CS480-Project-2

CSC Project 2 "Poker"

Poker Win Estimator using Monte Carlo Tree Search (MCTS)

This project estimates the pre-flop win probability of a given starting hand in Texas Hold’em Poker using Monte Carlo Tree Search (MCTS).

Each simulation plays out a full poker hand to showdown (opponent cards + flop + turn + river) and evaluates the result. The MCTS bot uses the UCB1 strategy to guide exploration and determine win rates based on random sampling.

From terminal or VSCode run:

• python main.py

It will output an estimated win probability for a hardcoded hand (default: ['As', 'Ks']):

Change the starting hand by going in the main.py and modify this line:

• state = {'hole_cards': ['Qs', 'Ks']}

You can change the number of simulations per estimate here in main.py:

• mcts = MCTS(root_state=state, num_simulations=10000)