Has anyone here ever tried to make a hold 'em robot?

[BottlesOf]

[sfer]

A HE robot would rule. Especially if it's like the one from Rocky IV and also brings you beer.

[jba]

Poki is the the best credible robot I've ever heard of. I'm about 99.999% sure this bot would own the "BEAT PARTY POKER FOR $$$$$$$$ WITH POKERBOT NOW@!!" spam dude.

[ QUOTE ]
Q: How good is Poki?
A: The older version of Poki
that plays in full 10-player games is better than a typical low-limit casino player, and wins consistently against average opponents; but it is not as good as most expert players. The newer programs being developed for the 2-player game are quite a bit better, and we believe they will eventually surpass all human players, perhaps within a few years, or less.

[/ QUOTE ]

[GemiNiveK]

When I was first coded I could only play microlimit. Slowly my AI adapted and now I play everything up to 10/20. I even have a mod that posts to forums. I'm working on a self-mod that screws my hacker's women while he's out.

[Saint_D]

[ QUOTE ]
You'd just have to program in every possible scenario and what action you'd take for it.

[/ QUOTE ]

You clearly don't know mutch about programming. Assuming you only play reasonably starting hands there are about 30 Million combinations to cover. This is just pre flop!

You would probably want to go more with a rules bases approached. Programming the starting hands chart from SSH wouldn't be too bad. Then you could program it to play based on pot odds. This is playing better than 50%+ of low limit players.

You could maybe use some genitic algorithms to discover interesting strategies.

Your real problem is detection. If you had a winning bot, you would want to run it a lot. However, the more you run it the more likely it is to get detected.

I would rather win the old fashioned way. Besides, if you can program this well, I am sure there is something you could make that would net you some profits in a legit way. Poker Tracker, PokerGrader, etc.

-D

P.S. If you have a great idea for the next LEGAL poker program, but lack the skills to make it, PM me. We can at least discuss how hard it would be to make. Maybe I can steal your idea and make millions. [img]/images/graemlins/tongue.gif[/img]

[dutchbrodymoss]

No.

[pokergrader]

It is not that hard at all to create a good poker playing bot in a 10 person limit game. Getting it to interface with an existing poker room would be a challenge, but not too difficult if a person has experience doing things like that.

Making the poker bot is really simple. If you wanted to be really lazy, you just could create a simple neural network and put 100,000 hands worth of your play to train it. If you wanted to create a specialty program, it would be even betting and still not too tough to create. In fact turning the software that pokergrader is based of into a winning poker bot would be incredibly simple to do, given that it can reduce every situation into a number ranking the strength of the current hand, and just add in rules telling it what to do for certain hand strengths.

[KingDan]

I'm sure it can be done for limit hold'em, but I think an all-in or fold preflop strategy in NL ring games would better. AA, KK, QQ, AK push preflop. Some of the crap that it'd get called by, I think it would be profitable.
Get a ton of hands in... good rakeback.

[waffle]

It is not as simple as you think. Please read these publications for more detail, especially "Opponent Modeling in Poker: Learning and Acting in a Hostile and Uncertain Environment."

http://www.cs.ualberta.ca/~games/poker/

[Chex]

[ QUOTE ]
dude I already made like infinity of those last summer at scout camp

[/ QUOTE ]

Ahahahaha, excellent timing! [img]/images/graemlins/grin.gif[/img]