The Crew Game Theory Analysis

Non-LLM agents compete in a cooperative trick-taking strategy lab

Tech Stack

About This Project

An interactive simulation and game-theory analysis of The Crew card game. Five non-LLM strategies—including information-set Monte Carlo tree search, reinforcement learning, greedy search, and rule-based planning—play matched, deterministic missions. The application compares mission success, task completion, computational cost, pairwise dominance, and replays the highest-scoring game.