Use these procedures as task-oriented reference. They describe public capabilities and verification steps, not private implementation. UI labels can vary by version.
Use the missed position, not only its score.
Identify whose turn it is and the immediate threat.
Compare the attempted move with the suggested line.
Find the defensive resource that changes the result.
Replay without the answer visible after a break.
Record the motif to revisit later.
You can explain why the move works and what your original choice missed.
Do not equate puzzle rating with an official playing rating or memorize only the first move.
Choose a positional exercise rather than a forced-win puzzle.
Assess king safety, piece activity and pawn structure.
Generate two candidate plans.
Choose a move and review the feedback.
Compare the resulting position, not just a numerical evaluation.
Repeat the position later and explain your choice.
The practice tests judgment instead of searching for a tactic in every position.
An engine evaluation depends on the position and analysis conditions; it is not a promise of improvement.
Record app/version, device/OS, selected mode, input conditions and the smallest failing step. Distinguish expected behavior from the observed result. Use synthetic examples and omit credentials or private records.
Documentation index · Capabilities and limits · Support and sources