Let's say you have $100,000 and your max tolerable drawdown is $25,000. You would then assume $25,000 is your total capital Then you would calculate the Kelly/Optimal F for your trades. You can then use 1/2 Kelly/Optimal F but you apply the fraction to the $25,000 not the $100,000 Correct me if I'm wrong but in a lot of systems, the Kelly/Opf, will generate a 99% drawdown if traded to infinity (but depending of the number of trades per year, that might take decades/centurities). By using 1/2, not only you protect against the usual stuff (estimation errors, fat tails) but also you make sure it becomes unlikely you will breach your max drawdown. At the same time you will maximize your capital growth given your risk tolerance (which might not happen if you use a fixed percentage rule of thumb like 0.5% regardless of the trade) Thoughts?
Imho, you can forget about the Kelly criterion in real-world trading application. However, the idea of using only 1/4 of the capital, if your max tolerated DD is 25%, sounds good to me, as given enough time, you do have a concrete possibility of hitting a 100% DD, with respect to used margin. It remains the problem that with 25K only you may be seriously undercapitalized in the mkts which make most sense to trade (futures).
The main factor that has to be taken into consideration is your performance stats. 1. Your average win rate. 2. Your average reward risk ratio. 3. Std deviation of them both. Assuming your system has run long enough, it should give you a good statistics. Another thing is, deciding whether or not to include a black swan event in your position sizing calculation. If you include, your reward would be diminished due to lower position size. If you take off the black swan events from including in the position sizing calculations, then you are prone to a larger drawdows. You have to find somewhere in between.
(1 - x)^y = ((100000 - 25000)/100000) I'd solve this for x knowing that y is the Max Drawdown ++ You Get x the Max % to Risk of your funds per trade.
It's home made. For exemple if the Max DD is 15 consecutive losses, Then y = 15. And x = 1.89% which means that, If you risk 1.89% per trade then, you'll reach Max DD (0.75%) after 15 straight losses.
in real-world, where average sharpe-ratio are in order of 0.5, continuous finance kelly criterion (f=1) works great. of course, in dream-world of data-snooper backtester kelly fails, but the fault, dear Brutus, is non in our stars , but in ourselves
I see what you're saying, but if any trader in the prop or hedge fund community risked according to Kelly, they would be in serious trouble with their employers if they could even afford the initial margin.
Thats why I mentioned 1/2 Kelly fraction of the max drawdown number. I believe that idea can be considered conservative
cjbuckley, E. Thorp in his famous 'The Kelly criterion in blackjack, sports betting and the stock market' (2006), shows that Warren Buffet is (was?) essentially a full kelly bettor. But probably Mr. Buffet is not real-world, is an other planet