Abstract:
|
In sports, the location of events is often of interest. These events often have a spatial component. For example, the location of shots in basketball, hockey or soccer are important indicators of performance. Recently, there has been a good deal of attention in the sports statistics/analytics realm to visual representations of these spatial locations. In this analysis, we consider the question of whether two sets of spatial locations on the same surface are statistically different. We propose and evaluate randomization tests of the spatial CDF's using the Kolmogorov-Smirnov and the Cramer- von Mises test statistics. We apply this test to data from the National Women's Hockey League to consider differences within a season for a given team and differences between teams.
|