Connect 4

Video Demo: https://github.com/Alireza-njt/Connect-Four/releases/download/video/CS50.Python.Final.Project_Connect.4.with.Minimax.Alpha-Beta.Pruning.mp4

This project represents my final submission for Harvard's CS50 Introduction to Programming with Python course - a fully-functional Connect 4 implementation featuring an intelligent AI opponent. Designed for terminal/IDE environments, the game combines classic gameplay with modern artificial intelligence techniques.

Project Origins

The architecture draws inspiration from:

• Project 2 of CS50's AI course (Tic-Tac-Toe implementation)
• Traditional Minimax algorithm concepts
• Alpha-beta pruning optimizations
• Classic board game implementations

Technical Implementation

Core Game Engine

initial_state()

Initializes the game board as a 6x7 grid (standard Connect 4 dimensions) with all positions empty. The state is represented as a 2D list containing:

• Red pieces (player/AI)
• Yellow pieces (player/AI)
• Empty slots

actions()

Dynamically calculates all valid moves by:

1. Scanning columns from bottom to top
2. Identifying the first available empty slot
3. Returning coordinates for valid moves
4. Enforcing game rules (pieces fall to lowest available space)

result()

Generates new game states by:

• Creating deep copies of the current state
• Validating move legality
• Applying player moves
• Preventing illegal state modifications

Game Logic

winner()

Implements victory detection through:

1. Horizontal line checks (4-connected in rows)
2. Vertical line checks (4-stacked in columns)
3. Diagonal checks (both ascending and descending)
4. Comprehensive board scanning
5. Early termination when victory detected

terminal()

Determines game conclusion by:

• Checking for any winning condition
• Verifying board fullness (draw condition)
• Providing immediate feedback for game over states

utility()

Quantifies game outcomes with:

• +1 for AI win
• -1 for player win
• 0 for draws/non-terminal states
• Supports Minimax evaluation

AI System

minimax()

The intelligent core featuring:

• Depth-limited search (5-ply)
• Alpha-beta pruning optimization
• Recursive state evaluation
• Move prioritization
• Adaptive strategy based on game phase

Development Challenges

Key obstacles overcome:

1. State Representation: Efficient 2D list management
2. Move Validation: Ensuring physical game rules
3. AI Optimization: Balancing depth and performance
4. Terminal Detection: Accurate game state evaluation
5. User Interface: Clean terminal visualization

Enhancements Over Base Requirements

1. Visual Improvements:

   • Colorama-powered colored output
   • Pyfiglet ASCII art headers
   • Formatted board display

2. Gameplay Features:

   • Player color selection
   • Turn-based alternation
   • Win/draw detection
   • Input validation

3. AI Capabilities:

   • Competitive gameplay
   • Strategic decision-making
   • Performance optimization

How to Play

1. Run python project.py
2. Select your preferred color (Red/Yellow)
3. On your turn, enter a column number (1-7)
4. Watch the AI respond with its move
5. First to connect four wins!

Red always moves first (whether player or AI). The game continues until a player wins or the board fills completely (draw).