![]() |
|
#51
|
|||
|
|||
|
Just with Pokertracker and Pokerstove. A bot could be programmed to make decisions based on hand ranges and pot equity.
Hell I might be able to program a winner using those two programs. What makes you think no one has already done this?? |
|
#52
|
|||
|
|||
|
Bot that beats 10/20 NL ? Oh my god, I Guess its possible but the person who code that bot has to be a ALOT better player than the bot, the bot has no idea how other people play their hands, the bot knows nothing about their styles, if they are tilting or what ever.
|
|
#53
|
|||
|
|||
|
I'm a compsci major at MIT but admittedly I know very little about this. Could you PM me some more info?
|
|
#54
|
|||
|
|||
|
[ QUOTE ]
What makes you think no one has already done this?? [/ QUOTE ] I dont think anybody did that but its just my guess. I dont have trouble believing that its possible or even that someone already programmed it. |
|
#55
|
|||
|
|||
|
[ QUOTE ]
I'm a compsci major at MIT but admittedly I know very little about this. Could you PM me some more info? [/ QUOTE ] You can google info about backgammon playing programs (JellyFish, Snowy) and some about neural networks. In theory its possible to program game playing program without even knowing the rules of the game. |
|
#56
|
|||
|
|||
|
Raised: Why would the bot know nothing about the other players styles? Bots can store data and adapt too them.
metamath: http://www.cs.ualberta.ca/~games/poker/ The university of alberta leads the field in poker AI development. There are a number of interesting papers listed there as well. If you want a go at writing your own bot try poker academy pro. It makes it easy to write a bot using your own algorithms and plug it into their software to test, you can also have your bot play on their poker server against real players. |
|
#57
|
|||
|
|||
|
when u guys find a bot that can beat shorthanded nl online games let me know. I'll play him for a large sample size, and I think I'll win. And truth be told im pretty much a donkey, nowhere near the great nl players
|
|
#58
|
|||
|
|||
|
for the sake of discussion, i know firsthand of bots killing 5/10 limit games...
imo it's only a matter of time til we see bots crushing nl |
|
#59
|
|||
|
|||
|
Aren't there scripts for AutoHotkey and such that automate common decision makings into hotkeys+mouseclicks? If all you have to do to c-bet is press F3, it's not too hard to play really, really fast.
|
|
#60
|
|||
|
|||
|
[ QUOTE ]
for the sake of discussion, i know firsthand of bots killing 5/10 limit games... imo it's only a matter of time til we see bots crushing nl [/ QUOTE ] 5/10 limit is a total joke. u could have a bot that understands agression and that is very detailed with respect to value betting that can beat that game. FWIW, i doubt this bot 'destroys' the game. That would imply the bot would be a winner at 10/20 or 15/30, games that are typicall drastically different online than 5/10. Admitedly, this is without considering a program like PTrack implemented into the bot, which im sure is possible. anyways i have a friend who has some experience with this, and I think its possible up to a certain extent(limit), but after that it would become v. difficult. I dont knwo enough abouit the intricaticies, but unless we're talkin about a super poker bot building team (in which case I can see this being done), this doesn't seem like it could be easily accomplished. I think things change A LOT when u consider no limit. Also, the 100/200 and above LHE games would [censored] kill a bot, no questions asked. The programming, testing and time necessary to make a bot on par w the avg player in these games would be sick. I don't know much of anything with regards to the programming part, so i'm not an expert on this topic. I do think it's possible, but I suppose what im saying is that it would be a very difficult and complicated endevaor-not sure if it will ever be done to the point of actually beating high stakes games. |
![]() |
|
|