5 min read
A practical post-game routine
Analyze fewer positions. Learn more from each one.
A useful game review is not a tour through every engine line. It is an investigation: what were you trying to do, what changed, which reply did you miss, and what cue could help you see the same idea next time? This routine keeps the engine in the role of evidence rather than answer key.
Start with your own thinking
Record what you noticed, what you calculated, and where you felt uncertain before an engine changes the memory.
Investigate turning points
Spend time on decisions that changed the game, not on automatically labeling every harmless inaccuracy.
End with a cue
Turn the lesson into something observable: a loose piece, an open file, a forcing reply, or a missing defender.
On this page
Pass one: replay the game without an engine
Go through the moves at a comfortable pace and stop wherever your decision felt difficult. Write down the candidate moves you remember considering, the reply you expected, and the part of the position that worried you. If you cannot remember a calculation, that is useful information too: the move may have been chosen from habit rather than a clear reason.
Do not try to prove that the result was deserved. A win can contain a serious miss, and a loss can contain a strong plan. The purpose of the first pass is to recover the decision process while it is still available.
- Mark uncertaintyFlag the moments where you spent time, changed your mind, or made a move because no alternative looked clear.
- Name the planA short note such as “trade queens” or “attack the pinned knight” makes later comparison much more concrete.
Pass two: find the positions that deserve analysis
Now add the evaluation and move grades. Look for a sharp swing, a missed tactic, a forced defensive move, or a position where your plan and the engine's preferred move point in different directions. Those are investigation candidates, not automatic verdicts.
Keep the shortlist small. One opening decision, one middlegame turning point, and one late-game technique are enough for a productive session. If several moves repeat the same mistake, study the earliest clear example and treat the rest as evidence of a pattern.
- Prioritize changed outcomesA move that turns a win into equality or equality into a loss deserves more attention than a tiny numerical dip.
- Include a good decisionKeep one position where your reasoning worked so the review reinforces a process worth repeating.
Compare candidate moves, not just the top line
Before opening the principal variation, ask which two or three moves belong on a sensible shortlist. Checks, captures, direct threats, defensive necessities, and moves that improve the least useful piece are good places to begin. Then compare what each candidate allows the opponent to do.
The engine's first choice is evidence that a move works under search. It does not tell you why another natural move fails, or whether the continuation is easy for a human to find. Comparing candidates exposes the exact branch where your original reasoning stopped too early.
Translate the line into a board-specific reason
After the line is visible, describe it without evaluation notation. Name the piece, square, file, diagonal, pawn break, or tactical relationship that makes the move work. “The knight on e5 has one defender and two attackers” is easier to retrieve later than “best move, +1.4.”
If the explanation could be pasted onto almost any position, it is not finished. Replace broad advice such as “improve king safety” with the concrete fact: which checking route exists, which defender is overloaded, or which file will open.
Finish with one action for the next game
Close the review by choosing one cue that can be checked over the board. Examples include “after every opponent move, ask what became loose,” “before a pawn push, scan the diagonal it vacates,” or “when the queens face each other, calculate the recapture before assuming a trade.”
Save the full line if it helps, but practice the cue. The goal is not to remember a tree of moves forever. It is to notice the feature that tells you when calculation is needed.
- One sentenceIf the lesson needs a paragraph, it will be hard to use with the clock running. Compress it to one observable trigger.
- One follow-upFind or create a similar position later in the week and test whether the cue brings the right candidates to mind.
Common questions
Clear answers before you try it.
Should I analyze every chess game?
Analyze games that contain a decision you want to understand. A short review of most serious games and a deeper review of selected games is more sustainable than treating every move as homework.
Should I use the engine immediately after a game?
First replay the important moments from memory. Then use the engine to test that reasoning. Opening it immediately can replace your original thought process before you have examined it.
How long should a game review take?
There is no required duration. Stop when you can explain the critical position, the missed or successful reply, and one cue to reuse. A focused 15-minute review can be more useful than an hour of passive line clicking.
What if the engine move makes no sense?
Compare two natural alternatives and follow the opponent's strongest or most likely reply after each. The difference usually reveals the hidden tactical or positional job of the engine move.
Try the routine on a board, not in a spreadsheet.
Play a free coached game or paste a PGN, then open the positions that changed the game and compare the likely replies.