Quote:
Originally Posted by TimM
Win rate and standard deviation were developed mostly for limit games, and have some problems when applied to big bet games. Two different games with the same blinds can play very differently, especially with very large or small stacks, straddles, bomb pots, etc.
I'm more interested in the type of stats Galfond and others want to collect from the streamed hands.
To demonstrate his cheating shows he had sight of others hands, his accusers would need to count how many times he did the following and how often he was successful in doing so. A normal poker player will sometimes get these wrong, by unexpectedly running into stronger hands. He makes his (unbelievably perfect) moves after the river, as he knows exactly what he is up against.
That is the key that people have not focussed on sufficiently:
He makes his (unbelievably perfect) moves after the river.
Responding
Folds when he knows he is losing to a much stronger hand that he cant bluff off the hand (need to count this - happens X times out of X hands when this scenario happens)
Value raises his winning hands. (wins X times out of X hands this scenario happens)
Does not try to value raise clearly second best hands (happens X times out of X hands this scenario happens)
Does not call.
Not calling is a red flag, calling is when one is unsure if one is ahead or behind, but he knows if he is ahead or behind, so it is no point him calling.
Leading
Bluffs against weak winning hands that he expects would fold to a sign of strength from him (wins X times out of X hands this scenario happens)
Does not bluff against much stronger hands(happens X times out of X hands this scenario happens)
There will be a few times when he does not follow the above pattern just to disguise his cheating, but his 94% win rate shows he followed this pattern in his streamed games, though his stats and results in unstreamed games
are markedly lower.