Most players analyse a lost game by turning on the engine, spotting the move marked as a blunder, and closing the tab. That feels productive, but almost nothing transfers to the next game. A useful review is a short investigation: you form an opinion first, then let Stockfish test it.
1. Replay the game without an engine
Open the game and click through the moves at your own pace. Stop at three moments: the point where you left your preparation, the point where you first felt unsure, and the point where the evaluation swung. Write down what you were thinking. This takes five minutes and gives the review a purpose.
If you cannot remember your thinking, that is already a finding. It usually means the moves were automatic rather than planned.
2. Name the mistake before the engine does
At each turning point, ask a concrete question. Did I miss a tactical resource? Did I trade into a worse endgame? Did I spend too much time on a quiet position? The engine will give you a better move, but only you can explain why you did not consider it.
3. Run a depth-controlled review
A full-game review should grade every move, not just the loudest blunder. On ChessUrge's game library you choose a target depth, see the maximum credit cost, and get accuracy plus blunder, mistake and inaccuracy markers for both sides. The important part is the pattern, not the percentage.
- If most errors happen after move 30, your problem may be fatigue or time management.
- If errors cluster in one opening, the issue is structural knowledge, not calculation.
- If you consistently choose the second-best move in sharp positions, slow down and calculate candidate moves before touching the piece.
4. Turn the worst moment into a puzzle
Analysed games can generate puzzles from the positions where you went wrong. A puzzle is better than a note because you have to solve it again, from the same position, without the engine line in front of you. ChessUrge files these as fix mistake puzzles for you and find tactics puzzles for other players.
5. Write one sentence, not ten
End the review with one actionable sentence: "In quiet positions, check my opponent's candidate captures before choosing a plan." One sentence you remember is worth more than a page of engine variations you will never reread.
The next step is to play another game with that sentence in mind, then review again. Improvement is a loop, not a single analysis session.
Frequently asked
How long should a game review take?
For a club game, 10 to 20 minutes is enough: five minutes without the engine, five to ten minutes checking the turning points, and one minute writing down the lesson.
Should I review every game?
Review the games where you were unsure or lost, and the games you won but did not understand. Blitz games can be reviewed quickly; classical games deserve a deeper pass.
See the whole game, not just the blunder.
Choose a review depth, see the maximum cost, and get move-by-move accuracy with blunder, mistake and inaccuracy markers.