Connect Four AI with Minimax Algorithm

An intelligent Connect Four game featuring a competitive AI opponent powered by the minimax algorithm with alpha-beta pruning. Built with Python and Pygame, this project demonstrates advanced game theory concepts and AI decision-making in a classic two-player strategy game.

Features

• Minimax AI with Alpha-Beta Pruning: Efficient game tree search that evaluates thousands of positions per move
• Dynamic Search Depth: Adapts search depth based on game state (5-10+ levels deep)
• Advanced Heuristic Evaluation: Custom scoring system that evaluates board positions, threats, and winning patterns
• Real-time AI Statistics Dashboard:
  • Positions evaluated per move
  • Search depth used
  • Average thinking time
  • Positions evaluated per second
• Smooth Animations: Piece drop animations with particle effects for wins
• Interactive UI: Column highlighting, move previews, and last AI move indicator
• Professional Game Design: 3D-styled pieces, gradient effects, and responsive interface

Technical Implementation

• Game Logic: Minimax algorithm with alpha-beta pruning for optimal move selection
• Heuristic Function: Multi-pattern recognition system evaluating:
  • Open three-in-a-row sequences (high priority threats)
  • Two-in-a-row patterns with potential
  • Strategic position values (center column preference)
  • Opponent threat detection and blocking
• Performance Optimization:
  • Shallow copy for board states instead of deepcopy
  • Early pruning to reduce search space
  • Efficient win detection through string pattern matching
• UI/UX: Pygame-based interface with custom graphics, animations, and sound effects

AI Performance

• Evaluates 10,000-50,000+ positions per move depending on game state
• Achieves 1,000-5,000 positions/second evaluation speed
• Plays at intermediate-to-advanced level with strategic blocking and winning moves
• Adapts search depth dynamically as game progresses

Technologies Used

• Python 3.x
• Pygame: Graphics rendering and game loop
• Custom Algorithms: Minimax with alpha-beta pruning, heuristic evaluation
• Game Theory: Zero-sum game optimization, adversarial search