README.md

---
license: cc0-1.0
pretty_name: Lichess Standard Rated Games
dataset_info:
  features:
  - name: Event
    dtype: string
  - name: Site
    dtype: string
  - name: White
    dtype: string
  - name: Black
    dtype: string
  - name: Result
    dtype: string
  - name: WhiteTitle
    dtype: string
  - name: BlackTitle
    dtype: string
  - name: WhiteElo
    dtype: int16
  - name: BlackElo
    dtype: int16
  - name: WhiteRatingDiff
    dtype: int16
  - name: BlackRatingDiff
    dtype: int16
  - name: UTCDate
    dtype: date32
  - name: UTCTime
    dtype: time32[s]
  - name: ECO
    dtype: string
  - name: Opening
    dtype: string
  - name: Termination
    dtype: string
  - name: TimeControl
    dtype: string
  - name: movetext
    dtype: string
configs:
- config_name: default
  data_files:
  - split: train
    path: data/**/train-*
tags:
- chess
- games
- game
- lichess
size_categories:
- 1B<n<10B
---

> [!CAUTION]
> This dataset is still a work in progress and some breaking changes might occur. In the meantime, please use https://database.lichess.org/#standard_games
>

# Dataset Card for the Lichess Rated Standard Chess Games Dataset

## Dataset Description

**6,111,210,071** standard rated games, played on [lichess.org], updated monthly from the [database dumps](https://database.lichess.org/#standard_games).

This version of the data is meant for data analysis. If you need PGN files you can find those [here](https://database.lichess.org/#standard_games). That said, once you have a subset of interest, it is trivial to convert it back to PGN as shown in the Dataset Usage section.

This dataset is hive-partitioned into multiple parquet files on two keys: `year` and `month`:

```bash
.
├── data
│   └── year=2015
│       ├── month=01
│       │   ├── train-00000-of-00003.parquet
│       │   ├── train-00001-of-00003.parquet
│       │   └── train-00002-of-00003.parquet
│       ├── month=02
│       │   ├── train-00000-of-00003.parquet
│       │   ├── train-00001-of-00003.parquet
│       │   └── train-00002-of-00003.parquet
│       ├── month=03
│       │   ├── train-00000-of-00003.parquet
│       │   ├── train-00001-of-00003.parquet
│       │   └── train-00002-of-00003.parquet
│       ├── ...

```


### Dataset Creation

```python

```
### Dataset Usage

Using the `datasets` library:

```python
from datasets import load_dataset
dset = load_dataset("Lichess/chess-evaluations", split="train")
```
Using the `polars` library:

Using DuckDB:

Using `python-chess`:


## Dataset Details

### Dataset Sample

One row of the dataset looks like this:

```python
{
  "Event":,
  "Site":,
}
```

### Dataset Fields

Every row of the dataset contains the following fields:

- **`Event`**: `string`, 
- **`Site`**: `string`,

### Notes

- About 6% of the games include Stockfish analysis evaluations: [%eval 2.35] (235 centipawn advantage), [%eval #-4] (getting mated in 4), always from White's point of view.

- The WhiteElo and BlackElo tags contain Glicko2 ratings.

- Games contain clock information as PGN %clk comments since April 2017.