Reply to thread

Message: [QUOTE="dj11, post: 1971429, member: 13851"] In the theme of 'poker is a non-ending game', sample size gets a funny rap. If you're a casual enthusiast who plays intently one night a week (online) after work, a significant sample size would be based on maybe a month's worth of hands and results. Longer would obviously be better, but 1 month is gonna give you enough to begin to understand things. If you are grinding out 1000 hands a day, then maybe that same 1 month's worth of hands would again hold true, so 30K hands. While I know about stats, and significant sample sizes, I disagree with how people tell us we need huge sample sizes before we can intelligently use those stats. Those static stats tell us history, and might suggest trends in static situations, but we want a more dynamic sample size to be useful. In most cases I think a time based sample size is better. Case in point, I have thousands of hands on many CC players for several years worth of play. I know these players are improved over those years, and if I use 1000 hands against player x to evaluate how they are playing today, I am screwing myself. I only need maybe their last 100 hands against me to get a much more accurate read on them. Likewise, if I use my lifetime stats on myself for some self-evaluation, I will be getting stats from when I didn't know or use this or that tactic, and again, will be screwing myself. So a month's worth (IMHO) of stats for either them or myself, will give me a fine sample size to evaluate either their play, or my own. You can evaluate your own play, via a tracker, by isolating hands into time based sets, and then compare different time based sets against each other. So you could compare your hands of Oct 2001, against your play in May 2012. So I start most games with full stats showing on my HUD (when I use it), and about 50 hands in, I change my HUD to show the last 50 hands, and keep adjusting those number to show how both they and I are playing now. If you play random games and mostly see random player, you won't have many hands (if any) against a large majority of the folks at your tables, and any HUD info you see will be small sample stuff that is still useable. [/QUOTE]

Verification

Search

Reply to thread

About CardsChat

Our Community

Our Site

Follow us