Quote:
Well yeah, but 1.4 million hands @0.2 BB = 60 buyins
Isnt that running bad?
You know how hard it is to pinpoint winrate with such accuracy ? Over such a sample size, if you want to get 99% confidence, it would be approximately +/- 2BB/100. -0.2bb/100 is very very close to the expected value (obviously the pokerev number only takes into account the variation from expected value in allin pots, which is not the complete picture).
As an aside, there was a slightly off-topic discussion in the previous thread about what you're talking about. As the number of hands grows, the number of buyins away from the expected value will NOT converge, contrarily to popular belief, as it has no real reasons to do so. However, The number of buyins divided by the sample size, that is the average difference with the expected value, WILL converge. Quick recap about this
in this probability thread.
edit : Awesome dotbum !