Fractional tackles: Leveraging player tracking data for within-play tackling evaluation in american football

football
Big Data Bowl
player-tracking data
Introducing fractional tackles, a new way to measure tackles with player-tracking data (2024 Big Data Bowl finalist)
Authors
Affiliations

Quang Nguyen

Department of Statistics & Data Science, Carnegie Mellon University

Ruitong (Larry) Jiang

Neuroscience Institute and Center for the Neural Basis of Cognition, Carnegie Mellon University

Meg Ellingwood

Department of Statistics & Data Science, Carnegie Mellon University

Ronald Yurko

Department of Statistics & Data Science, Carnegie Mellon University

Published

March 21, 2024

arxiv code

@article{nguyen2024fractional,
  title={Fractional Tackles: Leveraging Player Tracking Data for Within-Play Tackling Evaluation in American Football},
  author={Nguyen, Quang and Jiang, Ruitong and Ellingwood, Meg and Yurko, Ronald},
  journal={arXiv preprint arXiv:2403.14769},
  year={2024}
}

Abstract

Tackling is a fundamental defensive move in American football, with the main purpose of stopping the forward motion of the ball-carrier. However, current tackling metrics are manually recorded outcomes that are inherently flawed due to their discrete and subjective nature. Using player tracking data, we present a novel framework for assessing tackling contribution in a continuous and objective manner. Our approach first identifies when a defender is in a ``contact window’’ of the ball-carrier during a play, before assigning value to each window and the players involved. This enables us to devise a new metric called fractional tackles, which credits defenders for halting the ball-carrier’s forward motion toward the end zone. We demonstrate that fractional tackles overcome the shortcomings of traditional metrics such as tackles and assists, by providing greater variation and measurable information for players lacking recorded statistics like defensive linemen. We view our contribution as a significant step forward in measuring defensive performance in American football and a clear demonstration of the capabilities of player tracking data.