How to Analyze Your Own Chess Games (and Turn Mistakes into Training)
Training Tips

How to Analyze Your Own Chess Games (and Turn Mistakes into Training)

ChessPecker Team
September 21, 2026
9 min read

Most players finish a game, glance at the accuracy score, and queue up the next one. That is the single biggest waste in chess improvement. Your own games are the only source that tells you which patterns you miss, under your time controls, in your openings. Puzzles are the gym. Your games are the diagnosis that tells you what to lift.

Key takeaway: Analysis is not about finding the engine's best move. It is about finding the reason you didn't play it, and then training that reason away.

"The loss is already paid for. Analyzing it is how you get your money back."

Step 1: Write Down Your Thoughts Before the Engine

Right after the game, before you turn on any engine, spend five minutes going through the moves on your own. At each moment that felt important, note what you were thinking: what you were afraid of, what you calculated, and what you missed seeing entirely.

This step matters because the engine only tells you what was wrong. Your notes tell you why. "I didn't see the knight fork" and "I saw the fork but thought it lost a piece" are completely different problems with completely different fixes.

Tip: Keep it short. Three or four sentences per critical moment is plenty. Your goal is to capture your thinking while it is still fresh, not to write a book.

Step 2: Find the Critical Moments

A 40-move game usually has only three to five moments that decided it. Look for:

  • Evaluation swings: Moves where the position went from fine to losing (or winning to drawn)
  • Long thinks: Anywhere you spent a lot of clock time. Uncertainty is a signal.
  • Captures and checks you didn't consider: Forcing moves are where tactics hide
  • The transition moments: Leaving the opening, trading into an endgame, or your opponent's first real threat

Step 3: Now Turn On the Engine, and Ask Why

With your own notes written, check the critical moments against an engine. For every mistake, don't just read the best line. Ask: what did the engine see that I didn't? Play the best line out on the board until the idea is clear in words, like "the rook on the open file lands on the 7th rank with tempo."

Manifest Chess Lite showing a PGN file of 24,002 games, the board, and five Stockfish 18 engine lines side by side
Multi-line engine analysis next to the board, shown here in Manifest Chess Lite. Seeing several candidate moves at once helps you understand why the top move is best. (Image: Manifest Chess)

Looking at several engine lines at once is more useful than looking at one. If the top three moves all share an idea, like attacking the same weak pawn, that idea is the lesson.

Step 4: Label Every Mistake

This is the step almost everyone skips, and it is the one that makes analysis pay off. Give every real mistake a category:

Mistake TypeWhat It Looks LikeHow to Train It
Missed tacticYou never saw the fork, pin, or mateRepetition drilling on that theme
Calculation errorYou saw the idea but miscounted the lineHarder puzzles, solved without moving pieces
Opening problemBad position by move 12, every gameReview your repertoire file
Endgame techniqueWinning position converted into a drawEndgame-themed puzzle sets
Time troubleGood position lost with seconds on the clockTimed drills to build speed

After ten games you will have a tally. For most club players, "missed tactic" wins by a mile. That tally is the most valuable number in your chess life, because it tells you exactly where your training time should go.

Step 5: Save Your Annotated Games

Keep every analyzed game in a PGN file (the standard text format for chess games that every chess app can open), with your notes as comments. Months later, that file becomes a personal textbook: your own recurring mistakes, in your own openings, explained in your own words.

On a Mac, a dedicated PGN reader opens the file instantly and writes your comments straight back into it, so your notes live in the file rather than in one app's private storage.

Annotating a chess game in Manifest Chess Lite with comments, variations, and move symbols
Comments, variations, and move symbols saved back into your own PGN file. (Image: Manifest Chess)

The Tools: What to Use for Game Analysis

You don't need expensive software to do this well. Pick the tool that fits how you play and what computer you use.

Free, in the Browser

  • Lichess analysis board: Free, with a strong engine and PGN import. The best zero-cost option for most players.
  • Chess.com Game Review: Convenient if you play there. Free accounts get a limited number of full reviews.
  • ChessPecker analysis board: Stockfish running in your browser, with a verdict on each move you play. The engine is a Premium feature.
The free Lichess analysis board with Stockfish running on an Italian Game middlegame position
The free Lichess analysis board: paste a PGN or a position and Stockfish runs right in the browser. (Image: Lichess)

Desktop Apps

A desktop app is worth it once you have hundreds of games and want everything in one place, offline, and on your own machine.

