EsportsBench: A Collection of Datasets for Benchmarking Rating Systems in Esports

EsportsBench is a collection of 20 esports competition datasets. Each row of each dataset represents a match played between either two players or two teams in a professional video game tournament. The goal of the datasets is to provide a resource for comparison and development of rating systems used to predict the results of esports matches based on past results. Date is complete up to 2024-03-31.

Recommended Usage

The recommended data split is to use the most recent year of data as the test set, and all data prior to that as train. There have been two releases so far:

• 1.0 includes data up to 2024-03-31. Train: beginning to 2023-03-31, Test: 2023-04-01 to 2024-03-31
• 2.0 includes data up to 2024-06-30. Train: beginning to 2023-06-30, Test: 2023-07-01 to 2024-06-30

import polars as pl
import datasets

esports = datasets.load_dataset('EsportsBench/EsportsBench', revision='1.0')
lol = esports['league_of_legends'].to_polars()
teams = pl.concat([lol['competitor_1'], lol['competitor_2']]).unique()
lol_train = lol.filter(pl.col('date') <= '2023-03-31')
lol_test = lol.filter((pl.col('date') >'2023-03-31') & (pl.col('date') <= '2024-03-31'))
print(f'train rows: {len(lol_train)}')
print(f'test rows: {len(lol_test)}')
print(f'num teams: {len(teams)}')
# train rows: 104737
# test rows: 17806
# num teams: 12829

The granulularity of the date column is at the day level and rows on the same date are not guaranteed to be ordered so when experimenting, it's best to make predictions for all matches on a given day before incorporating any of them into ratings or models.

# example prediction and update loop
rating_periods = lol.group_by('date', maintain_order=True)
for date, matches in rating_periods:
    print(f'Date: {date}')
    print(f'Matches: {len(matches)}')
    # probs = model.predict(matches)
    # model.update(matches)
# Date: 2011-03-14
# Matches: 3
# ...
# Date: 2024-03-31
# Matches: 47

Data Sources

• The StarCraft II data is from Aligulac
• The League of Legends data is from Leaguepedia under a CC BY-SA 3.0
• The data for all other games is from Liquipedia under a CC BY-SA 3.0