Solving many exercises does not guarantee lasting improvement. Progress appears when every error provides information about a specific decision. Two mistakes in the same position can have different causes: one comes from missing a threat, while another comes from calculating an incomplete line.
Useful review does not look for blame or memorize answers. It reconstructs what you saw, identifies what was missing and defines one small action for the next session. This turns your mistake history into a personal training map.
Annotated position
Worked example: diagnose the error instead of memorizing it
White to move. Find a candidate that attacks two targets and record the defence you expect before revealing the line.
Be6 creates a fork involving the rook on c8 and another critical target. Black answers by moving the rook to e8 and the bishop then captures on f5. If you never saw Be6, the error was candidate generation; if you saw it but stopped after ...Re8 without calculating Bxf5, it was calculation. The same position needs a different remedy depending on where the process broke.
Example selected from the Rush64 Chess practice catalogue.
1. Record your first impression
Before checking the solution, write down what you thought was happening. Record your chosen candidate, the threat you wanted to create and the reply you expected. This information fades quickly after seeing the correct move because hindsight makes the solution appear obvious.
The note does not need to be long. A sentence such as “I looked for mate and missed that my queen was attacked” preserves the essential part of the process. The goal is to compare your mental model with the reality of the position, not to write a perfect annotation.
2. Classify the type of error
A simple classification separates perception, generation, calculation, evaluation and execution errors. In perception, you missed something on the board; in generation, you never considered the correct candidate; in calculation, you omitted a reply; in evaluation, you misjudged the outcome; and in execution, you moved too quickly despite understanding the idea.
This distinction matters because each category needs a different remedy. More timed puzzles will not fix poor endgame evaluation, just as studying theory will not correct a habit of moving without checking forcing moves.
- Did I see every immediate threat?
- Did I include the correct move among my candidates?
- Did I calculate the opponent’s best reply?
- Did I evaluate the final position correctly?
- Did I move before completing my final check?
3. Reconstruct the variation without help
After understanding the solution, close the explanation and replay the line from the initial position. Explain why your move fails and why the alternative works. If you only recognize the answer when you see it, you have not yet absorbed the tactical relationship behind it.
Change one detail of the position in your imagination: remove a defender, move the king or swap the colour of a piece. Ask whether the combination still works. This separates understanding from literal memory and helps transfer the pattern to future positions.
4. Write an operational rule
Turn the lesson into an instruction you can execute. “I must concentrate more” is too vague. “Before calculating my attack, I will check every opposing check” describes observable behaviour. Operational rules should be brief and tied to the exact moment when they are needed.
Keep only a few active rules. If you try to remember ten corrections during a game, none will receive enough attention. Choose one for the next week and use it in every challenge. Once it becomes automatic, replace it with the next priority.
5. Repeat after a delay
Reviewing the same position immediately tests short-term memory, not learning. Return to it after one or more days, mixed with different exercises. If you find the idea and can justify it, the pattern is beginning to settle. If you fail again, record whether the cause is the same or has changed.
Spaced repetition works best when you do not know the theme in advance. A mixed collection forces you to diagnose the position, while a set labelled “pins” already reveals part of the answer. Alternate thematic sessions for learning a technique with mixed sessions for practising recognition.
6. Measure decisions, not only correct answers
Accuracy is useful, but it does not tell the whole story. Also track how often you performed your final check, how long it took to find the first candidate and which error categories are declining. Your process can improve before the score reflects it consistently.
A good session ends with one concrete conclusion and enough energy to train again. Stop when attention drops and moves become impulsive. Consistent short, focused sessions produce more learning than occasional piles of exercises completed without review.