PsOpti
From PokerAI
PsOpti is a program designed to play only heads-up limit Texas Hold'em. Its playing strength is considered to be at an advanced level. PsOpti was built using a game-theoretic approach (i.e. building on the ideas of an equilibrium strategy for both players) for an abstract form of poker to make the problem tractable. In all experiments, PsOpti performed well and even held its own against world-class competition for 7,000 hands (which is small by computer standards, but long by human standards). However, given enough time, strong competition can still eventually find PsOpti's weaknesses and exploit them. The experiments conducted on PsOpti confirmed the presence of two problems inherent to programs that are designed and built using a "pseudooptimal" game-theoretic approach:
- In pseudo-optimal game-theoretic solutions, there are weaknesses due to approximations and those weaknesses are permanent. This means that once an opponent discovers the program's weaknesses, those weaknesses can be exploited forever.
- Because game-theoretic solutions are non-exploitive, strong players can adopt a style of play that probes for weaknesses without the fear that game-theoretic programs will punish them while they perform this (usually very predictable) activity.
See also
- Other popular Pokerbots
References
- Terence Schauenberg, Opponent Modeling and Search in Poker, 2006 [1]
- Darse Billings, Neil Burch, Aaron Davidson, Robert Holte, Jonathan Schaeffer, Terence Schauenberg, and Duane Szafron. Approximating game-theoretic optimal strategies for full-scale poker. In International Joint Conference on Artifcial Intelligence, pages 661-675, 2003.
