[Paul2432]

"your sample size is too small" is a common mantra here. We all know it takes hundreds of thousands of hands to have a high degree of confidence in win rate.

However there is a second way. That way is evaluating the quality of a player's decisions directly. I think a bad player can sometimes be identified after as few as 10 or 20 hands (for example if I see a player call a UTG PFR with junky off suit cards and then call down with bottom pair no kicker a couple of times, I am certain this player is a long term loser).

Identifying a winning player is probably more difficult. How many hands would an expert need to observe a winning player to be able to state definitively that that player is a winning player? (assume for purposes of this question that the player in question does not have tilt issues)

Paul