Convergence

[Cerril]

So there's a lot of discussion about various stats and how long they take to converge [with reality]. The most common of course is winrate (which takes 50k hands or so to yield even tenuous results), and many discussions involve VP$IP and PFR percentages (which converge in only a few thousand hands to pretty accurate numbers, provided no change in play).

The ones I'm curious about now are session win/loss rates (i.e. 52% winning sessions, 45% winning, so on) assuming sessions of equal length, and the various 'won at showdown' percentages (overall, and for each hand type... which is really the same question).

How many sessions before you start seeing a reasonably accurate number? How many hands shown down? Any thoughts?

[deacsoft]

[ QUOTE ]
How many sessions before you start seeing a reasonably accurate number? How many hands shown down? Any thoughts?

[/ QUOTE ]

I'm not saying that you're wrong here. However, many players who participate on this forum are looking at playing poker from a different perspective. They (myself included) don't believe in sessions because, in the end, it's all just one big game. In addition, a "session can mean so many different things to different players. For one it could mean two hours on single-tabling and for another it could mean six hours of 8-tabling. Therefore, BB/100 and such stats will remain the default settings on this forum for years to come.

P.S. You don't need pokertracker to figure your BBs/100. Keep good records and just do the math.

[Cerril]

Well sessions, if they're of equal length, are another way of dividing up your play arbitrarily. Say I play 3 hour sessions at 60 hands/hour and four table. So in a given workweek I might play ~3500 hands or 20 180 hand sessions.

But you're right... I can't think of a meaningful way to make that stat universal.