A Python implementation of 5x5 Tic-Tac-Toe, featuring an AI opponent that utilizes either the Minimax or Monte-Carlo Tree Search (MCTS) algorithms. This project was developed for UAF CS605 Artificial Intelligence, Spring 2021.
- 5x5 Tic-Tac-Toe: A larger version of the classic game, with unique win conditions.
- Custom Win Condition: The game now detects L shapes as win states for players, instead of the traditional straight lines or diagonals.
- AI Opponent: Play against an AI that can use either the Minimax algorithm with alpha-beta pruning or the Monte-Carlo Tree Search algorithm.
- Human vs Human: Play against another human locally.
- AI vs AI: Watch two AI players compete against each other.
- Game Replay: Replay previously saved games.
- Game Saving: Save your game progress to a file and resume later.
.
├── .gitignore
├── LICENSE
├── README.md
└── tictactoe/
├── __init__.py
├── ai.py
├── game.py
└── tictactoe.py
tictactoe/tictactoe.py: The main entry point for the game. Contains the user interface and handles game modes.tictactoe/game.py: Implements the core game logic, including board rendering, win condition checks (including the new L-shape logic), and game saving.tictactoe/ai.py: Contains the AI logic, including the Minimax algorithm with alpha-beta pruning and the Monte-Carlo Tree Search algorithm.tictactoe/__init__.py: Empty file to mark the directory as a Python package..gitignore: Specifies files and directories to be ignored by Git, such as virtual environments and cached files.LICENSE: The MIT License under which this project is distributed.README.md: Documentation for the project.
- Clone the repository:
git clone https://github.com/Farzan-Kh/TicTacToe.git cd TicTacToe
-
Run the game:
python tictactoe/tictactoe.py
-
Choose a game mode:
- [1] Play a new game: Start a fresh game.
- [2] Continue a game: Load a previously saved game.
- [3] Watch a previously played game: Replay a saved game.
- [4] Exit the game: Quit the program.
-
If playing against an AI, choose the algorithm:
- [1] Minimax: A deterministic algorithm that evaluates all possible moves.
- [2] Monte-Carlo Tree Search: A probabilistic algorithm that uses simulations to determine the best move.
-
Follow the on-screen instructions to play the game.
- The game is played on a 5x5 grid.
- Players take turns placing their marker (
XorO) on an empty cell. - The game ends when:
- A player forms a winning L shape (e.g., a vertical line with a horizontal line attached).
- The board is full, resulting in a draw.
- Explores all possible moves to determine the optimal one.
- Uses alpha-beta pruning to reduce the number of nodes evaluated in the game tree.
- Simulates multiple random games to evaluate the potential of each move.
- Balances exploration and exploitation using the Upper Confidence Bound (UCB) formula.
- Save a game: Enter
wqduring your turn to save the game and exit. - Load a game: Choose option
[2] Continue a gameand provide the path to the saved file.
- Choose option
[3] Watch a previously played gameto replay a saved game. - The game states will be displayed sequentially with a delay.
Welcome to 5x5 Tic-Tac-Toe!
At this juncture, you have several options:
[1] Play a new game.
[2] Continue a game (requires file input).
[3] Watch a previously played game (requires file input).
[4] Exit the game.
Choose an option 1/2/3/4: 1
Would you like to play against:
[1] A human player.
[2] An AI player.
[3] I'd like to watch two AI players play.
[4] Exit the game.
Please choose either 1, 2, 3, or 4: 2
Which algorithm would you like the AI to use?
[1] Minimax.
[2] Monte-Carlo Tree Search
Please choose either 1 or 2: 1
Based on the state of the board, I can see that it is X's turn.
This is the current board state:
| | | |
X | O | | |
_______|_______|_______|_______|_______
| | | |
O | X | | |
_______|_______|_______|_______|_______
| | | |
X | O | | |
_______|_______|_______|_______|_______
| | | |
O | X | | |
_______|_______|_______|_______|_______
| | | |
X | O | | |
| | | |
This project is licensed under the MIT License.
- Developed by Dylan Palmieri for UAF CS605 Artificial Intelligence, Spring 2021.