This is a project for CS229 at Stanford. The goal of our project is to implement an automated Texas Hold'em poker player that applies machine learning techniques (such as Reinforcement learning and Supervised learning schemes). Poker, compared to other strategy games like chess or checkers, is a more interesting game to analyze with machine learning techniques due to the amount of uncertainty in the game. Specifically, poker players may bluff or try to play in unpredictable ways as part of their strategy.

The player must play optimally (with the given information) in 1 on 1 poker games. We also plan to look at issues like opponent modeling and strategy. Variables that we have investigated so far include size of bets, game stages, and value of hands. Future variables to investigate are aggression, unpredictability of players, table position, chip count, and length of game. We also plan on imparting a personality on the poker application. The poker application's personality is an important characteristic to induce uniqueness and randomness in its play, so that it cannot be reverse-engineered and beaten trivially.