Play against a statistical shadow of anyone in the world.
Enter any chess.com username. This pulls that player's real blitz games, learns their opening choices and measures their blunder rate, then plays against you with a real chess engine (Stockfish) calibrated to a strength you choose — making your kind of mistakes at roughly the rate you actually make them.
Build a shadow
What this shadow learned
Games analyzed
0
Opening positions learned
0
Measured blunder rate
0%
Top first move
—
Opening repertoire
Building…
Time usage
Building…
Loading chess engine…
Heads up — Stockfish didn't load.
The game still works, but it's now using the simple built-in engine, which plays noticeably weaker chess. Refreshing the page often fixes this.
Engine strength18 / 20
On strength: this uses real Stockfish, not a hand-tuned approximation — but this particular build only exposes an older 0–20 "Skill Level" dial rather than a calibrated Elo target, and there's no precise, verified mapping from this dial to FIDE or chess.com ratings. Higher is stronger; adjust it to taste rather than trusting an exact number from us.
What's actually personalized: the opening book (this player's real most-played moves) and the blunder rate (measured from their real games) are genuine data. The engine strength dial and move quality in between are Stockfish, not that player specifically — Stockfish doesn't know how to play "like" someone, only how to play at a given strength.