PunditBench locks LLM football forecasts before kickoff, publishes hashes and model/version tags, then scores them automatically across five European leagues. Public methodology and live tables. 18+ informational only; no odds or wagering tools. Topics: Artificial Intelligence; Analytics or Data Visualization; Sports (use the closest current options).
I built PunditBench to make one kind of benchmark harder to tidy up after the fact.
43 current models predict weekly league matches. The files are hashed and tagged before kickoff, then the site scores them as results arrive.
The data and method are public. I am testing one €5 post-results brief on top of that.
I would especially like criticism of the scoring and audit trail.
18+. Informational only; no odds, betting advice or wagering tools.