Poker Bot Theory

[Larry_Csonka]

Seeing how it seems limit poker is a very mathematical game. I beleive it relies mainly on 3 virtues.
Patience
Math Skill
And the ability to rate your opponents.

All 3 of these skills a computer would be the master at. Using a system like poker tracker has to rate the opponents and adjust play accordingly.
So, My question is, does anyone else beleive it would be possible to make a computer bot that can play limit hold 'em as good if not better then any real life player?


PS. I may start trying to program something like that soon, if you interested in helping PM me.

[Iceman]

[ QUOTE ]
Seeing how it seems limit poker is a very mathematical game. I beleive it relies mainly on 3 virtues.
Patience
Math Skill
And the ability to rate your opponents.

All 3 of these skills a computer would be the master at. Using a system like poker tracker has to rate the opponents and adjust play accordingly.
So, My question is, does anyone else beleive it would be possible to make a computer bot that can play limit hold 'em as good if not better then any real life player?


PS. I may start trying to program something like that soon, if you interested in helping PM me.

[/ QUOTE ]

There's a group at the University of Alberta that has worked on poker bots. Their basic conclusion is that computers can be very strong heads-up players (using game theory and pattern tracking), but that it is much more difficult to create a program that would even be mediocre in full games. Human experts are much better at reading hands than any computer is, and they are much better at weighing factors that are not clearly quantifiable when making a decision - the raw math of a computer knowing that it has exactly a 16.6309% of making a flush when the human expert thinks it's 1/6 makes very little difference in real play.

Their page is here:

http://www.cs.ualberta.ca/~games/poker/
(scroll down to "Publications" in the right column).

Strategy, hand reading, and opponent reading are much more important aspects of limit poker than math.

[Wacken]

I personally believe i should be able to build a poker bot that plays at least decent and be winning at the low limmit games. I have rough ideas on how to make it and am planning to do so someday.

Higher limmits i don't know and i can't say because i haven't played them yet.

Tournament play would be extremely difficult and maybe practically impossible.

[ipp147]

[ QUOTE ]

PS. I may start trying to program something like that soon, if you interested in helping PM me.

[/ QUOTE ]

Why would you try and do this?

[Wacken]

Be real.
Everyone is fair and nice if its about small money. If you have the ability to build a bot that makes you half a million dollars over some time though, you might soon not be so fair and nice anymore. You may say it is bad, it is evel and whatever, but 90% of the population would probably do it if they could, and i can't say i don't understand them.

And even if you want to keep it fair, there is some bot tournament for 100k.

[abuljooj]

i just graduated with a bachelors in computer science, and for my software engineering class, 4 friends and I worked on creating a "poker bot" from scratch. we got as far as having it calculate pot odds perfectly for every possible scenario of hands. we wrote everything in java (for you fellow geeks who know what that is). as we kept working on it, and incorporating more and more techniques we realised that this is extremely hard to build the artificial intelligence into the bot to make it have the good poker player's "intuition", which will always make the bots beatable. the closest we got to the real thing was by incorporating a bayesian learning algorithm, but it still could never be too dependable for more than microlimits over thousands of hours to make a profit. certain set rules can easily be programmed into a bot which will make it gain a profit in the long run but once anyone figures out those rules, the bot will just be another fish! so all in all, its a nice concept, a very fun idea to try to build and perfect, but pointless if you just want to make a profit out of it, u might as well play yourself, you'll obviously have the knowledge and skill if you're going to program it into a robot! have fun! [img]/images/graemlins/smile.gif[/img]