Determining the probability that you're a winning player

[El_Hombre_Grande]

I'm not a math wizard, but the "long run" is a very, very long time with respect to live play. In fact, so long, that most winning players end up moving up prior to having the type of stats an online player can accumulate in a month.

So live I think its reasonable to monitor your results and make sure you are solidly beating the game, but i don't know about staying at 3-6 limit until you have a 100k under your belt. If you can be honest with yourself and point to numerous errors in your opponents play that you can and do exploit AND are actually winning over a substantial # of hands, that good enough for me.

[pzhon]

If you want to perform that type of Bayesian update, you want to use the density function of the normal distribution. Express the results as x standard deviations above the mean. The density is roughly

constant * e^-(x^2)/2.

The constant, 1/sqrt(2 Pi), doesn't matter for this type of calculation.

For example, the density at 2 standard deviations away from the observed result is e^-2 times the density at 0 standard deviations away from the observed results, so the observations would favor the latter by a factor of e^2 ~ 7.