Bayesian Analysis of player tendencies
Posted by DirtyD
Posted by
DirtyD
posted in
Gen. Poker
Bayesian Analysis of player tendencies
I was researching Bayes Theorem a bit after Phil mentioned it in a recent video, was wondering if someone who knows more about it can help with a question:
If the overall player pool 3bet frequency is x, and a player has 3bet y of z times, what is the best estimate of his true 3bet %?
This doesn't quite work to plug into the equations I've learned so far. Maybe the question is poorly formulated and I need to be more specific. For example, I know how to find how likely it is that the player's true 3bet % is one of two options I pick (say, 10% or 20%). Maybe it's also necessary to make some assumptions about the distribution of the player population?
Anyway, any help would be appreciated.
Loading 5 Comments...
This is the best free article i know of: http://www.husng.com/content/interpreting-small-sample-sizes-bayesian-estimators
Try to plug it into excel or a similar software.
Thanks, I don't 100% understand how this guy got his formulas, but I plugged them into a spreadsheet and they spit out numbers that seem plausible.
There is one simple answer wich doesnt require more info: if y/z > x then "actual" 3b% is less than y/z and if y/z< x then "actual" 3b% is more than y/z. This effect is usually quite big for small samples.
For more info you would need more assumptions, but I really don't think you need more than what is above at the tables ;).
http://www.runitonce.com/poker-training/videos/pudge1/
first video that sam greenwood made for rio
Be the first to add a comment