Understanding Pass Interference Through Catch Probability Distributions (Theme #2 – Propose a Rule Change)

This project relates to using the tracking data given by the NFL Big Data Bowl to build models related to pass interference and catch probability. In particular, we analyze how changing the passing distances in plays affects the distributions of catch probability. For instance, instead of a spot foul penalty, what would the catch probability distribution look like if we were to impose a 10-yard penalty instead?

The figure below shows the pipeline followed in our project:

By answering these kinds of questions, we propose a novel "two-level system" driven by transparency and data-driven models: a penalty for "flagrant fouls" that result in a spot foul and "common fouls" that result in a 10-yard penalty. This system, along with our models and data, is explained fully in our 12-page paper: pass-interference.pdf

Finally, our code, data, and figures are all included in this repository.