A game review is more useful than a win-or-loss verdict. A winning game can contain a weak decision, and a lost game can contain a good practical choice. Reconstruct what you were thinking, locate the few positions where your plan changed, and only then use external analysis to test a specific question. This routine gives you a lesson you can remember instead of a long list of computer moves.
| Rank | a | b | c | d | e | f | g | h |
|---|---|---|---|---|---|---|---|---|
| 8 | a8, empty | b8, empty | c8, empty | d8, empty | black rook on e8 | f8, empty | black king on g8 | h8, empty |
| 7 | a7, empty | b7, empty | c7, empty | d7, empty | e7, empty | f7, empty | g7, empty | h7, empty |
| 6 | a6, empty | b6, empty | c6, empty | d6, empty | e6, empty | f6, empty | g6, empty | h6, empty |
| 5 | a5, empty | b5, empty | c5, empty | white knight on d5 | e5, empty | f5, empty | g5, empty | h5, empty |
| 4 | a4, empty | b4, empty | c4, empty | d4, empty | e4, empty | f4, empty | g4, empty | h4, empty |
| 3 | a3, empty | b3, empty | c3, empty | d3, empty | e3, empty | f3, empty | g3, empty | h3, empty |
| 2 | a2, empty | b2, empty | c2, empty | d2, empty | e2, empty | f2, empty | g2, empty | h2, empty |
| 1 | a1, empty | b1, empty | c1, empty | d1, empty | e1, empty | f1, empty | white king on g1 | h1, empty |
Record your own view first
After the game, write the result, time control, and three moments: when you felt comfortable, became uncertain, and felt the result change. Add the move you wanted and the reply you expected. Do this before opening an engine so the review captures your decision process.
Do not annotate every move. “I thought the knight had two defenders” is useful because it gives you a testable belief. “The computer dislikes this” does not explain what to change next time.
- What was my plan in this phase, and which move was supposed to support it?
- What did I calculate, and what reply did I fail to consider?
- Was the problem a tactic, a rule misunderstanding, a plan, or a time decision?
Find decision points, not every inaccuracy
Mark a move when it changed something concrete: material, king safety, pawn structure, or whose pieces had the initiative. Then look one move earlier. The move that lost material is often only the final symptom; the important decision may be the earlier move that placed a piece on a vulnerable square or rejected a forcing reply.
At each marked position, write two or three human candidates before comparing them: the move you played, one you considered, and one quiet move that improves safety. For each, write the opponent’s most forcing reply. This keeps the review about choices rather than one mysterious best move.
Worked example: do not blame the last move
Imagine this opening sequence: 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O. White later loses a pawn after 9. h3 and writes, “h3 threw away my advantage.” That conclusion is too quick: the move may be useful, harmless, or a lost tempo depending on the plan that followed.
Compare 9. h3 with 9. d4 and ask what White expected after each. Does 9. d4 open the center before Black is ready? Does 9. h3 prevent a piece from using g4, or spend time while Black improves? The answer is about the position, not the eventual pawn loss.
Exercise: write the engine question
Return to the position after 8...O-O. Before using any tool, choose between 9. h3 and 9. d4. Write: “I expect my move to achieve ___, and Black’s strongest reply may be ___.” Which sentence is easier to verify?
Answer: either sentence works if it names a concrete expectation. A focused analysis request is: “After 9. d4, what is Black’s strongest reply, and does White’s center remain sound?” If 9. d4 was fine, perhaps you avoided an active plan; if there was a tactic, record the missed reply. Do not turn an incomplete position into a universal opening rule.
Turn one finding into the next drill
Choose one recurring cause and make a tiny assignment. If you missed attacks, pause after every opponent move and name the changed squares. If you rushed a central break, set up three positions and compare candidates. If time caused the error, review the clock decision separately from the chess move.
Use an engine or a stronger player after your own review to test a specific position, not to outsource the entire game. Never use live external assistance in a rated game. At your next review, check whether the same cause appeared; improvement is a repeated decision, not a single accuracy score.
Questions and answers
Should I use an engine before reviewing my own game?
No. First reconstruct what you thought, the move you wanted, and the reply you expected. Then use external analysis to test a focused question; a computer’s preferred move alone does not explain what to change.
How do I find the critical moment in a chess game?
Mark positions where material, king safety, pawn structure, or initiative changed, then look one move earlier. The final lost piece or pawn may be a symptom of an earlier decision that created the vulnerability.
Further reading: Chess.com: How to Analyze Your Chess Games