Title: Survival analysis in basketball: An analysis of the NBA players' offensive performance
Authors: Ambra Macis - Universita' degli studi di Brescia (Italy) [presenting]
Marica Manisera - Universita' degli Studi di Brescia (Italy)
Paola Zuccolotto - University of Brescia (Italy)
Abstract: Over the years, statistics have been widely used in sport analytics. Many questions have been answered concerning, for example, player performance, outcome prediction of a match or a tournament and factors identifying successful and unsuccessful teams. Moreover, many studies have been developed to study athletes' dropout rate or their return to sport after an injury. These last studies have been developed through survival analysis. However, up to now, to the best of our knowledge, survival methods have not been used for analyzing player performance. The offensive performance of NBA players is investigated in terms of the amount of points gained during a season segment using statistical methods for survival analysis. More in detail, the available players' statistics in the pre-All-Stars game segment have been used as baseline covariates. Then, the post-All-Stars game season segment has been observed to analyze whether the players exceeded or not a given amount of points and how many minutes they played for experiencing or not the event of interest. The final aim is to identify the features associated with the probability of exceeding the given threshold during a fixed time period.