Newswise — What determines a team’s home advantage, and why does it change with time? It’s a question that’s been the subject of many studies. But now, Austin Harris, a doctoral student at the University of Wisconsin-Milwaukee, is using data science to find the answer in National Basketball Association games.
Harris collected season performance statistics for all NBA teams across 32 seasons (1983-84 to 2017-18). Data were also obtained for other potential influences identified in the literature, including variables like stadium attendance and team market size.
Using a data science method called an artificial neural network (ANN), a team’s home advantage was diagnosed using team performance statistics only. When data from possible influences were applied, the ANN identified only one associated with larger advantages at home: Teams that made more two-point and free-throw shots.
Given the rise in three-point shooting in recent years, this finding partially explains the gradual decline in home advantage observed across the league over time.
Neural networks operate in a way that is similar to the human brain with certain data turned “on” or “off,” like a neuron firing. The key difference between ANN and similar regression techniques is that ANN can pick up on the nuanced, non-linear connections in the dataset. ANNs are most beneficial when relationships in the data are complex.
The paper will be published July 31 in the journal PLOS ONE at https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0220630