How to Read a Chess Game Review and Turn It Into Drills

What the four labels in a game review actually mean

A game review runs every move you played through a chess engine. The engine compares your move with the move it would have chosen. It measures how much the evaluation dropped and puts a label on the gap.

The evaluation is shown in pawns. A score of +1.0 means White is better by roughly the value of a pawn. A negative score favors Black. When a move is flagged, the review is telling you that its drop was large compared with the alternatives.

Accuracy

Accuracy is a single score that summarizes how close your moves stayed to the engine's choices over the whole game. Each site uses its own formula, so an accuracy number from one platform can't be compared with one from another.

It also depends on the game itself. A quiet, simplified position is easier to play accurately than a sharp tactical fight. Treat accuracy as a mood indicator for one game, not a grade.

Blunder, mistake, inaccuracy

• Blunder: the biggest drop. Usually a piece left hanging, a tactic missed, or a mate allowed. It often changes the result.
• Mistake: a clear drop that hands over a real advantage, without necessarily deciding the game.
• Inaccuracy: a smaller slip. Your position gets worse, but you're still in the game.

The exact cutoffs between these labels aren't standardized, and they vary from one site to another. Some reviews also weigh the context: losing material when you're already completely winning may be labeled more gently than the same loss in a balanced position. For the exact rules, check your platform's help page. For a fuller look at what a free review covers, see our guide to free chess game reviews.

How to read the review without fooling yourself

The trap is to scroll through the labels, nod at the engine's suggestions, and close the tab. You learn very little that way. Use this order instead.

1. Play through the game yourself first, without the engine. Note the moves where you felt unsure.
2. Open the review and find the turning point: the move where the evaluation swung the most. That's usually your blunder, and it matters more than everything else in the game.
3. Look at the clock. Check how much time you had when you played each flagged move. A blunder with ten seconds left is a time-management problem, not a calculation problem.
4. Before reading the engine's suggestion, cover it and try to find it yourself. Then compare.
5. Ask why. If you can't explain in one sentence why the engine's move is better, play the line forward until you can.

Leave the accuracy score for the end. It can't tell you what to fix. The flagged moves can. For a step-by-step routine covering the whole game, our move-by-move analysis method goes deeper, and this guide to the analysis board shows how to test lines without letting the engine think for you.

Turn each flagged move into a concrete drill

Every flagged move is a position you got wrong. That makes it training material cut to your own game: it comes from your openings, your habits and your blind spots. Here's how to drill each label.

Blunders: rebuild the habit that failed

First, name the type. Did you leave a piece undefended, miss your opponent's threat, or miss a tactic of your own?

Then set up the position just before the blunder and play it out against the engine from your side until you can hold it. Save the position in a list. Re-solve it a few days later, then again a week after. Finally, add one line to your pre-move checklist that would have caught it, such as “What does their last move attack?”

Mistakes: find the move and the reason

Set up the position, hide the engine, and give yourself a few minutes to find the engine's choice. Write the reason in one sentence. “Trade the bishop because their knight has no squares” is a reason. “The engine says so” is not.

Then play the position from the correct move against the engine a few times. Watch how the advantage gets converted.

Inaccuracies: drill only what repeats

One inaccuracy is noise. The same inaccuracy in five games is a weakness. Compare your move with the engine's and name the theme: piece activity, pawn structure, king safety or development.

Keep a tally across games. Only when a theme keeps showing up does it earn a drill, usually a short session on model positions built around that theme.

A worked example: one game, three drills

Here's a made-up case to show the process. Your review flags one blunder, two mistakes and several inaccuracies.

• The blunder: on move 24 you pushed a pawn and left your knight undefended. You had plenty of time on the clock. Diagnosis: no threat check. Drill: save the position, re-solve it later, and add “Is every piece defended?” to your checklist.
• The two mistakes: both came in the middlegame, when you traded off your active bishop. Diagnosis: you trade pieces when you feel uncomfortable. Drill: play both positions from the engine's move and write down why keeping the bishop was better.
• The inaccuracies: mostly slow developing moves in the opening. Drill: none yet. Note the theme and watch for it in your next games.

Result: two drills worth doing today and one theme to watch. That's a review you can act on. To fit this into your week, see our simple weekly routine.

When one game isn't enough: find the pattern

A single review shows what went wrong in one game. It can't tell you whether that error is a one-off or a habit. For that, you need to look across many games and count how often each type of error comes back. You can do it by hand with a spreadsheet, but it takes time.

That's the job Rookprint was built for. The Report takes up to your last 200 online games, analyzes them with Stockfish, and uses AI to write up your 3 recurring weaknesses with numbers, 50 exercises built from your own mistakes, and a 30-day plan. On the day of writing, the pricing page lists the Report at $29. Progress, at $49, adds 100 exercises and a second report 30 days later that compares before and after. You can check the format on the sample page. It isn't a coach: it won't explain moves live or watch you play, and the exercises only help if you actually do them.