Engine literacy guide

Chess accuracy, evaluation and move grades explained.

The numbers are useful when they lead you back to the board. Learn what each signal means, where it can mislead, and which positions deserve a closer look.

9 minute readUpdated July 19, 2026For players at every level

A game review contains several layers: an evaluation of each position, a comparison between the played move and stronger alternatives, and a whole-game summary. Read them in that order. A single percentage should never erase the story of the game.

What a chess evaluation means

A numerical evaluation estimates which side has the better position with strong play. In the common convention, a positive value favours White and a negative value favours Black. A value near zero suggests approximate balance, not necessarily an easy draw.

+0.2Roughly balanced
+1.0White has a meaningful edge
-2.5Black has a large advantage
MateA forced mating sequence is seen

The number is not a literal pawn count. Material is only one factor. King safety, activity, space, pawn structure, initiative, tactical threats and endgame features all affect the position.

Evaluation is objective; difficulty is practical. A position evaluated at +1.5 may be hard to convert, while a nominally equal position can be unpleasant to defend. Always inspect the plan and the concrete line.

How to read mate scores

A mate score means a forced checkmate has been found against best defence. The displayed number may refer to moves or half-moves depending on the interface. The essential question is not the exact count: it is whether the mating sequence is forced and what tactical idea makes it work.

If you missed mate, look for the first position where checks, captures, or threats created the net. Training the pattern is usually more valuable than memorising the final sequence.

What an accuracy score tells you

Accuracy is a whole-game summary of how closely a player's moves preserved the quality of the position. A high number usually means the player avoided costly losses. It does not automatically mean the game was difficult, creative, or played at a particular rating level.

Interpret accuracy with context:

  • Game length: a short game provides fewer meaningful decisions.
  • Position difficulty: forced recaptures and obvious moves are different from complex choices with several plausible candidates.
  • Starting advantage: preserving a winning position and defending an inferior position present different practical challenges.
  • Sample size: one game is evidence about that game, not a stable measure of your overall strength.

Use accuracy to compare the two journeys through the game, then open the move list and study where the evaluation, plan, or tactical demands changed.

How move grades should be used

Move grades organise the review. Their names and exact boundaries can differ between products, so treat them as navigation labels rather than universal chess laws.

Grade familyPractical reading
Best / strongThe move preserves the strongest available continuation or stays very close to it.
Good / acceptableThe move remains playable but may allow a better plan, easier defence, or small concession.
InaccuracyThe position worsens enough to deserve attention, though the game may remain balanced or manageable.
MistakeA significant opportunity is lost or the opponent receives a substantial advantage.
BlunderA major tactical or positional failure changes the likely course of the game.
Great / exceptionalA particularly strong decision in a demanding context; the exact label depends on the review method.

A "best" move is not always the only reasonable human move. An "inaccuracy" may reveal a deep strategic issue or a small technical detail. Open the position before drawing a conclusion.

Why reliability matters

A summary is more trustworthy when the game contains enough relevant decisions and the important positions have clear, stable conclusions. Short games, missing phases, incomplete game records, or highly unstable tactical positions can reduce confidence in aggregate scores.

When a review reports insufficient data, that is useful information. A ten-move opening loss cannot provide credible evidence about endgame technique. Study the decisive opening and tactical choices without inventing conclusions about phases that never occurred.

The best order for reading a game review

  1. Start with the result and opening transition.Recall what you expected and when the game left familiar territory.
  2. Scan the evaluation flow.Locate turning points, recoveries, and stretches where the pressure steadily increased.
  3. Open the critical moves.Compare your candidates with the strongest replies and resulting plans.
  4. Check phase summaries.Look for a repeated weakness in opening decisions, middlegame plans, or endgame technique.
  5. Choose one training priority.Convert the most useful diagnosis into a board task or study theme.

Frequently asked questions

Does 90% accuracy mean I played like a grandmaster?

No. Accuracy describes move quality within one game and its positions. It is not a direct rating conversion. Difficulty, game length, opponent errors, and the number of meaningful decisions all matter.

Why can a winning move receive a low grade?

The move may still win while giving away a large part of the advantage or allowing unnecessary counterplay. Compare the resulting position and the opponent's strongest reply.

Why does a quiet move rank above a check or capture?

Forcing moves must be calculated first, but they are not automatically good. A quiet move may prevent the opponent's threat, improve the worst piece, or create a stronger threat that cannot be met.

Can two analysis tools give different grades?

Yes. They can use different search settings, alternative lines, thresholds, and classification methods. Compare the chess evidence in the position before comparing labels.

Read the numbers on a real board.

Upload a PGN to see the evaluation flow, critical decisions, alternatives, and coaching in context.

Review your game →