A practical guide to what ChessIQ measures, why labels can change with deeper analysis, and how review becomes training — without exposing proprietary internals.
What is ChessIQ actually measuring?
It grades each decision against stronger engine alternatives, then summarizes game quality and decision quality in a readable way.
Why can results change during deeper analysis?
Because deeper search can reveal tactics or defenses that were not visible earlier. ChessIQ refines results over time instead of pretending the first pass is final.
What does ChessIQ turn into training?
Missed opportunities and major mistakes from your analyzed games become puzzle positions in your local training lanes.
ChessIQ is not trying to guess talent or predict your rating. It measures move quality in context: what the position allowed before your move, what changed after your move, and how large that swing was.
Move labels (such as Best, Good, Inaccuracy, Mistake, Blunder, or Miss) are interpretation layers over engine evaluation data. Accuracy is a summary signal for decision quality, not a full description of everything that happened in the game.
In ChessIQ, the pre-move evaluation represents the position you had to solve. That is where best-move alternatives come from.
The post-move evaluation represents the position you actually created. That is what powers swing charts, practical outcome context, and many critical-moment signals.
Initial review is a fast pass. After that, continuous analysis can keep deepening the focused position. As depth improves, best lines, evaluations, and sometimes labels can be updated.
ChessIQ intentionally gates harsher re-labeling behind stronger stability support so transient shallow reads are less likely to flash misleadingly.
Best move means the strongest continuation found for the pre-move position at the current analysis strength.
Accuracy summarizes move quality over the game and de-emphasizes low-information situations like forced sequences and routine moves in already decided positions.
Critical moments are high-impact swings, missed conversions, initiative flips, and similar turning points surfaced from move-by-move analysis signals.
After review, ChessIQ can generate training positions from your own game history. The current puzzle lanes separate missed opportunities from blunders/major mistakes so practice stays targeted.
Puzzle priority is adaptive and local-state-driven (attempt history, repeats, due timing, and recurrence), so weak spots can resurface until they stabilize.
Core analysis runs in your browser using Stockfish WebAssembly. Core review history, training state, and many preferences are stored locally (including IndexedDB/local storage paths used by the app).
External calls are still used for public-game imports (Chess.com/Lichess) and production analytics, so local-first should be read as a trust boundary — not an absolute "no network ever" claim.
Use Privacy for storage/service boundaries and FAQ for common interpretation questions.