Generating the initial dataset chess puzzles took more than 50 years of CPU time. We went through 300,000,000 analyzed games from the Lichess database, and re-analyzed interesting positions with Stockfish 12/13/14/15 NNUE at 40 meganodes. The resulting puzzles were then automatically tagged. To determine the rating, each attempt to solve is considered as a Glicko-2 rated game between the player and the puzzle. Finally, player votes refine the tags and define popularity.

Dataset Usage

Using the datasets library:

from datasets import load_dataset
dset = load_dataset("Lichess/chess-puzzles", split="train")

Dataset Details

Dataset Sample

One row of the dataset looks like this:

{'PuzzleId': '000hf',
 'GameId': '71ygsFeE/black#38',
 'FEN': 'r1bqk2r/pp1nbNp1/2p1p2p/8/2BP4/1PN3P1/P3QP1P/3R1RK1 b kq - 0 19',
 'Moves': 'e8f7 e2e6 f7f8 e6f7',
 'Rating': 1575,
 'RatingDeviation': 75,
 'Popularity': 92,
 'NbPlays': 674,
 'Themes': ['mate', 'mateIn2', 'middlegame', 'short'],
 'OpeningTags': ['Horwitz_Defense', 'Horwitz_Defense_Other_variations']}

Dataset Fields

Every row of the dataset contains the following fields:

PuzzleId: string, the puzzle's unique identifier. The puzzle would be live at https://lichess.org/training/{PuzzleID}.
GameId: string, the unique identifier of the specific game and move the puzzle was extracted from. The game would be accessible at https://lichess.org/{GameId}
FEN: string, the FEN string of the position before the opponent makes their move.
Moves: string, the solution to the puzzle. All player moves of the solution are "only moves", i.e. playing any other move would considerably worsen the player position. An exception is made for mates in one: there can be several. Any move that checkmates should win the puzzle.
Rating: int, the Glicko-2 rating of the puzzle.
RatingDeviation: int, the Glicko-2 rating deviation of the puzzle.
Popularity: int, a number between 100 (best) and -100 (worst), calculated as 100 * (upvotes - downvotes)/(upvotes + downvotes). Votes are weighted by various factors such as whether the puzzle was solved successfully or the solver's puzzle rating in comparison to the puzzle's.
NbPlays: int, the number of times a puzzle was played.
Themes: list, a list of puzzle themes.
OpeningTags: list, a list of openings. This is only set for puzzles starting before move 20.

Additional Information

For a list of all possible puzzle themes and their description: puzzleTheme.xml
To better understand puzzle themes, check out this study: https://lichess.org/study/viiWlKjv
GitHub Repo: https://github.com/lichess-org/database
Official Website: https://database.lichess.org/#puzzles
For a list of all possible openings: https://github.com/lichess-org/chess-openings
For the blog post introducing puzzles from 2020: blog/new-puzzles-are-here
For the blog post analyzing Lichess puzzle usage from 2021: blog/some-puzzling-analysis

Downloads last month: 12

Total file size:

476 MB