A command-line Mancala game powered by AI using Minimax and Alpha-Beta Pruning, with configurable search depth for adjustable difficulty.
Benchmark AI performance through automated simulations that reveal strategic strength over time.
Player 1 | Player 2 | Depth (Plies) | Games | Description |
---|---|---|---|---|
random_player |
random_player |
– | 100 | Baseline comparison of randomness |
minimax_player_with_variable_plies(5) |
random_player |
5 | 100 | Minimax vs random |
alpha_beta_player_with_variable_plies(5) |
random_player |
5 | 100 | Alpha-Beta vs random |
alpha_beta_player_with_variable_plies(10) |
random_player |
10 | 100 | Deeper Alpha-Beta vs random |
minimax_player_with_variable_plies(5) |
alpha_beta_player_with_variable_plies(5) |
5 | 100 | Minimax vs Alpha-Beta |
python mancala_ai.py
At the end of each simulation, the program outputs win rates, average move counts, and tie frequency — offering insight into each AI’s performance.
Built on AIMA’s adversarial search framework.