How to Analyze Your Own Chess Games

Most players finish a game, click the engine button, nod at the coloured bar, and learn nothing. The method coaches teach is the reverse: record what you were thinking, analyse the game with your own head, and only then let the engine check your work. It is slower, and that is exactly why it works.

Why the engine goes last, not first

The engine tells you what the best move was. It cannot tell you why you didn't play it, and that second question is the only one that improves your chess. If you open the evaluation graph straight after the game, you outsource the thinking, feel briefly informed, and repeat the same mistake next week.

This is not a new idea. Mikhail Botvinnik, the sixth world champion, built the modern school of chess training on annotating your own games with full effort before anyone — or anything — else weighs in. He called chess the art of analysis, and he meant your analysis, not a machine's. The engine is a proofreader. You still have to write the essay.

Capture your thoughts while they're still warm

Do this within a few hours of the game, the same day at the latest. Your memory of what you were actually thinking decays fast, and once it's gone the most valuable data from the game is gone with it.

Play through the game once, quickly, and write short notes at the moves that mattered:

• What you calculated and what you rejected
• Moves where you felt unsure, uncomfortable, or short of a plan
• Where you spent a long time — a five-minute think is a flag even if the move was fine
• What you thought the evaluation was at three or four points in the game

Two or three sentences per moment is enough. You are not writing a book; you are preserving evidence.

Find the three to five critical moments

You do not analyse all forty moves. A typical game between improving players turns on three to five moments, and your job is to find them. Look for:

• The point where the opening ended and you had to think for yourself
• Any move where the initiative or the pawn structure changed hands
• Every capture or exchange you allowed or offered
• The moment you first felt worse — or first felt winning
• The move you regretted the instant your hand left the piece

Mark these positions before doing anything else. The rest of the method operates on this shortlist, which is why a proper analysis takes 30 to 45 minutes rather than half a day.

Analyse the critical moments with your own head

Now the real work. Set up each critical position and treat it as a fresh problem, as if it were from someone else's game. Spend five to ten minutes per position. Ask three questions:

• What were the candidate moves? List at least three before calculating any of them.
• What did my opponent want? If you can't answer this, you've likely found the actual mistake — most blunders are failures of attention, not calculation.
• What was the position asking for? Attack, defence, exchange, improvement of the worst piece. Name it in words.

Write your conclusion down: the move you now think was best and why. Committing to an answer in writing is the step most people skip, and it's the step that makes the engine check useful rather than passive.

Then, and only then, the engine

Switch the engine on and check your shortlist — not the whole game. For each critical moment, compare its top choice against your written conclusion. Three outcomes are possible:

• You agreed. Good. Your evaluation of that type of position is sound. Move on.
• You disagreed and the engine's move makes sense once you see it. This is gold. Work out which of your three questions you answered wrongly, and note it.
• You disagreed and the engine's move makes no sense to you. Don't force it. A line you can't understand at your level is not a lesson; it's noise. Flag it for a coach or stronger player and move on.

Ignore small evaluation swings. A drift from +0.4 to -0.2 decides nothing at club level. Care about the moments where a clearly better position became equal or lost.

Keep a mistake journal

One analysed game teaches you a little. Twenty games logged in the same notebook teach you who you are as a player. After each analysis, write one line per real mistake: the date, the position or move number, and — most importantly — the category of error. Useful categories:

• Missed the opponent's threat
• Calculated a line but stopped one move too early
• Wrong exchange — traded a good piece for a passive one
• Played fast in a critical position
• Drifted with no plan after the opening
• Time trouble caused by earlier slow play

GM Axel Smith popularised this as a List of Mistakes, and the principle is simple: you cannot fix an error you haven't named. After ten to fifteen games, read the whole list in one sitting. One or two categories will dominate. That is your training plan, handed to you by your own games.

Turn the pattern into one habit

Don't try to fix five weaknesses at once. Pick the category that appears most often and attach a single concrete habit to it, applied in your next games:

• Missing threats? Before every move, say the opponent's last move out loud (in your head, over the board) and name what it changed.
• Stopping calculation early? Add one forced move to the end of every line before you commit.
• Moving fast at critical moments? When a capture or pawn break is possible, sit on your hands for a full minute.

Run one habit for ten games, check the journal to see whether that category shrank, then pick the next. This loop — play, analyse, log, adjust — is the whole engine of improvement, and no amount of puzzle-solving replaces it.

Which games, and how often

Analyse your slow games — anything from about 15 minutes per side upwards, and every over-the-board game you play. Blitz and bullet aren't worth the full method; your mistakes there are mostly reflexes, not decisions. A quick 5-minute scan for tactical blunders is plenty.

Losses teach the most, but don't skip wins: many are full of mistakes your opponent failed to punish, and those errors are the sneakiest because the result hides them. Draws where you pressed and failed are often the richest of all.

One properly analysed game per week beats seven games skimmed with an evaluation bar. If a student brings a coach one game with honest written notes, the lesson practically runs itself — which is exactly why coaches assign this before anything else, and why tools like Grandmaster Gains' Academy build assignment tracking around it.

Common questions

How long should analysing one game take?

About 30 to 45 minutes for a slow game: a quick replay with notes, five to ten minutes on each of three to five critical moments, a short engine check, and a journal entry. If you're spending two hours, you're analysing too many moves; if you're done in ten minutes, you've only run the engine.

Should I analyse blitz games?

Not with the full method. Blitz mistakes are mostly reflexes and time-pressure guesses, so deep analysis of them teaches little. Give blitz games a five-minute scan for outright blunders, and save the proper method for games of roughly 15 minutes per side or longer.

What if I can't remember what I was thinking during the game?

Analyse sooner — same day, ideally within a couple of hours. If the memory is already gone, treat the critical positions as fresh puzzles and record what you'd play now; that's still far better than passive engine review. For over-the-board games, jotting clock times on your scoresheet helps reconstruct where you struggled.

Do I need the engine at all?

Yes, but only as a final check on the moments you've already analysed yourself. It catches tactical oversights your own review missed and confirms or corrects your evaluations. What it can't do is diagnose why you erred — that part only comes from comparing its verdict against conclusions you wrote down first.

Where to go from here

Everything above is the practical layer — the part that decides whether a club survives, whether a child keeps playing, and whether a coaching business pays. It is also the part almost nobody writes about, because it is less interesting than opening theory and far more useful.

If you are teaching, the two things that move the needle fastest are structure and measurement: a session that has the same shape every week, and a number each student can watch go up. Neither requires you to be a strong player. Both require you to be organised, and organisation is the thing you can actually control.

The rest of this Library covers the chess itself — twenty-one openings written from both sides, seven complete guides to tactics, endgames, checkmate patterns and strategy, and twenty variants with a guide each. All of it is free, and all of it is playable on a live board, which matters more than reading it.

More on coaching and scholastic chess

• Blitz vs Rapid vs Classical: What Each Trains
• Does Chess Count as Screen Time?
• Best Chess Books for Kids, Matched to Level