There's a pattern in almost every online chess account: a loss, followed by "New Game" within seconds, followed by another loss in the same way. Hundreds of games later, the rating hasn't moved. It's not a lack of talent. It's that none of those games were ever looked at again.
Chess has one big advantage over most games: everything that happened is written down in the move list. A focused 15-minute review of one game usually teaches more than five more games played on autopilot. Here's a routine that works.
Minutes 0–5: your own review first
Don't open the engine yet. If you see the engine's verdict first, you'll nod along and forget it by tomorrow. Finding your own mistakes is what builds memory.
Replay the game from the start and answer three questions:
- When did I first feel uncomfortable? There's usually a moment where "something felt off." That's the most valuable spot to study.
- Which moves did I play without thinking? Automatic, fast moves are where errors hide.
- What did my opponent do that surprised me? Surprises mean your understanding of the position was incomplete.
Note the move numbers of two or three moments. That's all you need.
Minutes 5–10: ask the engine the right question
Now open the analysis. On ChessDigi, the Review game button runs Stockfish through every move, grades each one from Best to Blunder, and lists the moves that cost you the most.
Don't read every comment. Do two things:
- Compare the engine's list with yours. If it flags the same moment you noted, your intuition was right — you just didn't find the solution. If it flags a move you thought was perfectly normal, you've found a blind spot. Blind spots are the most valuable thing a review can show you.
- Focus on blunders and mistakes only. Inaccuracies almost never decide games below 1800. One or two big errors usually do.
Minutes 10–13: classify the mistake
For each big error, ask why you played it. Nearly every mistake falls into one of four categories:
| Type | What it looks like | How to fix it |
|---|---|---|
| Missed threat | The opponent had a tactic you didn't see | Ask "what does their last move threaten?" before every move |
| Calculation error | You saw the idea but miscounted the result | Solve puzzles; calculate to the end before moving |
| Planning error | No tactic, but the position slowly got worse | Study pawn structures and middlegame plans |
| Time pressure | The error happened with little time left | Play slower games; budget your clock |
At club level, "missed threat" dominates. A classic example:
Minutes 13–15: write one lesson
End with one sentence you'll take into your next game. Not "play more carefully" — that changes nothing. A good lesson is specific and actionable:
- "When their bishop and queen both aim at a square next to my king, I check that square before anything else."
- "I don't push my g-pawn after castling kingside unless I have a concrete reason."
- "Under one minute on the clock, I choose safe moves over ambitious ones."
Write it down — a notebook or a note on your phone. After twenty games, some lessons will repeat. Those repeats are your real weaknesses, and they're what to train next.
Review your wins too
Wins deserve a look as well, because many wins are really the opponent's mistakes. The engine often shows that you were lost on move 18 if your opponent had found one strong move. Knowing that protects you from false confidence.
When the engine isn't the lesson
Sometimes the engine's best move is a ten-move sequence no human would find, keeping a tiny edge. That isn't the lesson for you. The lesson is usually the simplest move that doesn't lose — often the engine's second or third choice. Use the evaluation as a compass, not a map.
Building a mistake journal
The single lesson you write after each game becomes far more valuable when you keep them together. A simple journal needs only four columns:
| Date | Game | Type of mistake | Lesson |
|---|---|---|---|
| Mon | 15+10 vs. computer, Medium | Missed threat | Check the f7 square whenever their bishop is on c4 |
| Wed | 10+5 online | Time pressure | Keep 3 minutes for moves 30–40 |
Every two or three weeks, read through it. Two things usually jump out:
- One type of mistake dominates. For most club players it's missed threats. That tells you to spend more of your training on tactics and on the "what does their move want?" habit.
- The same lesson appears again and again. That's a habit that hasn't stuck yet. Write it on a sticky note next to your screen for the next week of games.
The journal takes thirty seconds per game and turns scattered losses into a clear training plan.
Common review mistakes
- Reviewing only with the engine running from move one. You learn much less than when you think first.
- Studying every variation. Chasing deep side lines turns a 15-minute review into an hour of confusion.
- Reviewing only losses to blitz games. Reviewing slow games teaches more, because you actually had time to decide.
- Never writing the lesson down. Without a written note, the same mistake returns next week.
Start now
If you just finished a game, try the routine on it right away. If not, play the computer on ChessDigi at a level where you win about half your games, then press See why you lost or Review game when it ends. For what the grades mean, read blunder, mistake, inaccuracy.
Frequently asked questions
Do I need to review every game?
No. Review slow games, painful losses, and games where one moment clearly decided the result. Two or three reviews a day are plenty.
Should I use an engine while playing?
Never in a game against a human — that's cheating. Even against the computer, using an engine during the game means you learn nothing. Use it only after the game.
How long should a review take?
About 15 minutes for a 10–15 minute game. Longer reviews tend to drift into deep variations that aren't useful yet. The goal is one lesson, not total understanding.
Is it better to review with a stronger player than with an engine?
When you can, yes — at least sometimes. A stronger player can explain the plan behind a move and ask what you were thinking, which an engine can't. The ideal combination is your own review first, the engine to check tactics, and occasionally a coach or clubmate to discuss the ideas.
What if I can't understand why the engine prefers a move?
Play the position out against the computer from that point, trying both your move and the engine's suggestion. Seeing how each line develops usually makes the reason clear.
