Puzzle ratings and game ratings often diverge, sometimes by hundreds of points. That is not a bug. They measure different skills in different conditions.
What puzzles measure
A puzzle gives you a position, a hint that there is a tactic, and no clock pressure. The task is to calculate a forced line. Strong puzzle solvers are good at pattern recognition, candidate moves and verification. Those skills matter in games, but they are not the whole game.
What games measure
A game requires you to create the tactical opportunity, manage the clock, handle opening theory, and stay focused for dozens of moves. A player who sees the tactic instantly in a puzzle may never reach the position in a game because the opening plan was wrong.
Why the gap runs both ways
- High puzzle, lower game: strong calculation, weaker opening/positional play or time management.
- High game, lower puzzle: good practical play and experience, but less exposure to artificial tactical positions.
Neither rating is "true". They describe different samples of your chess.
How to train both
Use puzzles to build calculation and pattern recognition. Use games to build plans, clock habits and opening knowledge. Review your games to see which puzzle themes actually appear in your losses. Then choose puzzles that match those themes.
On ChessUrge, the puzzle set is seeded with 15,000 hard Lichess puzzles and supplemented with puzzles generated from your own analysed games. That combination connects the clean pattern to the messy game.
Do not chase a puzzle rating. Chase the types of positions you keep getting wrong.
See the whole game, not just the blunder.
Choose a review depth, see the maximum cost, and get move-by-move accuracy with blunder, mistake and inaccuracy markers.