212,300 Chess Puzzles by Theme: Frequency and Rating Distribution
While building this site I curated 212,300 puzzles from the Lichess open puzzle database (CC0), filtering out low-popularity and rarely-played ones. When I aggregated that set by tactical theme, the counts suggested a few things about what a beginner might learn first. Here are the actual numbers.
Forks and sacrifices lead the non-mate tactics
Share of themes across all 212,300 puzzles, leaving aside the mate tags and the broad tags discussed below:
- Fork: 17,829 (8.4%)
- Sacrifice: 17,332 (8.2%)
- Pin: 13,506 (6.4%)
- Discovered attack: 12,694 (6.0%)
- Deflection: 11,486 (5.4%)
- Attraction: 10,059 (4.7%)
- Hanging piece: 7,801 (3.7%)
A single puzzle can carry several themes, so the shares add up to more than 100%. Among named tactics the order is still clear: forks and sacrifices lead, then pins and discovered attacks. The broad tags left out above describe the setting or the kind of move rather than a tactic, and several of them are larger: kingside attack (14,749) and advanced pawn (14,144) both outnumber pins, and defensive move (13,336) outnumbers discovered attacks. Quiet move (9,603) and exposed king (7,943) also outnumber hanging pieces. These counts cover the whole curated set; the result does not describe a beginner-only set or the frequency of tactics in actual games.
The mate tags sit above all of them and are counted separately here, since they describe what the puzzle asks for rather than the motif that gets you there: mate covers 57,285 puzzles, mate in 2 25,089, and mate in 1 23,166. Those get their own guides.
Which tactics are easy, and which are hard
Where each theme's puzzles sit on the Lichess puzzle rating scale shows which themes lean easy and which lean hard. Here "beginner" means a puzzle rating under 1200, and "hard" means 1900 or above.
Clustered in the beginner range (recognition mates):
- Smothered mate: 83% under 1200, 3 puzzles (0.5%) at 1900+
- Back-rank mate: 79% under 1200, 2% at 1900+
Smothered mate and back-rank mate have relatively recognizable piece arrangements. That may help explain their lower puzzle ratings, but the aggregate does not establish the cause. They are reasonable candidates for introducing mating shapes. Among non-mate tactics, skewer leans the same way, less strongly: 36% under 1200 and 22% at 1900+.
Clustered in the hard range (named tactics with the largest shares at 1900+, leaving aside the mate tags and the broad tags above; the first two are very small samples):
- Castling: 97% at 1900+ (61 puzzles)
- Underpromotion: 75% at 1900+ (32 puzzles)
- En passant: 0% under 1200, 67% at 1900+
- Zugzwang: 63%
- Clearance: 51%
- Discovered check: 42%
- Double check: 41%
- Attraction: 40%
- Sacrifice: 39%
- Trapped piece: 38%
En passant puzzles skew unusually hard. The counts do not say why. One guess: the capture is legal only right after a two-square pawn advance, and the point of the puzzle usually lies in the line it opens, not in the capture itself. Trapped-piece tags also lean toward the higher-rated band in this sample. Interference (cutting a defender's line) has no 1900+ figure in the table; a separate count puts 37% of its 1,010 puzzles at 2000 or above. Neither trapped piece nor interference is necessarily quiet or positional in every puzzle.
The fork sits in the middle: 24% under 1200 and 25% at 1900+, so substantial shares appear at both ends of the range. It is the most common non-mate tactic, and you keep meeting it from low-rated puzzles to high-rated ones.
So what should you learn first
Using this curated set alone, one possible starting order looks like this. The data does not compare how well each order teaches:
- Look at forks and pins first. They are frequent in this sample and include puzzles below 1200, making them reasonable introductory candidates.
- Learn the shapes of back-rank and smothered mate. Their high share of lower-rated puzzles makes them useful for basic pattern practice.
- Special rules like en passant can wait. The puzzles are rare and skew hard, so it is enough to know the rule precisely and keep it lower on your tactics-training priority.
