AI Playtesting

The objective is to use AI agents to playtest and subsequently help balance card games. To that end, we are building multiple prototypes of card games designed by us and implementing different reinforcement learning algorithms to train an AI agent to play. Through the semester, we plan to incorporate various mechanics into our game making it progressively more complex. A trained AI agent can go through hundreds of thousands of iterations of the game and identify dominant strategies to win the game. This can provide valuable statistics to a game designer and help make educated game design decisions.

Project Instructors: Mike Christel, Scott Stevens

• Ram Iyer
• Ziheng Xiao