How to create puzzles from your own chess games

Puzzle databases are excellent, but they are someone else's mistakes. Your own games contain the positions you actually misjudged, with your own habits and time pressure baked in. That makes them unusually efficient training material.

How generated puzzles work

When a full-game review finds a clear mistake, ChessUrge stores the position before the move, the move you played, and the engine's preferred continuation. That becomes a puzzle. You see the position, find the best move, and compare it with what you did in the game.

For the player who owns the game, the puzzle is filed as fix mistake. For other players, it appears as find tactics. The position is the same; the framing is different.

Why the threshold matters

Not every difference from the engine is a useful puzzle. A 20-centipawn preference is not a training position. ChessUrge only generates puzzles from larger losses, so the set stays focused on real errors rather than engine nitpicks.

A simple routine

1. Play a game and run a full review at a moderate depth.
2. Open the generated puzzles for that game.
3. Solve each one without moving the pieces first. Say the candidate moves out loud.
4. Compare your answer with the game move and the engine line.
5. Write down the motif or the blind spot.

Repeat the same puzzles a few days later. If you still miss them, the pattern is not learned yet.

Do not over-generate

Twenty focused puzzles from your last ten games are more useful than two hundred from every game you have ever played. Review the games that mattered, generate the positions that surprised you, and let the rest go.

Start from your game library, then open Learn to solve the generated set.