Owen P. Hall, Jr., PhD, PE, PhD, Professor of Decision Sciences
A Hierarchical Linear Modeling Approach for Assessing NBA Player and Team Performance, International Journal of Computer Science in Sport, 14(2), 2015.
Abstract
With teams’ annual payrolls nearing $100 million and valuations for some teams exceeding $2 billion, the National Basketball Association is big business. That being the case, many pro-basketball general managers are turning to analytics to discover ways to improve organizational performance. The purpose of this paper was to highlight the results of an analytics-based assessment of both player and team performance, using data from the regular 2012-2013 NBA season. The analytical paradigm described in this paper consisted of a two-tiered hierarchical linear modeling design that combined several specific on-court and off-court factors. The analysis also introduced a relatively new sports performance metric—entropy, which can be used to measure the degree of disorder at both the player and the team level. The target variable was Hollinger’s Player Efficiency Rating. The results of the analysis revealed that players’ ages, entropy, and compensation were among a number of factors that were found to be statistically significant. To this end, NBA general managers can use this modeling approach to evaluate both trade and draft opportunities.