Izvorni znanstveni članak

https://doi.org/10.26582/k.51.1.9

Key game indicators in nba players’ performance profiles

Rubén Dehesa ; VALFIS Research Group (IBIOMED), FCAFD. University of León, León, Spain
Alejandro Vaquera orcid.org/0000-0003-1018-7676 ; VALFIS Research Group (IBIOMED), FCAFD. University of León, León, Spain
Bruno Gonçalves orcid.org/0000-0001-7874-4104 ; Research Center in Sports Sciences, Health Sciences and Human Development, CIDESD, CreativeLab Research Community. Vila Real, Portugal
Nuno Mateus orcid.org/0000-0001-7275-9161 ; Research Center in Sports Sciences, Health Sciences and Human Development, CIDESD, CreativeLab Research Community. Vila Real, Portugal
Miguel-Ángel Gomez-Ruano orcid.org/0000-0002-9585-3158 ; Universidad Politécnica de Madrid, Spain
Jaime Sampaio ; University of Trás-os-Montes e Alto Douro, Vila Real, Portugal

Sažetak

The aim of the present study was to identify and describe players’ performances in NBA games using individual and team-based game variables. The sample was composed by 535 balanced games (score differences below or equal to eight points) from the regular season (n=502) and the playoffs (n=33). A total of 472 players were analysed. The individual-based variables were: minutes on court, effective field-goal percentage, free-throws/field-goals ratio, offensive rebound percentage, turnover percentage and playing position. The team-based variables were: team points minus opponent’s points (on and off court), NET score (player’s on values minus his/her off values), maximum negative and positive point difference, team’s winning percentage, game pace, defensive and offensive ratings. A two-step cluster analysis was performed to identify the player’s profiles during regular season and playoff games. The results identified five performance profiles during regular season games and four performance profiles during playoff games. The profiles identified were mainly characterized by the game quarter and the negative NET indicator (players’ performance on court minus their performance off court) in regular season games and the positive NET indicator during playoff games and second and third game-quarters. Coaching staffs can fine-tune these profiles to develop more team-specific models and, conversely, use the results to monitor and rebuild team formation under the constrained dynamics of the game and competition stages.

Ključne riječi

collective behaviour; decision-making; game statistics; machine learning; cluster analysis; elite basketball

Hrčak ID:

218235

URI

https://hrcak.srce.hr/218235

Datum izdavanja:

30.6.2019.

Posjeta: 2.617 *