Great Lakes Analytics in Sports Conference
Quantifying Foul "Tilt" of NBA Players
In the NBA, certain players such as DeMarcus Cousins are well known for accruing a large number of fouls per game and seem to “tilt” in the sense that as they accrue fouls, they are more likely to continue to foul. We model foul rates using both basic analysis and a stratified Cox model and find that Cousins indeed tends to “tilt,” even after controlling for other potentially confounding variables. We further consider the effect substitutions can have on players that exhibit this “tilting” behavior, so that coaches can make more intelligent substitutions for certain players, especially as they approach the foul limit.
Katherine Evans is a fifth year graduate student in Biostatistics at Harvard University. Her academic research is in semi-parametric methods for missing data in causal inference. While she mostly applies these methods to comparative effectiveness research using electronic medical records, she has far more fun applying them to sports data.
Back to Great Lakes Analytics in Sports Conference