Noughts and crosses against TypeSafe's Jev model, with every move checked against a perfect solver. Shows how board format and question wording change its decisions.
Stars measure the whole repository, including work unrelated to Jev.
Documented records the linked documentation or source. JevHunt has not independently run or benchmarked this project.
This is a weekend experiment, not an authoritative evaluation. One model version (`jev-1.13.0`), one game, mostly one sample per position. The working notes, including everything that did not pan out, are in [NOTES.md], and the experiments not yet run are in [docs/next-steps.md].
Evidence commit: fb64fd3e286b449e57266413435583c536ce70ef
Discovered through: github-search.
How we review →