I am always confused by the "too small a sample size" comment. What exactly constitutes a good sample? I agree that 3K is too small, but what about 10K? I would think most things would average out by then. I know that doesn't constitute "long term", and that if you look at specific examples like number of AA hands out of a 10K sample, you don't get a high enough sample (ie I have 52 AA hands out of 12K).

But can't you get a feel for where your game is headed after 10K hands?