five

Kusya1/Lichess_Games_and_Metrics

收藏
Hugging Face2025-04-17 更新2025-10-25 收录
下载链接:
https://hf-mirror.com/datasets/Kusya1/Lichess_Games_and_Metrics
下载链接
链接失效反馈
官方服务:
资源简介:
这个数据集包含了超过500万个国际象棋游戏,这些游戏来源于Lichess.org。数据集包含了Stockfish引擎的评价、时钟时间、移动和游戏元数据。数据集由两个文件组成:`lichess_games_moves`,包含了传统的PGN风格的元数据和移动序列,存储为Parquet格式以便快速查询;`lichess_moves_sample.parquet`,包含了每个玩家每场比赛的性能指标,这些指标是从评价和计时数据中得出的,包括每一步的兵价值损失、移动准确性、时间使用等。每个游戏都通过一个唯一的游戏ID(UUID)来标识,这使得两个数据集之间的无缝连接成为可能。这种结构允许研究人员在游戏级别和玩家级别上分析玩家行为,支持作弊检测、玩家分析、技能进步分析、以及AI性能基准测试等任务。数据集涵盖了广泛的棋局类型、评级等级和事件格式,提供了对数百万个国际象棋游戏中决策制定、准确性和时间管理的详细视图。

This dataset contains over 5 million chess games sourced from Lichess.org, including Stockfish Engine Evaluations, clock timings, moves, and game metadata. The dataset is composed of two files: `lichess_games_moves`, which contains traditional PGN-style metadata and move sequences in Parquet format for fast querying; and `lichess_moves_sample.parquet`, which contains per-player, per-game performance metrics derived from evaluations and timing data. Each game is uniquely identified by a game_id (UUID), enabling seamless joins between the two datasets. This structure allows for analysis of player behavior at both the game and player levels, supporting tasks such as cheat detection, player profiling, skill progression analysis, and AI performance benchmarking. The dataset covers a wide range of game types, rating tiers, and event formats, offering a granular view into decision-making, precision, and time management across millions of chess games.
提供机构:
Kusya1
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作