When you open a game review, every move gets a label: Best, Good, Inaccuracy, Blunder and so on. The labels are a quick way to see where the game was decided, but they are only useful if you know what they measure. This article explains each label, how a review decides between them, and why a "Brilliant" move is much more than a move that wins.
What the labels measure
A game review analyses every position with a chess engine and compares the move you played with the best move it found. The key question for each move is: how much did this move change your chances?
That change can be measured in pawns (centipawns) or, more usefully, in expected score: the share of points a player can expect from the position, from 0% (lost) to 100% (won). Expected score matters because the same loss in pawns can mean very different things:
- Going from +0.5 to -0.5 turns a pleasant position into a worse one. That is a real error.
- Going from +8.0 to +7.0 changes nothing. You are still completely winning.
So reviews usually judge moves by how much winning chance was lost, not by raw pawns. The chessGM game review works on the same principle and adds a coach's explanation for the key moments, so you see why a move earned its label.
The positive labels
Best
The move the engine considers strongest in the position, or one so close to it that the difference is negligible. Many positions have several moves that are practically equal; any of them can be marked Best.
Excellent
A move that is almost as good as the best one. You lost a tiny fraction of your winning chances, too little to affect the result. Strong players' games are full of Excellent moves.
Good
A sound move that keeps the position roughly where it was but is noticeably weaker than the best. It does not damage your game, but there was something better available.
Great
A Great move is a strong move that was hard to find, usually because it was the only move, or one of very few, that kept your advantage or saved the position. If every other reasonable move would have lost ground and you found the one that did not, that deserves more credit than Best. Typical examples are a quiet defensive move that refutes an attack, or the only move that holds a difficult endgame.
Brilliant
The rarest label. A Brilliant move is usually a sound sacrifice: you give up material, the move is among the best in the position, and it was not obvious. All three conditions matter, and the next section explains them in more detail.
The negative labels
Inaccuracy
A move that loses a small but real share of your chances. Your position gets worse, but usually not enough to change the expected result. One inaccuracy rarely decides a game. Several in a row often do.
Mistake
A move that gives away a significant part of your advantage, for example turning a clearly better position into an equal one, or an equal position into a clearly worse one. Mistakes are often positional (a bad trade, a weakened king, a misplaced piece) or tactical oversights that cost a pawn or the initiative.
Blunder
A move that changes the likely result of the game: a winning position becomes drawn or lost, or a drawn position becomes lost. Hanging a piece, missing a mate, or walking into a mating attack are typical blunders.
Note the asymmetry. If you are already completely lost, a move that loses another pawn may not be labelled a Blunder, because your chances were already close to zero. The labels measure how much the move changed the game, not how bad it looks.
White: pawn on f2, pawn on g2, pawn on h2, rook on d1, king on g1. Black: king on g8, pawn on f7, pawn on g7, pawn on h7. White to move.
How a brilliant sacrifice is judged
Players love Brilliant moves, so it is worth understanding what earns one. A good review asks several questions.
1. Was material given up? The move must offer or leave material that the opponent can take: a piece, the exchange, or occasionally a pawn in a striking way. A quiet move, however strong, is not a sacrifice. Trading a knight for a knight is not a sacrifice either.
2. Is the sacrifice sound? The move must be among the best moves in the position. A sacrifice that the opponent can accept and then win with is not brilliant. It is a blunder, however spectacular it looks. This is why reviews check the sacrifice against the engine's evaluation, not just whether you won the game.
3. Was it hard to see? A recapture or a sacrifice that wins immediately with a one-move mate is strong but not remarkable. Brilliant moves usually look risky at first glance and only work because of a deeper idea.
4. Was your position not already completely winning? If you are a queen up, sacrificing a rook for a quicker mate is pleasant, but many reviews do not give Brilliant for moves in positions that were already decided.
A classic example
White: bishop on c4, pawn on e4, knight on c3, knight on f3, pawn on a2, pawn on b2, pawn on c2, pawn on d2, pawn on f2, pawn on g2, pawn on h2, rook on a1, bishop on c1, queen on d1, king on e1, rook on h1. Black: rook on a8, knight on b8, queen on d8, king on e8, bishop on f8, knight on g8, rook on h8, pawn on a7, pawn on b7, pawn on c7, pawn on f7, pawn on h7, pawn on d6, pawn on g6, pawn on e5, bishop on g4. White to move.
The black bishop pins the knight on f3 to the queen. White plays 5.Nxe5!, apparently dropping the queen. If Black accepts with 5...Bxd1, White mates: 6.Bxf7+ Ke7 7.Nd5#.
White: bishop on f7, knight on d5, knight on e5, pawn on e4, pawn on a2, pawn on b2, pawn on c2, pawn on d2, pawn on f2, pawn on g2, pawn on h2, rook on a1, bishop on c1, king on e1, rook on h1. Black: rook on a8, knight on b8, queen on d8, bishop on f8, knight on g8, rook on h8, pawn on a7, pawn on b7, pawn on c7, king on e7, pawn on h7, pawn on d6, pawn on g6, bishop on d1. Black to move.
If Black does not take the queen, for example 5...dxe5 6.Qxg4, White has simply won a pawn. So the sacrifice is sound, it is the best move, and it looks impossible at first glance because it hangs the queen. That combination of conditions is what a Brilliant label rewards. This pattern is known as Legal's mate.
Now imagine the same sacrifice where Black has an extra defender covering d5. Then 5.Nxe5 would just lose material. The move would look identical, but the review would label it a Mistake or Blunder, because the idea does not work.
Other labels you may see
- Book: a known opening move. Reviews do not grade these, because they are established theory.
- Forced: the only legal move, or the only move that does not lose immediately. There was no choice to judge.
- Missed win: you had a winning move, often a tactic or a mate, and played something else.
How to use labels in your review
- Start with Blunders and Mistakes. They decide most games. Find out why each one happened.
- Do not chase Brilliants. They appear when the position offers them. Looking for sacrifices in every game leads to unsound ones.
- Learn from your Great moves too. If you found an only-move under pressure, understand why the alternatives failed. That builds the same skill for next time.
- Treat Inaccuracies lightly. At club level, a few small inaccuracies per game are normal. Fixing your blunders first is far more valuable.
- Practise the patterns. If your blunders keep coming from the same motif, work on it in puzzle themes.
The labels are shorthand for one idea: how much each move changed the game. Read them that way and the review becomes a map of exactly where your result was won or lost.
Published 2026-11-01 · 22Softwares