Why is accuracy missing on some of my chess.com games?

chess.com only calculates accuracy when a game goes through Game Review. Until that happens the game has no accuracy value at all. Open the game on chess.com, run Game Review, and the score will exist from then on. Tools that read chess.com's public data can only show the accuracy number that has already been computed, though Rookbook can run its own free move-by-move engine review of any game regardless.

What does chess accuracy actually measure?

It compares your moves against a strong engine's preferred moves and compresses the result into a score out of 100. It is heavily affected by the type of position: quiet positions score high for everyone, sharp tactical positions score lower, and it does not distinguish one game-losing blunder from many small inaccuracies.

What is a good chess accuracy score?

There is no universal figure, because faster time controls produce lower accuracy for every player. The meaningful comparison is against your own accuracy in the same time control over time, and against your opponent's accuracy in the same game.

rookbook

Analyze my games · Reading · Glossary

Why is accuracy missing on some chess.com games?

Some of your games show an accuracy score and some show nothing at all. It isn't a bug, and there's a specific reason — along with a few things the number does and doesn't tell you. Rookbook shows yours across your whole history.

Why some games have no accuracy score

chess.com doesn't calculate accuracy for every game automatically. It's computed when a game goes through Game Review, and until that happens the game has no accuracy value attached — not a zero, not a low score, nothing.

So a blank isn't a bad result. It means nobody has run the review on that game yet. Open the game on chess.com, run Game Review, and the number appears from then on. Any tool reading chess.com's public data, including Rookbook, can only show the accuracy number chess.com has already computed. What Rookbook can do is run its own free engine review of the game — blunders, mistakes and missed mates, move by move in your browser — even when the accuracy field is blank.

If you're looking at your history and half the accuracy column is empty, that's normal for most accounts. Reviews are run per game, usually only on the ones people care about.

Rookbook's game log with accuracy and opponent accuracy columns, several rows showing ? because those games never went through Game Review. — The blanks aren't zeros or bad scores. Those games haven't been through Game Review, so no accuracy figure exists for them.

What accuracy actually measures

Accuracy compares your moves against what a strong engine would have played, then compresses the result into a single 0–100 figure. Higher means your moves tracked the engine's preference more closely.

Two things follow from that, and both matter more than people expect:

It's position-dependent, not just skill-dependent. A quiet, locked position where few moves change the evaluation will score high almost regardless of who's playing. A sharp tactical melee punishes small errors heavily. The same player scores differently in different games without playing better or worse.
It doesn't distinguish a fatal mistake from a harmless one. Twenty tiny inaccuracies can cost you more accuracy points than one blunder that lost the game outright. Accuracy is an average, and averages hide the moment that decided the result.

What's a good accuracy?

Less useful a question than it sounds, because the honest answer is "compared to what." Faster time controls produce lower accuracy for everyone — there's less time to find the right move — so a blitz figure and a rapid figure aren't comparable numbers.

The comparison that means something is against yourself in the same format, over time, and against the opponents you actually face. Rising accuracy in rapid over a few hundred games is real improvement. A single high-accuracy game is mostly a quiet position.

A more useful way to read it

Rather than chasing a target number, the two comparisons worth making are:

Your accuracy against your opponent's, in the same game. Winning while scoring lower than your opponent usually means they made one decisive error and you didn't — which is a different story from outplaying them steadily.
Your accuracy in wins against your accuracy in losses. If they're close, your results are being decided by single moments rather than by overall quality, and reviewing those moments beats grinding general study.

Rookbook shows both automatically across your history, alongside the games where accuracy is missing so you know what's included in the average. And for the blanks, Rookbook's own free game review fills the gap: an engine in your browser walks the game move by move, no chess.com membership required.

Free, no account, and your games never leave your browser.

Review any game move by move, free, on rookbook.net →

Common questions

Why is accuracy missing on some chess.com games? Because that game hasn't been through Game Review. Run the review on chess.com and the number will exist from then on — or open the game in Rookbook's free Review tab and get a move-by-move engine verdict without a membership.

Does higher accuracy always mean I played better? No. Quiet positions inflate it and sharp ones deflate it, so the same quality of play can produce very different scores across games.

Why is my blitz accuracy so much lower than my rapid? Less thinking time produces more inaccuracies for everyone. Compare within a time control, never across.