Board Game Rating Prediction Dataset

Columns

• game_id: A unique identifier for each game recorded.
• nickname: The player's name used on the online Scrabble platform.
• total_turns: The total count of turns taken by the player during the game, typically ranging from 9 to 50 turns.
• first_five_turns_points: The accumulated points scored by the player within the initial five turns of the game.
• max_points_turn: The highest point total scored by the player in a single turn, observed values go up to 311 points.
• min_points_turn: The lowest point total scored in a single turn, which can be negative (as low as -221).
• max_min_difference: The difference in points between the highest and lowest scoring turns.
• first: Indicates whether the player took the first turn in the game.
• time_control_name: The classification of the game speed, primarily ‘regular’ (83%) or ‘rapid’ (9%).
• game_end_reason: Specifies the rule or condition that caused the game to conclude (e.g., STANDARD, RESIGNED).
• winner: A numerical representation of the game result: 1 for a win, -1 for a loss, and 0 for a draw.
• created_at: The timestamp indicating when the game was played.
• lexicon: The dictionary used for word verification, primarily CSW21 (61%) or NWL20 (29%).
• initial_time_seconds: The amount of time each player starts with, ranging from 15 seconds up to 3600 seconds.
• increment_seconds: The seconds gained by the player per turn, generally 0 seconds.
• rating_mode: Indicates if the game results affected player ratings (RATED, 74%) or not (CASUAL, 26%).
• max_overtime_minutes: The maximum permitted overtime before the player loses due to a timeout.
• game_duration_seconds: The total length of the game in seconds.
• time_used: The ratio representing the game duration divided by the initial time.

Distribution

Usage

• Developing supervised machine learning models to predict a player's rating or final winning status based on in-game performance metrics.
• Feature engineering based on game statistics like turn efficiency (points per turn) and time management.
• Performing statistical analysis on professional Scrabble game dynamics, player style, and the impact of different time controls or lexicons.
• Creating visualisations to understand player behaviour across different stages of a game (e.g., early game scoring via first_five_turns_points).

Coverage

License

Who Can Use It

• Data Scientists: For training and evaluating classification or regression models focused on competitive gaming outcomes.
• Game Analysts: To study optimal strategies, player consistency, and the statistical impact of high-scoring or low-scoring turns.
• Academic Researchers: Conducting studies on skill acquisition, strategic decision-making, and performance measurement in complex board games.

Dataset Name Suggestions

• Scrabble Player Performance Metrics
• Aggregated Scrabble Game Features
• Board Game Rating Prediction Dataset

Attributes