Some more comments on Win Rate and other stats.

I agree with most here that a sample of 7500 will mot necessarily encompass all the swings that a player's bankroll will undergo so there is some error. But I would guess that when you do have such a sample the accuracy is somewhere around +/-10% at least 90% of the time or better.

But here's another point, if an adequate sample is only around 100,000 hands, Win rate becomes meaningless except maybe for those who play 1000 hands per day.
Otherwise, (and I would suspect this is the case for most players who are learning) Win Rate measured over long periods becomes meaningless because it fails to account for many other elements such as change in play, stakes, etc.

And the question of sample size applies to all systematic analysis of players. In another note, wish lists of player stats for software to display were requested. I think people requested fold, raise rates, etc. How many hands do you need to sample before you get a predictable 3rd st call rate? 4th st? 5th?