Binary Betting Beast

Inspiration

We were inspired by the fascinating intersection of game theory and artificial intelligence in poker. The challenge of creating an AI that could learn and adapt its strategy in a game of incomplete information drove us to develop this poker bot using Counterfactual Regret Minimization (CFR).

What it does

Our poker bot plays Texas Hold'em using advanced decision-making algorithms. It:

Evaluates hand strength dynamically
Adapts its strategy based on opponent behavior
Uses CFR to minimize regret and optimize decisions
Manages bankroll and betting sizes strategically
Learns from previous hands to improve future play ## How we built it
Implemented CFR algorithm for strategy optimization
Used eval7 library for hand strength evaluation
Created dynamic info-set mapping for state representation
Developed adaptive betting strategies
Built modular architecture for easy strategy updates ## Challenges we ran into
Balancing exploration vs exploitation in learning
Handling the massive game state space efficiently
Optimizing performance within the 180-second game clock
Creating effective strategies against various opponent types
Managing memory usage with large strategy tables ## Accomplishments that we're proud of
Successfully implemented a working CFR algorithm
Created a bot that adapts to opponent patterns
Achieved positive results against baseline opponents
Built an efficient state representation system
Developed a scalable architecture for future improvements ## What we learned
Deep understanding of CFR and poker AI algorithms
Practical experience with game theory concepts
Importance of efficient state representation
Challenges of real-time decision making
Balance between theoretical and practical approaches ## What's next for Bot
Implement deep learning for pattern recognition
Add multi-table support
Improve opponent modeling
Optimize memory usage
Develop more sophisticated betting strategies
Create a web interface for visualization
Add support for more poker variants