What is the difference between a winning and a losing team: insights from Euroleague basketball

ÇENE E.

INTERNATIONAL JOURNAL OF PERFORMANCE ANALYSIS IN SPORT, cilt.18, sa.1, ss.55-68, 2018 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 18 Sayı: 1
  • Basım Tarihi: 2018
  • Doi Numarası: 10.1080/24748668.2018.1446234
  • Dergi Adı: INTERNATIONAL JOURNAL OF PERFORMANCE ANALYSIS IN SPORT
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.55-68
  • Anahtar Kelimeler: Classification tree, Bayesian model averaging, BMA, machine learning, team sports, game outcome, GAME-RELATED STATISTICS, OLYMPIC GAMES, DISCRIMINATE, PERFORMANCE, CHAMPIONSHIP, INDICATORS, QUALITY, PLAYERS
  • Yıldız Teknik Üniversitesi Adresli: Evet

Özet

The purpose of this study is to determine which game-related factors have the most influence on game outcome for basketball matches based on the 2016-2017 Euroleague season. Games are divided into three groups with cluster analysis based on final score differences. First, independent samples t-test is used to detect differences between winning and losing teams among game-related variables. Later, Bayesian Model Averaging is employed to determine key candidate variables. Finally, conditional interference classification trees are constructed for all game groups. According to classification tree results, true shooting percentage, steals and committed fouls separates winners from losers for the close games. While 2-point field goal made, 3-point field goal made, steals and defensive rebounds are crucial for the balanced games; 2-point field goal made and defensive rebounds are the most influential game statistics on the game outcome for the unbalanced games. The results from the classification trees implied that in close games, quality of the shots are more important than the quantity of shots, whereas, the inverse deduction can be made for balanced and unbalanced games. These results may show guidance to basketball coaches and players in terms of training and game preparation.

The purpose of this study is to determine which game-related factors have the most influence on game outcome for basketball matches based on the 2016-2017 Euroleague season. Games are divided into three groups with cluster analysis based on final score differences. First, independent samples t-test is used to detect differences between winning and losing teams among game-related variables. Later, Bayesian Model Averaging is employed to determine key candidate variables. Finally, conditional interference classification trees are constructed for all game groups. According to classification tree results, true shooting percentage, steals and committed fouls separates winners from losers for the close games. While 2-point field goal made, 3-point field goal made, steals and defensive rebounds are crucial for the balanced games; 2-point field goal made and defensive rebounds are the most influential game statistics on the game outcome for the unbalanced games. The results from the classification trees implied that in close games, quality of the shots are more important than the quantity of shots, whereas, the inverse deduction can be made for balanced and unbalanced games. These results may show guidance to basketball coaches and players in terms of training and game preparation.