Why this exists

Everybody does their puzzles. The rating goes up. Then, in a game, a knight lands on c7 and it is only after the rook is gone that the word “fork” arrives. The puzzle told you there was something to find; the game did not.

So this trains three separate things instead of one: knowing a motif, recognising it when nobody announces it, and seeing it in your own games, where it actually costs you. Each motif is a journey: a lesson, a drill with the theme announced, then a test with it hidden and mixed with the motifs you already know. The gym gives you a rating that is verified move by move. And the scan reads your games with an engine, finds the tactics you walked past, names them, and hands them back as puzzles until you have fixed each one twice.

Where the puzzles come from

From the Lichess puzzle database, released into the public domain, five million positions from real games, each rated by hundreds of players. We keep the best-rated ones and map Lichess's themes onto our motifs. The lessons are the only thing written by hand, and the tactics from your own games are found by Stockfish on our servers.

Why not just use Lichess?

You should, for volume: it is free and excellent. What it does not do is tell you which motifs you miss in your games, put them on a path, and bring them back a day later. That is what this is for, and the free tier is generous enough that you never have to leave for a rating.

Who makes it

A small independent studio that also builds ChessEndings, ChessOpenings and DarkSquares.