Took me a few minutes admittedly to figure out, but this "SPS", lol, is just the expected value per dollar risked!!!!!
The problem is that markets are not static, plus on a lower win % system, you lose on commissions. So in reality, I would not trade either system since over time both would lose money.
According to SPS, lol, system C (see above) is even more attractive with a SPS of 1.12. But tbh, i wouldn't want to trade a system where the prob of a winner is 0.1%.
The ideal system for me has the following: win rate of at least 30% profit factor of at least 2 expected value per dollar risked (or SPS lol) of $1 or more
That's not the general case, just the special case of exactly two outcomes (only one win amount and only one loss amount). For the general case of multiple outcomes, I prefer the formula I wrote (k*W) using the exact value for k, because the Kelly approximation (= p/|L| - q/W) is always an overestimation, which means you're overtrading. Not good to overtrade.
SPS(C) = 1.221 actually. Despite the low winrate, the payoff makes the system best of the three. Would you refuse to play one of the multi-state lotteries even if the jackpot reached a billion dollars? Lol
PF = pW/qL = .3W/.7L >= 2 : W >~ 5L SPS >~ .3*5-.7 = .8 < 1 Looks like SPS is your bottomline factor. Lol
kut2k2, I made an equivalent claim in: http://www.elitetrader.com/vb/showthread.php?s=&postid=3028937#post3028937 Any references to and/or proofs of the claim would be greatly appreciated. The proof I have uses Jensen's Inequality for conditional expectations which I believe to be beyond the mathematical level of most traders. Thanks, Jim Murphy