1. Chat message data - kbo2011_naverbasevall_comment.sql: chat messages in Naver Sports that users posted in 2011 - kbo2012_naverbasevall_comment.sql: chat messages in Naver Sports that users posted in 2012 Each MySQL table consists of the following six columns. 1. comment_id: Integer variable to identify each comment 2. gid: Integer variable to identify each game 3. uid: Integer variable to identify each user 4. category_id: A team name that a user selected when a chat message is posted 5. content: Chat message text 6. post_time_relative: Elapsed seconds since the game starts 2. KBO game information - game_basic.sql: KBO game information in 2011/2012 (This data was collected from here, http://sports.news.naver.com/schedule/index.nhn?uCategory=&category=kbo&year=2010&month=03) This MySQL table consists of the following 12 columns. 1. gid: Integer variable to identify each game 2. game: Game ID used in Naver Sports 3. year: year of the game play 4. month: month of the game play 5. date: date of the game play 6. time: time of the game play 7. day: day of the game play 8. team_away: name of the away team 9 team_home: name of the home team 10. score_away: score by the away team 11. score_home: score by the home team 12. location: location of the game play 3. KBO game play-by-play data - game_playbyplay.sql: playbyplay data in 2011/2012 (This data was collected from Naver Sports; EX, http://sports.news.naver.com/gameCenter/gameRecord.nhn?category=kbo&gameId=20100327HHSK0) This MySQL table consists of the following 21 columns. 1. play_no: Integer variable to identify each play 2. gid: Integer variable to identify each game 3. inning: inning of the play 4. pitcher: current pitcher name 5. batter: current batter name 6. batter_no: current batter order 7. pitch: the pitch result 8. strike: current number of strikes 9. ball: current number of balls 10. out_cnt: current number of outs 11. score_home: current score of the home team 12. score_away: current score of the away team 13. 1st_base: player name on the 1st base 14. 2nd_base: player name on the 2nd base 15. 3rd_base: player name on the 3rd base 16. runner_state: overall state of the three bases (ex: 1 --> A player is on 1st base, but 2nd and 3rd bases are empty, 13 --> Two players are on 1st and 3rd bases respectively, but 2nd base is empty) 17. runner_cnt: the number of players on bases 18. result_r: r by this play 19. result_rbi: rbi by this play 20. result_score_home: total home score after this play 21. result_score_away: total away score after this play 4. Winning rate data - game_winrate.sql: winning rates of the home/away teams for an inning (This data is calculated based on the playbyplay data. For the details, please refer to our paper--currently under review) This MySQL table consists of the following 21 columns. 1. play_no: Integer variable to identify each play 2. gid: Integer variable to identify each game 3. inning: inning of the play 4. away_win: probability that the away team will win this game 5. draw: probability that the game will finish a tie 6. home_win: probability that the home team will win this game If you have any question about the dataset, please contact Minsam Ko (msko@kaist.ac.kr).