Dynamic Network DEA approach to basketball games efficiency

Abstract

Although Data Envelopment Analysis (DEA) has been widely applied to sports, not many studies are related to basketball. Two types of approaches have been developed to measure the efficiency in basketball so far: those focused on the assessment of players and those that assess the performance of the teams. Assuming that the number of points scored in a basketball game greatly influences the appeal of a game, in this paper, a new approach focused on the measurement of the scoring efficiency of the two teams that play a game is addressed. To do that, the performance of each team in each quarter and the carry-overs between successive quarters must be taken into account. This leads to a Dynamic Network DEA model with two subprocess (corresponding to the home and visitor teams) running in each quarter. A scoring efficiency can be computed for each team in each quarter as well as for each team overall, for each quarter overall and for the whole game. The proposed approach is applied to the matches played during the 2014–2015 NBA season.

Keywords:

• Basketball game
• scoring efficiency
• Dynamic Network DEA
• quarter and game efficiency