How to analyse your own chess games

Most players "analyse" a game by switching on an engine, watching the bar jump around and clicking through the red moves. Ten minutes later they have forgotten almost all of it. The engine told them what was wrong, but not why they chose the wrong move, and the why is what decides your next game.

The method below takes 20 to 30 minutes per game. It has three stages, and the order matters: your own thinking first, the engine second, patterns last.

Stage 1: Review the game without help

Do this as soon as possible after the game, while you still remember what you were thinking. A game you analyse the next morning is already half reconstructed from the moves rather than from your memory.

Play through the game on a board or in the analysis board with the engine switched off, and write short notes at three kinds of moments:

• Where you spent a long time. Long thinks usually mean the position was critical or you were unsure. Write down what you were choosing between.
• Where you felt the game turn. "I thought I was better here, and five moves later I was worse." Mark the move where that feeling changed.
• Where you were surprised. Any opponent's move you did not consider is worth a note, even if it turned out to be harmless.

Keep the notes plain: "Wanted to play Nd5 but was afraid of ...c6 and ...b5." These notes are the raw material for the whole review. Without them, you can only learn what the right move was. With them, you can learn what went wrong in your thinking.

Then try to find improvements yourself. Pick the two or three moments you marked as most important and spend a few minutes on each, as if it were a puzzle. You will often find the answer on your own, and a lesson you find yourself sticks far better than one an engine gives you.

An example from the opening

Suppose you had Black here and lost quickly after 6.Nxf7 Kxf7 7.Qf3+. In Stage 1 your note might say: "I recaptured on d5 automatically. Didn't think about f7." That one sentence already contains the lesson. The knight on d5 is loose, the king is stuck on f7 after the sacrifice, and 5...Na5 is the normal way to avoid this. You do not need an engine to understand that; you need to admit that the recapture was a habit, not a decision.

Stage 2: Bring in the engine

Now run a full review. The free game review gives each move a label, shows the evaluation graph and lets a coach explain the key moments in plain language. Use it to check your Stage 1 notes, not to replace them.

Work through it in this order:

1. Compare the turning points. Does the engine agree with where you thought the game changed? If you felt the game turn on move 24 but the graph shows the real drop on move 19, you misjudged a position, and that is more important than any single move.
2. Look at the biggest drops first. Blunders and mistakes are where games are decided. An inaccuracy that costs a tenth of a pawn rarely matters at club level.
3. For every big error, ask what you missed. Was it a tactic (a fork, a pin, a back-rank weakness)? A plan? A move by your opponent you did not see? Write the answer in one word if you can: "tactic", "king safety", "time", "calculation", "didn't look".
4. Play out the engine line. Do not just read "Best move: Rd8". Make the move on the board and play the next few moves until you understand why it works. If you cannot see why, ask the coach for an explanation or set up the position and try the alternatives yourself.

A diagram like this is the kind of thing Stage 2 often reveals. A player loses a rook endgame, and the review shows that ten moves earlier there was a back-rank mate available, or that their own king was in the same danger and they never made luft with h3 or g3. The engine finds the move; your job is to file it under "back rank" and remember it.

Be careful with tiny differences

If the engine prefers one move at +0.6 and your move gave +0.4, that is not something to study. Both are fine. Focus on moves where the evaluation changed by a pawn or more, or where the result of the game changed: a win became a draw, a draw became a loss.

Stage 3: Find the patterns

A single game tells you what went wrong once. Ten reviewed games tell you what goes wrong regularly. This is the stage most players skip, and it is the one that actually changes results.

Keep a simple log. A spreadsheet or a notebook works. For each game record:

• the opening and whether you were comfortable after it,
• the number of serious mistakes and blunders,
• a one-word cause for each,
• how much time you had left at the critical moment.

After ten games, count the causes. A typical club player's list looks something like this:

• Missed opponent's tactic: 7
• Low on time: 5
• Unfamiliar opening position: 3
• Bad endgame technique: 2

Now you know what to work on, and it is rarely what you expected. Many players believe they need more opening theory, then find that most of their points disappear to simple tactics in the middlegame.

Turn the patterns into training

• Missed tactics: solve themed puzzles on the motifs you missed. If you keep missing knight forks, train forks specifically in puzzle themes. Daily mixed puzzles keep the general skill sharp.
• Opening trouble: if the same opening keeps leaving you worse by move 12, drill the first moves in opening training and study one model game in that line.
• Time trouble: look at where the time went. Usually it is one or two very long thinks in the middlegame, not slow play throughout.
• Endgames: pick the type of endgame you lost and learn its basic positions before learning anything else.

A short checklist

Before you close a game, you should be able to answer these questions:

1. What was the most important moment of the game, and did I recognise it during the game?
2. What was my worst move, and why did I play it?
3. Did my opponent's moves surprise me, and which ones?
4. Was there one idea I could have known in advance (an opening plan, an endgame technique)?
5. What one thing will I do differently next game?

If you only answer question 5, you have still done more than most players.

Common mistakes in analysis

• Only reviewing losses. Wins hide mistakes too. Your opponent may have missed your blunder; the next one will not.
• Reviewing too many games badly. One game reviewed properly beats five games clicked through.
• Arguing with the engine. If you think the engine is wrong, test it: play your move and let it show you the refutation.
• Ignoring the clock. A move played with 10 seconds left is a different problem from one played after 10 minutes of thought.

Analysis is a skill like any other. The first few reviews feel slow. By the twentieth, you will see the same kinds of errors in your games before the engine points them out, and that is the moment you start making fewer of them.

Published 2026-11-01 · 22Softwares