Belangrijkste concepten
Analyzing the impact of playing style compatibility on scoring efficiency in basketball lineups.
Samenvatting
This study focuses on analyzing the impact of playing style compatibility on scoring efficiency in basketball lineups. It employs shooting style clustering and offensive role clustering to provide insights into effective player combinations for increased scoring efficiency. The analysis includes dataset descriptions, preprocessing methods, clustering techniques, variable creation for lineup analysis, and lineup analysis results using regression models.
Abstract:
- Scoring efficiency is crucial in basketball due to repeated possessions.
- Effective collaboration among players with diverse styles enhances scoring efficiency.
- Analyzing player compatibility quantitatively can contribute to winning games.
Introduction:
- Scoring efficiency is vital in basketball due to the increasing number of possessions per game.
- Regression analysis is used to assess player compatibility based on performance metrics.
Methods:
- Shooting Style Clustering: Utilizes Wasserstein distance for shot tendencies between players.
- Offensive Role Clustering: Uses playtype data and advanced stats for soft clustering of offensive roles.
- Lineup Analysis: Predicts scoring efficiency using Bayesian models based on cluster combinations.
Results:
- Shooting Style Clusters: 13 clusters identified with unique player characteristics and shooting tendencies.
- Comparison of Shot Efficiency within Clusters: Top players from different clusters show varying TS% values.
- Lineup Analysis: Regression models predict OFFRTG with varying accuracy based on shooting style clusters.
Statistieken
In a basketball game, the average number of team possessions has increased over the years (less than 100 to more than 100).
The study uses tracking data from SportVU for shooting style clustering.
Playtype data obtained from NBA.com is used for offensive role clustering.