opposite side.) There is more to this than the casual observer may realize. In fact, I know of one trader who trades 3 separate systems (position trading, no intraday). He does not take every trade in each system. Instead, he waits for 2 losing trades and then takes the next trade. In some cases he is long & short at the same time, but that happens less than 10% of the time he says. I do not know the specific systems but know that two are trend following in nature (one breakout, the other buy dips/sells rallies within defined trend). Don't know what the third system is. He trades indexes & currencies. He is in effect, system-trading his systems. I haven't look at the math on it but it works for him. Of course, you first need three systems that you are willing to trade.
As I said above, you can trade your system, but only if it's inefficient. And if that's the case, why not simply change the system? I wonder if there are any situations when trading your system is easier or more efficient then simply altering it. voodoo
voodoo, I believe altering the system in effect that is this guy is doing. Most trend following systems suffer from the problem of multiple whipsaws before a nice trend hits. That is the general nature of trending/consolidation. If you were trading just one system on one market you'd want to take every signal because you need that one or two nice trends to payoff all those whipsaws in a given year. He is in effect altering each system by saying "wait for X amount of losers, then take next signal". Whether that actually improves his probability of hitting a winner on the X+1 signal or not, I don't know. There is some correlation between subsequent signals in most trading systems (unlike in games of chance where there each outcome is independent), but I don't know if it is significant enough to make a difference in the long-term results or not. Since waiting for X consecutive losses may take awhile, he trades 3 different systems this way, thereby generating about the same # of signals as if he took every signal on one system. He diversifies across indexes and currencies. He claims his backtesting showed lower drawdowns than typical trend-following systems (he directly compared Turtles & Aberration to his approach). Remember this is position trading capturing long-term trends, not a short-term approach. Since I don't know all the details, I'll see if I can get him to post his methodology here. He's been doing this 2 years now.
dottom, I looked at that approach some time back and found that I ended up missing too many big winners, but I was not using multiple systems. That may give you a better chance. I wonder if your friend did any sort of study of the distribution of losers or just worked from the win/lose ratio?