Blackjack Strategy Analysis

A comparative study of three approaches to optimal Blackjack strategy:

1. Probability-based calculation - Exact expected value computation
2. Monte Carlo simulation - Empirical estimation through repeated play
3. Q-Learning - Reinforcement learning approach

Game Rules

• Single deck (52 cards)
• Player actions: Hit, Stay, Double Down
• Soft/Hard hand handling (Ace = 1 or 11)
• Card counting enabled (non-independent draws)
• Dealer hits until ≥ 17

Project Structure

blackjack-strategy/
├── core/           # Game engine
├── strategies/     # Three strategy implementations
├── analysis/       # Comparison and visualization
└── tests/

Status

🚧 Work in progress

Author

Roger Xia - UCL Mathematics