How to analyze a chess game online: 7 steps that teach you more than a red Blunder badge

Finished a game and the engine says Blunder? Good — but that is only the start. Use this 7-step MoveForge workflow to understand your decisions and turn mistakes into training.

The engine found a mistake. Great. What did you actually learn?

You finish a game, open the analysis and a few seconds later the verdict arrives: an inaccuracy here, a mistake there, then the position jumps from +1.2 to -3.4. An arrow shows the engine move. You click forward. Another arrow. Another number.

And this is where it is easy to fall into a trap. Watching an analysis is not the same thing as analyzing your game.

If you close the review remembering only that Stockfish preferred a knight move instead of a bishop move, but you still do not know why you made the bad decision, the training value is limited.

At MoveForge we prefer a different workflow. Break the game down like material in a forge: your decisions first, critical moments second, engine third. At the end, something concrete should remain to train.

Step 1. Start with one of your real games

You do not need a masterpiece from an elite tournament. Some of the best training material is already sitting in your own game history.

MoveForge Game History can work with your public games from Chess.com and Lichess. Choose a game you actually remember, preferably one containing a moment you did not understand while playing.

It might be:

• a loss from a position that felt good,
• a win where the opponent suddenly dropped material and hid your earlier problems,
• a strange opening that made you uncomfortable from the beginning,
• an endgame where you had no idea what plan to choose,
• or simply a game that left you thinking: what actually happened here?

Do not review only losses. A win may contain several serious mistakes that your opponent simply failed to punish.

Step 2. You first. No Stockfish yet.

This is the most important step and probably the easiest one to skip.

Replay the game from the beginning without asking the engine for the best move. Do not try to prove that you played well. Reconstruct what was genuinely happening in your head.

At important moves ask yourself:

• What was I trying to achieve?
• What reply did I expect?
• Which other candidate moves did I consider?
• What was I worried about?
• Did I calculate a line or did the move simply look natural?

If you look at the position now and think, “that was obvious”, be careful. A board seen after the game is not the same board you saw under pressure. Reconstruct your thinking before you reveal the engine answer.

Step 3. Mark 2–4 positions where you actually had to make a decision

Not every move deserves five minutes of investigation. If you made a natural developing move and the engine slightly prefers another reasonable move, there is no need to turn it into a crime scene.

We are looking for decision points.

Stop when:

• you had several reasonable options,
• the position suddenly became tactical,
• you did not know what plan to choose,
• the opponent's move completely surprised you,
• you spent a long time thinking and still felt unsure,
• or you moved instantly and knew seconds later that something felt wrong.

If you extract three positions like this from one game, you already have strong training material. You do not need thirty red icons.

Step 4. Now bring in the engine

This is where your idea meets computer analysis.

Do not stare only at Stockfish's first move. Start by checking whether the engine agrees with your assessment of the position. Were you actually better? Was the attack real? Was the exchange you avoided really bad?

An engine evaluation is an assessment of the position, not a simple material counter. Instead of obsessing over tiny differences such as +0.4 and +0.6, concentrate on meaningful evaluation swings and the reason behind them.

If your move turns a calm position into a clearly worse one, do not stop at the word “mistake”. Explore the continuation and identify the mechanism: lost material, a tactical shot, king safety, a weak square, a bad ending or simply giving away the initiative.

Step 5. Read the graph as the story of the game, not a stock-market chart

MoveForge Game Review uses the evaluation graph to show where the character of the position changed. Alongside it you can inspect accuracy, move classifications and the move list.

The goal is not to panic every time the line moves.

Use the graph like a map:

• a long quiet section — the character of the position probably remained broadly stable,
• a sudden jump — stop here and find out what changed,
• several drops in a row — the problem may not be a single move but a poor plan or a difficult position you failed to stabilize.

A move label helps you locate a position worth checking. It is not the full explanation. “Blunder” tells you that the move seriously damaged the position. Your job is to find out why you played it and what you failed to notice.

Step 6. Go back one move before the Blunder

This is one of my favorite analysis tricks.

The review shows a huge red Blunder badge? Do not immediately ask for the best move. Go back to the position and make the decision again.

Check:

1. Which candidate moves can you find now?
2. What is the opponent threatening?
3. Which pieces are undefended or overloaded?
4. Are there checks, captures or forcing moves?
5. What changed after the opponent's previous move?

Only after your own attempt should you reveal the engine line.

If the best move looks completely unnatural, do not memorize it. Follow the continuation and find the idea. Maybe it removes a piece with tempo. Maybe it protects a key square. Maybe it prepares a queen trade. Maybe the opponent simply runs out of good replies.

A move without understanding is an answer to a test. The idea behind the move is training material.

Step 7. Turn the conclusion into training or you will probably repeat it

This is the step missing from many quick reviews.

Do not finish with “I should have played the rook”. Finish with one concrete problem.

Examples:

• I do not check the opponent's threats before starting my own idea,
• I regularly leave pieces undefended,
• I start flank play before finishing development while my king is still unsafe,
• I fail to recognize a pin on an important file or diagonal,
• I move too quickly in endings that require accurate pawn-race calculation.

If the critical position is useful to replay, turn it into a task and return to it later in Tactical Forge. This time you are not looking at the answer — you have to find the idea again on the board.

The cycle becomes:

game → self-analysis → engine → cause of the mistake → training task → another attempt.

That is when your game history starts working for your next result.

The weakest possible analysis? “The engine says this move is better.”

An engine is an extremely powerful tool, but a number on the screen does not repair your decision-making process.

After a useful review, you should be able to finish this sentence:

“I played badly because…”

Not: “…Stockfish showed a different move.”

Instead:

“…I focused entirely on my attack and never checked the opponent's reply.”

Or:

“…I rejected the trade because it looked passive, but the resulting ending was actually favorable.”

That is knowledge you can carry into the next game.

Important: analyze after the game, not during it

When you play another person on an external chess platform, use engine analysis only after the game has finished. Engine assistance during ongoing live games conflicts with the fair-play rules of major chess platforms.

In that situation, use MoveForge as the post-game workshop: break the position down, understand the mistake and prepare the next training session.

Open your game history

Turn a position into training