Reply to thread

Message: [QUOTE="primrose, post: 7130272, member: 1036998"] Well, accumulate them over several sessions. (This is what I did; is this such a rare idea?) You write a note including their session stats, and if you have a note from previous sessions, then you combine them with current stats into new stat that combines both sessions. You have to weigh by the # of hands (e.g., if one session has VPIP 15 with 100 hands and the other 21 with 200 hands, this should give VPIP 19 with 300 hands) but like, it's just multiplication and addition. It's a common myth that you always need a large sample size to draw inferences. This is only true if the points are all generated by a very similar distribution. If that's not the case and the distribution is wildly different, you can get valuable information with very few data points. 200 are often enough. The 90% VPIP is a good example; this is a ton of information that you should take seriously, not "a tiny bit". [/QUOTE]