The full table
The sections above pick out what stood out. Here is the whole set, for checking a figure or citing one. Ordered by puzzle count.
| Tactic | Puzzles | Share of DB | Under 1200 | 1900+ |
|---|---|---|---|---|
| Mate in 2 | 25,089 | 11.8% | 54% | 6% |
| Mate in 1 | 23,166 | 10.9% | 82% | 0% |
| Fork | 17,829 | 8.4% | 23.8% | 25.1% |
| Sacrifice | 17,332 | 8.2% | 13.9% | 39.3% |
| Kingside Attack | 14,749 | 6.9% | 32.6% | 22.5% |
| Advanced Pawn | 14,144 | 6.7% | not measured | not measured |
| Pin | 13,506 | 6.4% | 12.5% | 33.6% |
| Defensive Move | 13,336 | 6.3% | not measured | not measured |
| Discovered Attack | 12,694 | 6% | 22.8% | 30.2% |
| Deflection | 11,486 | 5.4% | 19.8% | 33.4% |
| Attraction | 10,059 | 4.7% | 6.5% | 40.4% |
| Quiet Move | 9,603 | 4.5% | not measured | not measured |
| Exposed King | 7,943 | 3.7% | 8.5% | 49.7% |
| Hanging Piece | 7,801 | 3.7% | 28.3% | 30.4% |
| Mate in 3 | 7,359 | 3.5% | 32% | 20% |
| Promotion | 5,568 | 2.6% | 23% | 35.7% |
| Skewer | 5,183 | 2.4% | 35.7% | 22.3% |
| Discovered Check | 4,933 | 2.3% | 15.1% | 42.1% |
| Queenside Attack | 3,515 | 1.7% | not measured | not measured |
| Back Rank Mate | 3,458 | 1.6% | 79.4% | 1.9% |
| Clearance | 3,373 | 1.6% | 9.4% | 51.3% |
| Trapped Piece | 3,187 | 1.5% | 3.9% | 37.6% |
| Intermezzo | 3,090 | 1.5% | not measured | not measured |
| Zugzwang | 2,670 | 1.3% | 7.5% | 62.5% |
| Opera Mate | 2,025 | 1% | 55.7% | 10.8% |
| Pillsbury's Mate | 1,979 | 0.9% | 53.8% | 11.3% |
| Remove the Defender | 1,809 | 0.9% | 7.9% | 32.6% |
| Double Check | 1,389 | 0.7% | 14.9% | 41% |
| Attacking f2/f7 | 1,388 | 0.7% | 48.4% | 17.4% |
| Mate in 4 | 1,378 | 0.6% | 4% | 49% |
| Interference | 1,010 | 0.5% | not measured | not measured |
| X-Ray Attack | 800 | 0.4% | 25.4% | 23.8% |
| Epaulette Mate | 772 | 0.4% | 58.4% | 8.4% |
| Smothered Mate | 617 | 0.3% | 83.3% | 0.5% |
| Corner Mate | 443 | 0.2% | 52.8% | 12.9% |
| Hook Mate | 379 | 0.2% | 53% | 8.2% |
| En Passant | 346 | 0.2% | 0% | 67.3% |
| Triangle Mate | 340 | 0.2% | 50.3% | 9.1% |
| Swallow's Tail Mate | 331 | 0.2% | 54.1% | 14.8% |
| Morphy's Mate | 314 | 0.1% | 49% | 13.1% |
| Collinear Move | 305 | 0.1% | not measured | not measured |
| Mate in 5+ | 294 | 0.1% | not measured | not measured |
| Blind Swine Mate | 235 | 0.1% | 56.2% | 11.5% |
| Arabian Mate | 217 | 0.1% | 60.8% | 4.1% |
| Kill Box Mate | 209 | 0.1% | 31.1% | 17.2% |
| Anastasia's Mate | 198 | 0.1% | 48% | 1% |
| Dovetail Mate | 147 | 0.1% | 43.5% | 6.8% |
| Vukovic Mate | 101 | under 0.1% | 14.9% | 27.7% |
| Double Bishop Mate | 64 | under 0.1% | 93.8% | 1.6% |
| Balestra Mate | 62 | under 0.1% | 25.8% | 14.5% |
| Castling | 61 | under 0.1% | 0% | 96.7% |
| Boden's Mate | 56 | under 0.1% | 85.7% | 1.8% |
| Underpromotion | 32 | under 0.1% | 12.5% | 75% |
Method and limits
The numbers aggregate the whole curated set for the 53 themes in the table above (shares and per-band counts are exact counts, not estimates). But "difficulty" is based on Lichess puzzle rating, so it reflects how often solvers miss a puzzle rather than the intrinsic difficulty of the tactic. Different curation rules (popularity and rating-deviation filters) would shift the numbers. Theme frequency here is not real-game frequency, and a puzzle-rating band is not the same thing as a player-rating band. The aggregate describes distributions; it does not prove why one motif is easier or harder.
This site draws its training puzzles from that set and schedules reviews by tactical concept. The how is covered in Applying SM-2 to Chess.
References
- Lichess Puzzle Database, the source data for these aggregates (CC0). The curation, by popularity and rating deviation, was done by this site.
- Lichess Puzzle Themes, the standard taxonomy for tactical themes.
Ready to start training?
Practice what you just read in our spaced repetition puzzle trainer.
Start Free