What does my kid's Chess.com accuracy score actually mean?
The short answer: accuracy measures how closely your kid's moves matched what the engine would have preferred, averaged over that one game. It is not a skill rating — it swings heavily with game length, opponent strength, and how complicated the position got. A single number can't be judged good or bad without that context. Trends over many games at the same time control mean something. Any single game's score mostly doesn't.
What the number actually measures
After each game, Chess.com compares every move your kid played to the engine's preferred move. A bigger gap between what was played and what the engine wanted is a bigger hit to the score. Moves that closely match the engine's choice score high. Average those gaps across the game and you get a percentage.
The number swings with conditions, not just skill. Short, quiet games — where nothing is ever particularly wrong — score high. 90-plus is not unusual when both sides play 20 solid moves and someone resigns before anything dramatic happens. Long, sharp, complicated games give the engine more chances to disagree, and the score drops even when the chess is excellent.
Context matters more than the number itself. Beating a weaker opponent who blunders on move eight can produce 95%+ without your kid doing anything special. A hard-fought loss in a sharp position — where both sides navigated real complications — can score lower than a lopsided win where nothing was ever at stake.
The move labels, translated
Chess.com labels every move after the game: Brilliant, Great Move, Best, Excellent, Good, Inaccuracy, Mistake, Blunder, Miss. Most parents look at the red ones. Here is what they actually mean:
- Brilliant — a sacrifice or resource the engine approves of; genuinely rare and worth noticing when it appears
- Great Move / Best / Excellent / Good — fine; don't overread the differences between these labels
- Inaccuracy — a small slip; the position is still basically the same
- Mistake — real cost; the evaluation shifted meaningfully
- Blunder — changed the likely result of the game
- Miss — a winning move was available and wasn't played
One blunder is often all it takes. A 92% game with a single blunder in the critical moment loses; an 80% game with no blunders can win. The count of blunders per game, tracked across 20 or 30 games at one time control, is more informative than any single accuracy score.
What the numbers can't tell you
The accuracy score tells you something happened. It does not tell you why — whether your kid moved too fast, miscalculated, had no plan, or got rattled. It doesn't tell you whether it's the same mistake recurring across games. And it doesn't tell you what to practice.
Two kids with identical 82% accuracy can need opposite fixes. One blunders in time pressure late in the game; the other drifts in the middlegame after the opening runs out. The label says "blunder" for both. The diagnosis — and the practice — are completely different.
The numbers are a symptom log. They say something went wrong here. Reading across many games and asking what those moments have in common — that is the diagnosis. That is the hard part, and it is genuinely hard to do from move labels alone.
How to actually use it as a parent
Ignore single-game accuracy. The variance is too high and the context too different from one game to the next to draw conclusions.
What holds up over time:
- Track blunder count per game across 20-plus games at the same time control — this trends more cleanly than accuracy percentage
- Compare only the same time control to itself; 10-minute accuracy and 3-minute accuracy are measuring different things
- After a bad game, ask "was this the same kind of mistake as last time?" — that question is hard to answer from move labels alone
- Look at when in the game the blunders happen — early (nerves or opening), middlegame (planning), late (stamina or time pressure) each has a different cause
The last question — whether a mistake is recurring — is where reading across the full game history earns its keep. It's also where a single game's accuracy score is least helpful.
Quick answers
Is 85% accuracy good on Chess.com? It depends entirely on game length, sharpness, and the opponent. The same quality of play can score 75 in a sharp tactical game and 92 in a quiet one. Trends over 20-plus games at one time control mean something. Single games don't.
Why did my kid get high accuracy and still lose? Accuracy measures per-move closeness to the engine, averaged across the whole game. One blunder decides the result. An otherwise clean game with a single blunder at the critical moment can score 90% and still lose.
Should we try to raise the accuracy number? Not directly. The number is a side effect of what your kid does. Train the habit that causes blunders — moving too fast, not checking for opponent threats — and the number follows. Chasing the score without addressing the underlying habit doesn't work.
Where ChessIntel fits: the move labels tell you a mistake happened. They don't name the habit behind it. ChessIntel reads across all the games, finds the recurring pattern — the specific kind of mistake that keeps showing up — and describes it in plain language, checked against a real engine. The first report is free. Related: why your kid's online rating runs higher than their USCF rating.