AI Plays TicTacToe

Overview

This project demonstrates an AI agent playing TicTacToe against a human player using a magic square representation and the minimax algorithm. The AI plays optimally, making it impossible to beat—your best outcome is a draw!

---

Jupyter Notebook

For a full walk-through with code and 🎮 play online (no setup required), see the Jupyter Notebook:

---

Features

• Interactive GUI: Click-based gameplay using ipywidgets with a visual grid
• Magic Square Logic: Uses a 3×3 magic square for efficient win detection (any winning line sums to 15)
• Minimax Algorithm: AI makes optimal moves by exploring all possible game states
• Visual Feedback: Color-coded moves (red for AI's ❌, blue for player's ⭕)
• Flexible Start: Choose whether you or the AI goes first

Requirements

ipywidgets
IPython

How to Play

1. Run all cells in the notebook
2. Choose whether to play first (Y/n)
3. Click on empty cells to place your ⭕
4. The AI will respond with its ❌ move
5. Try to achieve a draw—the AI won't let you win!

Technical Details

Magic Square Representation

┌───┬───┬───┐
│ 8 │ 3 │ 4 │
├───┼───┼───┤
│ 1 │ 5 │ 9 │
├───┼───┼───┤
│ 6 │ 7 │ 2 │
└───┴───┴───┘

Each position maps to a magic square value, enabling quick win detection by checking if any three positions sum to 15.

Game Logic

• AI (X): Maximizes score (+1 for win)
• Human (O): Minimizes score (-1 for win)
• Draw: Score of 0

The minimax algorithm recursively evaluates all possible game states to select the optimal move.

File Structure

• Magic square setup and mapping
• Board display functions with ipywidgets
• Game logic:
  • Win detection, available moves
  • Minimax algorithm implementation
• Interactive gameplay function

Challenge

Since the AI plays perfectly using minimax, can you force a draw? Good luck! 🎮