Can somebody point me to a site with the math for positon sizing optimized for not blowing out based on expected win percent gain, expected loss percent and expected percentage of winners? I used to use an online calculator and now I can't find it, the guy at the site hasn't answered my email about it. I would like to get the math down and incorporate it in my strategy code. Something of note came from using the online calculator and looking at the math: no strategy merits a bet size greater than 100% of your account, that is the theoretical maximum. That rules out margin I suppose, perhaps if you are spreading money across more than one system somehow it is justified but I think not. Max
1% on each trade is what i prefer. I don't like concepts like optimal f because you never know what you biggest loss could be. This is an interesting article. http://www.futuresmag.com/futuresclassroom/articles/fcr_a_7.html
Hi Max, Even if both sites don't offer any free online calculation tools (and are trying to sell some custom made programs), I would anyhow recommend them for the information they offer : => http://unicorn.us.com/trading/prosizer.html The Prosizer program from Alex Matulich may include the kind of calculations you are looking for when it comes to simulate the impact of different kind of position sizing strategies on a given trading system. The documentation page of the program is worth a look as it offers a quick review of the main common risk strategies and an alternative way of calculating the expectancy of a trading system (versus the original approach developed by Van Tharp) => http://www.tradelabstrategies.com/ Get there the free book of Larry Sanders and be prepared for an hour of two of a very intense and interesting reading. His explanation of the famous marble game and the subsequent introduction to MonteCarlo simulation concepts (in order to analyze the impact of different risk strategies) are really good and quite exhaustive : this is by far the best introduction/presentation I ever read on these subjects. Eddy