Improvement · 2 min read
How to analyse your own games, before the engine
An engine's evaluation graph teaches little on its own. Write down what you thought, then compare it with what was true. The gap is what you study.
The four-pass routine
- From memory: note your plan at moves 10, 20 and 30, and mark every move you were unsure about.
- Turning point: find the one or two moves that decided the game.
- Classify the mistake: a missed tactic, a bad trade, a weakened structure, poor clock use or an opening gap.
- Engine check: confirm your conclusions, starting with the moments you marked.
Review your scoresheets
Over-the-board games are the most useful to review, because you had time to think. Photograph the scoresheet, turn it into PGN, and step through it with move insights on.
Write one sentence per game
Name the mistake that mattered most, as a rule. 'I traded my active bishop to avoid a threat that was not real' is useful. 'Blundered on move 24' is not. Reread the list before each tournament round.
Key takeaways
- Annotate from memory before you open the engine.
- Find the turning point rather than listing every inaccuracy.
Frequently asked questions
- How long should analysing one chess game take?
- About twenty to thirty minutes for a serious over-the-board game. Blitz games rarely need more than a quick look for repeated patterns.
- Should I analyse games I won?
- Yes. Wins hide mistakes that a weaker opponent missed and a stronger one will punish.
Practise the related tactics
The ideas behind this
- Evaluating PositionsMaterial, kings, pawns, piece activity: the checklist that gives a plan.
- Prophylactic ThinkingBefore every move, ask what your opponent wants to do.
- Memorizing Key LinesLearn the ideas first, then memorise only the forcing lines.