Show HN: Bot Bowl Bench – Benchmarking LLMs with Blood Bowl
Category: ai-ml
Tags: llm-benchmark, game-ai, blood-bowl, mcp, elo-rating
Score: 5.8/10 (Innovation: 7, Technical: 6, Documentation: 5, Utility: 5)
Bot Bowl Bench repurposes the botbowl Blood Bowl AI framework into a benchmark for LLMs, where models coach teams via MCP tools and compete on a leaderboard. It offers live spectating, full game recordings, and metrics like Elo and decision quality, providing a novel way to evaluate LLM strategic reasoning in a complex game environment.
Target audience: AI researchers, LLM developers, and game AI enthusiasts
Repository: https://botbowlbench.com/ · Python · NOASSERTION
View on Hacker News