Abstract:
|
A valuable skill for NFL coaches is to be able to anticipate whether the opposing team will call a pass or run play. If the offensive play type can be predicted, the defensive team can call for example, a blitz or coverage play to gain an advantage. Based on various factors we develop a classification model, trained using play by play data for the 2012 and 2013 NFL seasons, to predict future play types. The factors used for initial consideration are: average yards per pass play, average yards per run play, previous 3 plays, distance to first down, distance to goal line, down, point differential,quarter, and time remaining. Factors used in the final model are determined using variable selection techniques. We compare the accuracy of multiple classification methods including logistic regression and decision trees to determine the optimal model. Once a final model is determined, an interactive visualization is created using the R Shiny package. This visualization is an intuitive and easily interpretable way to make real time decisions given the current game conditions.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.