lapp0/100M_random_chess_boards

该数据集包含100百万个随机采样的国际象棋棋盘位置。生成过程包括模拟100百万场随机游戏，每场游戏的每一步都是随机决定的，直到游戏结束（如将死、逼和或和棋）。然后，从每场游戏中随机选择一个棋盘状态。数据集中的每一行除了棋盘位置的FEN表示（fen），还包含了其他统计信息，如从游戏开始到采样位置的移动次数（n_moves）、游戏结束前的总移动次数（term_n_moves）、游戏结束的原因（term_reason）以及白方是否获胜（term_white_wins）。

This dataset contains 100 million randomly sampled chess board positions. The process to generate each board state involves simulating 100 million random games, where each move is determined randomly until the game ends (e.g., checkmate, stalemate, or draw). Then, a single random board state is selected from each game. Each row in the dataset, besides the FEN representation of the board position (fen), includes other statistics such as the number of moves made from the start of the game to reach the sampled position (n_moves), the total number of moves in the game before it terminated (term_n_moves), the reason the game ended (term_reason), and whether white won the game (term_white_wins).