How to Analyse Your Own Game Without Wasting an Hour
Most engine reviews are read once and forgotten. A repeatable method that produces two or three things you will actually remember, in about fifteen minutes per game.
· 3 min read
Running a game through an engine takes ten seconds and produces a page of numbers. Most people read it, nod at the blunder, and learn nothing. The problem is not the tool, it is that the output is a list of corrections and corrections are not memorable.
What follows is a method that produces two or three specific things per game. It takes about fifteen minutes.
Before you switch anything on
Play through the game once and write a sentence at every point where you had to think. Not analysis, just what you believed at the time: "I thought his king was weak here", "I did not know what to do so I improved a piece", "I calculated the sacrifice and it looked unclear".
Four or five sentences is enough. This step is the one people skip and it is the one that makes the rest work, because it converts the review from "here is a better move" into "here is where your thinking was wrong". The second version is what you remember in your next game.
Then run the engine
Now analyse, and resist reading the summary. Go straight to the list of your moves sorted by how much each lost.
| Look at | Why |
|---|---|
| Your three largest losses | The whole content of the game |
| The first move where equality ended | Games are decided before the obvious moment |
| Any position you wrote a sentence about | Where your judgement is testable |
That is between four and eight positions, and several will overlap.
Work each position properly
For each one, do three things in order.
Look at the position before you look at the answer. Give yourself two minutes to find something better than what you played. If you find it now, the problem was time or attention rather than knowledge, and that is a different fix.
Play the refutation out on a board. Not read it. The line the engine gives after your move is the reason your move was bad, and the reason is usually four or five moves in. Reading notation does not put it in your memory; watching the pieces arrive does.
Name the pattern in words. "My knight on f6 was the only defender of h7." "I opened the file my opponent's rook was already on." A named pattern transfers to other positions. A remembered move does not.
What to write down
One or two lines per game, in a file you actually revisit:
- The pattern you missed, named.
- The judgement that was wrong, if your sentences from step one turned out mistaken.
After ten games you will have a list, and the list repeats itself. That repetition is the finding. Nobody discovers their real weakness from one game, because in one game every mistake looks like bad luck.
What to skip
The accuracy percentage. It is a summary of a summary and tells you nothing you cannot get from the move list.
Inaccuracies. They are frequently the engine preferring something you had no reason to consider.
Opening moves that were theory. If the engine dislikes move six of a mainline, either the analysis is shallow or you are looking at a position where humans have already decided the engine's preference is impractical.
Positions after the game was decided. Once you are a rook down, how you played the remaining moves is not the lesson.
The one habit worth building
Analyse your wins as often as your losses. A won game frequently contains the same error as a lost one, with an opponent who failed to punish it. Only reviewing losses means only seeing the errors that were exploited, and those are a biased sample of the errors you make.
Common questions
- Should I analyse a game before or after running an engine?
- Before, always. Write down what you thought was happening at three or four points while you can still remember your reasoning. The engine then tells you where your reasoning was wrong, which is far more useful than a list of better moves with no context.
- How long should analysing one game take?
- Fifteen to twenty minutes for a game you played seriously. Beyond that you are usually re-reading the same conclusions. Depth of attention on three positions beats shallow attention on forty.
Try it on your own position
The analysis board runs the same engine in your browser, free. The app adds what a web page cannot: reading a position straight from a photograph, saved boards, and analysis with no connection.
