that's interesting, but i'm not sure i follow or agree - i can't put numbers on it, but intuitively his variance feels a bit like, but less than, poker's varian

that's interesting, but i'm not sure i follow or agree - i can't put numbers on it, but intuitively his variance feels a bit like, but less than, poker's variance. Poker graphs look like a sawtooth of small losses and big wins. This means he would have to get spectacularly unlucky to be down after 50 sessions, given his positive pre-tax expectation[1]. so it depends how he structures his taxes. if he can get his taxes calculated yearly instead of per session, taxes make him a bit less of a winner. if his taxes are calculated per session, he is -EV. this is the same problem that poker players face, IIRC the precedent there is that if you file as a small business you can sum your wins and deduct your losses annually. (/me used to be a small-time poker player) [1] expectation in blackjack is computable because perfect play is deterministic, expectation in poker is harder because it depends on human factors like your mood and your opponent's skill level and his mood. you can only measure past expectation. A lot of players think they are winners but are really breakeven or losers, but blame it on bad luck.