[Talking Poker]

My best guess: 100,000 hands.

I can say for sure that 10,000 hands isn't enough, becauese looking at groups of 10,000 hands I've played (at the same game), I see vastly different results. I don't have a big enough sample size to say if my 100,000 hand chunks will be the same, but I think that's a good estimate.

I suppose 50k hands might be enough... I guess it depends on your definition of "negated."

Do yourself a favor and throw your sessions into Excel and graph them. Hands on the X Axis and $ on the Y. Then throw in a trend line, and if you really want to get fancy, figure out your $(or BB for multiple limits)/hand average and graph that too. The first graph should show a nice upward slope. The second one will have a lot of variance to it in the beginning, but over time, it will level off. When it's barely moving at all over groups of 1000s of hands, IMO you've reached the long run - or have at least mostly negated variance. It's hard to say, but for me, that number seems to be somewhere around 35k hands. It still moves after that, but only about a 3 cent per hand max swing over the course of thousands of hands. Sounds like negated variance to me.