Manifest Chess Lite app icon

Manifest Chess Lite (Mac): a fast, offline PGN reader. Double-click a PGN file and it opens on the board, with Stockfish 18 (up to five lines) and annotations saved back into the file. No account and nothing sent anywhere. One-time purchase on the Mac App Store (€19 at time of writing).

Manifest Chess Studio app icon

Manifest Chess Studio (Mac): the full database version. Keeps every game you own in searchable databases with position search, syncs your Lichess and Chess.com games automatically, and queues them for review. One-time purchase (€129 at time of writing).

Game review in Manifest Chess Studio paused at a critical moment, asking the player to find a better move for White
Studio's game review stops at a critical moment and asks you to find the better move yourself, which is Step 3 of this routine done for you. (Image: Manifest Chess)
  • ChessBase '26: The long-standing professional standard on Windows, with the largest ecosystem of databases and training material. €199.90 for the program alone, €249.90 for the Starter Package.
  • En Croissant: A modern, free, open-source database for Windows, Mac, and Linux.
  • SCID vs. PC: A long-standing free, open-source database for Windows, Mac, and Linux. Older interface, lots of features.
ToolPlatformCostBest For
Lichess analysisBrowserFreeQuick review of any game
Chess.com Game ReviewBrowserFree, limitedGames played on Chess.com
Manifest Chess LiteMac€19 one-timeFast offline PGN reading and annotation
Manifest Chess StudioMac€129 one-timeA personal database with automatic game sync
ChessBase '26Windows€199.90 program only, €249.90 Starter PackageProfessional preparation
En CroissantWin / Mac / LinuxFree, open sourceModern open-source database
SCID vs. PCWin / Mac / LinuxFree, open sourceLong-standing free database

Step 6: Turn Your Mistakes into a Training Set

Analysis without follow-up is just reading. If your tally says "missed tactic," the fix is to see that tactic so many times that you can't miss it again. That is exactly what the Woodpecker Method is for.

  • Check Theme Mastery: Your ChessPecker theme breakdown should confirm what your game analysis found. If you keep missing forks in games and your fork accuracy is low, you have found your weakness from two directions.
  • Build a targeted set: Create a puzzle set focused on that theme, at a difficulty slightly below your rating so you can solve fast.
  • Drill it in cycles: Solve the same set repeatedly, getting faster each time, until the pattern is automatic.
  • Re-check your games: After a month, analyze ten new games and redo the tally. The category you trained should shrink.
Weekly routine: Analyze your two or three most important games each weekend (Steps 1 to 5, about 20 minutes each). Spend weekday sessions drilling the theme your tally points to. Analysis picks the target; drilling hits it.

Frequently Asked Questions

Should I analyze every game?
No. Analyze losses, draws you should have won, and games at your main time control. Bullet games rarely teach enough to be worth 20 minutes.

How long should analysis take?
15 to 30 minutes for a rapid or classical game. If it takes longer, you are probably reading engine lines instead of looking for the reason behind each mistake.

Is the engine's accuracy score useful?
As a rough mood indicator, sure. But a 92% accuracy game with one missed mate matters more than an 80% game full of small inaccuracies. Look at the critical moments, not the average.

What is a PGN file?
Portable Game Notation, the plain-text format for chess games. Every chess site lets you download your games as PGN, and every analysis tool can open them.

What is the best chess database for Mac?
ChessBase is the standard, but it is Windows software and doesn't run natively on macOS. The native Mac options are Manifest Chess Studio (€129, one-time) and the free, open-source En Croissant and SCID vs. PC. Studio is the easiest if you want Lichess and Chess.com sync and position search without setup; En Croissant is the pick if you want something free.

What is the best PGN reader for Mac?
For opening and annotating PGN files, Manifest Chess Lite (€19, one-time) is a fast native option: double-click a file and it opens on the board with Stockfish 18, and your annotations save back into the same file. Lichess's import is the free alternative if you don't mind working in a browser.

Start With Your Last Loss

Open your most recent loss, write down what you were thinking, find the moment it turned, and label the mistake. Then go train that mistake until it's gone.

Ready to fix what your games reveal? Check your Theme Mastery on ChessPecker, build a set around your weakest theme, and start your first cycle.

// TOPICS

how to analyze chess gameschess game analysispost-game reviewPGN readerchess database Maclearn from chess mistakes

// APPLY_KNOWLEDGE

Ready to Put This
Into Practice?

> Start training with the Woodpecker Method on ChessPecker today. Build lasting pattern recognition.