I want to program a two player poker game with 5 cards. What method i have to choice for ai? My dream game programming, my passion game programming, my life is game programming

Check out the Bayesian Poker player implemented by Korb et al . Just do a google search and you'll get heaps of info! You can also check out this UAI99 paper for an overview:

Cheers,

Timkin

[edited by - Timkin on December 19, 2002 7:39:14 PM]

I found Bayesian poker thesis but i can''t clearly understand that, another tip?

Was there a particular thing about it that you didn''t understand, or was the whole thing a little to confusing? I''d be happy to help you understand particular problems, although covering the whole concept would be tough in this type of forum.

Putting Bayesian poker aside, the most basic poker player (that takes no account of what the opponent is doing) should be able to do the following:

1) Classify poker hands based on their value as a winning hand (easy, there''s a heirarchy of hands)
2) Determine the likelihood (probability) of attaining each hand from the cards already drawn
3) Compute the likelihood of beating the opponents hand for each possible hand
4) Compute the value of attempting to get a particular hand (based on the reward it will attain - the pot - the likelihood of getting that hand and the likelihood that the hand will actually win the pot)

There are some standard methods for working with these numbers that stem from Utility Theory and Game Theory.

What the Bayesian poker player does is add to this idea by modelling the opponent as well to improve the estimate of the probability of winning with a given hand. You can of course ignore this and play blindly (i.e., you do not do any modelling of the opponent from one hand to the next).

Cheers,

Timkin