Chessy
An inaccuracy is a slightly suboptimal move — one that lets some advantage slip without being a clear mistake. It's the mildest of the three error grades engines use to review your games: inaccuracy, mistake, and blunder, in rising order of severity.
On this page
Where it sits on the scale
Analysis engines classify almost every non-best move on a spectrum. An inaccuracy costs a small amount of evaluation — often just a fraction of a pawn — and doesn't hand the opponent anything concrete. A mistake gives up more, and a blunder typically throws away material or the game outright. Most inaccuracies wouldn't even be noticed by a human opponent in real time.
Why they still matter
A single inaccuracy rarely loses a game, but a string of them adds up. Games between strong players are often decided not by one dramatic blunder but by whoever accumulates fewer small inaccuracies — imprecise piece placement, a slightly loose pawn move, a missed chance to improve the position. Reviewing where your inaccuracies cluster (openings, time pressure, quiet positions) is one of the most efficient ways to improve.
Reading an engine report
When a game-review tool flags a move as an inaccuracy, it's telling you a better move existed and the gap was measurable but small. It's worth pausing on these moments even though they didn't cost the game — they usually reveal a pattern (a plan you didn't consider, a resource you overlooked) that's cheaper to fix than a genuine blunder.
Frequently asked questions
Is an inaccuracy the same as a mistake?
No. An inaccuracy is milder — it loses a small amount of advantage. A mistake loses more, and a blunder is the most severe, often losing material or the game.
Can inaccuracies lose a game?
Rarely on their own, but several inaccuracies in a row can slowly erode an advantage until a real mistake or blunder becomes possible.
How do I reduce inaccuracies in my games?
Review your games with an engine and look for patterns — certain phases, piece types, or time-pressure moments where imprecise moves cluster — then drill those specific situations.