TacticBarTL;DR: Among puzzles that require exactly two solver moves, quiet moves average 1,956 and defensive moves 1,860. The overall two-move average is 1,391.
Comparing two-move puzzles
If I simply ranked every theme by rating, themes with longer solutions would get an unfair advantage. Restricting the comparison to two-move puzzles is a crude control, but it makes the comparison a bit less bad.
Quiet moves
Checks and captures give you somewhere obvious to start. In a quiet-move puzzle, the first move may not look tactical at all. There may be no forcing move to put at the top of your candidate list.
The data cannot prove that quietness causes difficulty. Tags overlap, and puzzle ratings come from the players who attempt them. But even after comparing puzzles of the same length, the difference is still large.
Defensive puzzles are difficult too
Chess puzzles train us to expect an attack. A position whose answer is prophylaxis, simplification or the only saving move works against that expectation. Defensive moves appear in 1.08% of puzzles below 1,200 and 22.38% of puzzles rated 2,400 or higher.
Forks and mates show up at lower ratings
Among two-move puzzles, forks average 1,165 and mating puzzles average 1,111. I do not think that means forks or mates are inherently easy. More likely, their simplest and most recognisable forms are well represented at lower ratings.
Method and limitations
I analysed the official Lichess export dated 10 September 2026. Theme labels are generated by Lichess and a puzzle can carry several, so these groups overlap. Holding length at two moves removes one major confounder, not every difference between positions.
Download the derived JSON or read the companion analysis of how themes change by rating band.