Update README.md
Browse files
README.md
CHANGED
|
@@ -46,9 +46,9 @@ tags:
|
|
| 46 |
<!-- Provide a quick summary of the dataset. -->
|
| 47 |
## Dataset Description
|
| 48 |
|
| 49 |
-
**5,
|
| 50 |
|
| 51 |
-
This dataset is updated monthly, and was last updated on
|
| 52 |
|
| 53 |
### Dataset Creation
|
| 54 |
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](https://github.com/ornicar/lichess-puzzler/tree/master/tagger). 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.
|
|
@@ -70,23 +70,23 @@ dset = load_dataset("Lichess/chess-puzzles", split="train")
|
|
| 70 |
One row of the dataset looks like this:
|
| 71 |
|
| 72 |
```python
|
| 73 |
-
{'PuzzleId': '
|
| 74 |
-
'
|
| 75 |
-
'
|
| 76 |
-
'
|
| 77 |
-
'
|
| 78 |
-
'
|
| 79 |
-
'
|
| 80 |
-
'
|
| 81 |
-
'
|
| 82 |
-
'
|
| 83 |
-
'GameUrl': 'https://lichess.org/seIMDWkD#25'}
|
| 84 |
```
|
| 85 |
### Dataset Fields
|
| 86 |
|
| 87 |
Every row of the dataset contains the following fields:
|
| 88 |
|
| 89 |
- **`PuzzleId`**: `string`, the puzzle's unique identifier. The puzzle would be live at `https://lichess.org/training/{PuzzleID}`.
|
|
|
|
| 90 |
- **`FEN`**: `string`, the FEN string of the position before the opponent makes their move.
|
| 91 |
- **`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.
|
| 92 |
- **`Rating`**: `int`, the Glicko-2 rating of the puzzle.
|
|
@@ -95,8 +95,6 @@ Every row of the dataset contains the following fields:
|
|
| 95 |
- **`NbPlays`**: `int`, the number of times a puzzle was played.
|
| 96 |
- **`Themes`**: `list`, a list of puzzle themes.
|
| 97 |
- **`OpeningTags`**: `list`, a list of openings. This is only set for puzzles starting before move 20.
|
| 98 |
-
- **`PuzzleUrl`**: `string`, the link to where the puzzle can be played.
|
| 99 |
-
- **`GameUrl`**: `string`, the link to the specific game and move the puzzle was extracted from.
|
| 100 |
|
| 101 |
## Additional Information
|
| 102 |
- For a list of all possible puzzle themes and their description: [puzzleTheme.xml](https://github.com/ornicar/lila/blob/master/translation/source/puzzleTheme.xml)
|
|
|
|
| 46 |
<!-- Provide a quick summary of the dataset. -->
|
| 47 |
## Dataset Description
|
| 48 |
|
| 49 |
+
**5,524,871 chess puzzles**, rated and tagged. See them in action on [Lichess](https://lichess.org/training/themes).
|
| 50 |
|
| 51 |
+
This dataset is updated monthly, and was last updated on November 12th, 2025.
|
| 52 |
|
| 53 |
### Dataset Creation
|
| 54 |
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](https://github.com/ornicar/lichess-puzzler/tree/master/tagger). 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.
|
|
|
|
| 70 |
One row of the dataset looks like this:
|
| 71 |
|
| 72 |
```python
|
| 73 |
+
{'PuzzleId': '000hf',
|
| 74 |
+
'GameId': '71ygsFeE/black#38',
|
| 75 |
+
'FEN': 'r1bqk2r/pp1nbNp1/2p1p2p/8/2BP4/1PN3P1/P3QP1P/3R1RK1 b kq - 0 19',
|
| 76 |
+
'Moves': 'e8f7 e2e6 f7f8 e6f7',
|
| 77 |
+
'Rating': 1575,
|
| 78 |
+
'RatingDeviation': 75,
|
| 79 |
+
'Popularity': 92,
|
| 80 |
+
'NbPlays': 674,
|
| 81 |
+
'Themes': ['mate', 'mateIn2', 'middlegame', 'short'],
|
| 82 |
+
'OpeningTags': ['Horwitz_Defense', 'Horwitz_Defense_Other_variations']}
|
|
|
|
| 83 |
```
|
| 84 |
### Dataset Fields
|
| 85 |
|
| 86 |
Every row of the dataset contains the following fields:
|
| 87 |
|
| 88 |
- **`PuzzleId`**: `string`, the puzzle's unique identifier. The puzzle would be live at `https://lichess.org/training/{PuzzleID}`.
|
| 89 |
+
- **`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}`
|
| 90 |
- **`FEN`**: `string`, the FEN string of the position before the opponent makes their move.
|
| 91 |
- **`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.
|
| 92 |
- **`Rating`**: `int`, the Glicko-2 rating of the puzzle.
|
|
|
|
| 95 |
- **`NbPlays`**: `int`, the number of times a puzzle was played.
|
| 96 |
- **`Themes`**: `list`, a list of puzzle themes.
|
| 97 |
- **`OpeningTags`**: `list`, a list of openings. This is only set for puzzles starting before move 20.
|
|
|
|
|
|
|
| 98 |
|
| 99 |
## Additional Information
|
| 100 |
- For a list of all possible puzzle themes and their description: [puzzleTheme.xml](https://github.com/ornicar/lila/blob/master/translation/source/puzzleTheme.xml)
|