Re #1: While I enjoyed Mr Seykota's interview in Market Wizards, I don't particularly follow him, nor am I in one of his tribes, which I find bizarre at best. I also don't like the fact that he's charging for seminars when he can supposedly make so much trading. But I will take the little gems he has offered in his interviews, thank you very much. Re the jerking off, sounds to me that that's more than likely a freudian slip on your part, buffoon, because your posts on ET are a lot like your daily jerk off routine, i.e., "Much ado about nothing" #2: No jackass, if YOU had a drop of common sense in your clown's body, you would realize that the sarcasm applied to you and not to ET. I agree that it is hopeless sometimes on ET, but hey, I love to pick my battles and crush morons like you. Now pull your pants back up, stop surfing for porn all day, and please above all, puh leaze, quit the mental masturbation... wait.... it's too late moron, you're already blind and no one can help you.
I read the first posting in this topic and then skipped to the last posting (above) thinking that the thread might have developed into something interesting and I see that we've gone off topic.
for 20 pages, this thread sure has gotten a lot of views. if it weren't for distractors like nononsense who need to be corrected once in a while, the thread would be pure information. I still think we can carry on with the discussion on using a fraction of kelly to size positions, that is assuming we don't get hijacked again with non productive posters.
I don't think it's correct to assume that traders maximize their utility by "fixed fractional" betsizing. People love to disagree about whether the proper fraction is "the Kelly criterion" or "Vince's Optimal f" or "half-optimal-f" or "one third Kelly". They forget to challenge themselves on their assumption that FIXED fractional betsizing is the way to go. Why must betsize be LINEARLY proportional to bankroll? Where is that carved in stone? Why cannot betsize grow more slowly than linear? As you get richer you become more risk-averse ... why is that obviously incorrect and sub-optimal? Why can't people incorporate their (highly nonlinear) utility functions, into their betsizing schedule, thereby making betsize a nonlinear and non-fixed-fractional function of current bankroll?
Good point, I was reading up on that... the only other option I know of that can change that linear relationship is Ryan Jones' Fixed Ratio approach, but from one article I've seen online, the method could be outperformed by the fixed fractional approach depending on what the sequence of trades is here is the link: http://www.breakoutfutures.com/Newsletters/Newsletter0703.htm If it's the case that one needs to know the sequence of trades to decide whether to use a linear or non linear approach, then it seems like there's little hope left? What am I missing?
I see no good reason for using the fixed ratio method or other methods that change the sizing algorithm based on past results. They assume that future results will either mean revert or trend, and I think that there's no statistical evidence to confirm it. As long as you know your system and it behaves as expected, use its parameters to calculate the Kelly fractional, then based on your acceptable drawdown determine your risk size per position. When you size your trades this way your absolute risk and position size will self adjust, and it won't matter in what succession your winning and losing trades will come. You'll just have the benefit of compounding. If you're 100% sure of your trading system there's no reason not to use the full Kelly. If you're only reasonably sure of it use 1/5 or 1/6 Kelly. If you're not sure, paper trade or backtest intensely until you get enough data to trust it or abandon it..
barney_collier, You seem to be familiar with several sizing methods. Would you care to elaborate on your post? Also, what do you man by incorporate their (highly nonlinear) utility functions, and how do you use or propose to use them?
These underlined things are hyperlinks. You can click on them. Introduction (1) Introduction (2) Explanation (1) Explanation (2) Pictures To learn more