When we started building Brass Pawn ( source on GitHub, AGPLv3 ), we needed a large offline tactics, positional judgement, and endgame dataset that ran completely on-device—both in the browser and natively on iPhone, iPad, and Mac.
Transcribing chess puzzles by hand or scraping unverified problem collections breaks down fast. Two failure modes show up immediately in real training:
Cooked solutions: A position has a flashy sacrifice, so a human annotator marks it as winning—missing a quiet defensive resource at depth 18, or missing a second equally winning move that punishes a student for playing an objectively strong alternative.
Horizon flips: A shallow engine pass (depth 10–12) flags a move as +2.10 , only for depth 22 to reveal perpetual check ( 0.00 ).
Instead of transcribing puzzles, we built a deterministic mining and independent verification pipeline powered by Stockfish 18 .…