NFL Play Classification Using GNN and Player Tracking Data

• Carlos E. G. R. Souza UFOP
• Eduardo J. S. Luz UFOP
• Vander L. S. Freitas UFOP

DOI: https://doi.org/10.5753/eniac.2025.14217

Abstract

This study addresses the classification of offensive plays in the National Football League (NFL) as either pass or rush, utilizing official NFL datasets from 2022-2023 season. The central hypothesis posits that the relative spatial positioning of players on the field, along with their interactions, directly influences the offensive strategy. To investigate this, we generate graph-based representations connecting players, which serve as input for a Graph Convolutional Network (GCN) for the subsequent classification of these graph structures. The empirical results demonstrate that the GCN-based framework outperforms conventional machine learning architectures, namely Random Forest and Multilayer Perceptron.

References

Biro, P. and Walker, S. G. (2022). A reinforcement learning based approach to play calling in football. Journal of Quantitative Analysis in Sports, 18(2):97–112.

Bunker, R. and Susnjak, T. (2022). The application of machine learning techniques for predicting match results in team sport: A review. Journal of Artificial Intelligence Research, 73:1285–1322.

Cormen, T. H., Leiserson, C. E., Rivest, R. L., and Stein, C. (2022). Introduction to algorithms. MIT press.

Eager, E., Brown, B., Chahrouri, G., Riske, T., Spielberger, B., Yui, L. S., and Seth, T. (2022). Using tracking and charting data to better evaluate nfl players: A review. In 2022 MIT Sloan Sports Analytics Conference.

Fernandes, C. J., Yakubov, R., Li, Y., Prasad, A. K., and Chan, T. C. (2020). Predicting plays in the national football league. Journal of Sports Analytics, 6(1):35–43.

Goyal, U. (2020). Leveraging machine learning to predict playcalling tendencies in the NFL. Doctoral dissertation, Massachusetts Institute of Technology.

Heiny, E. L. and Blevins, D. (2011). Predicting the atlanta falcons play-calling using discriminant analysis. Journal of Quantitative Analysis in Sports, 7(3).

Kipf, T. N. and Welling, M. (2016). Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907.

Reyers, M. and Swartz, T. B. (2023). Quarterback evaluation in the national football league using tracking data. AStA Advances in Statistical Analysis, 107(1):327–342.

Salim, M. O. and Brandao, W. C. (2018). Predicting the success of nfl teams using complex network analysis. In International Conference on Enterprise Information Systems (ICEIS), volume 1, pages 135–142.

Teich, B., Lutz, R., and Kassarnig, V. (2016). Nfl play prediction. arXiv preprint arXiv:1601.00574.

Xenopoulos, P. and Silva, C. (2021). Graph neural networks to predict sports outcomes. In 2021 IEEE International Conference on Big Data (Big Data), pages 1757–1763. IEEE.

Xia, V., Jain, K., Krishna, A., and Brinton, C. G. (2018). A network-driven methodology for sports ranking and prediction. In 2018 52nd Annual Conference on Information Sciences and Systems (CISS), pages 1–6. IEEE.

Ötting, M. (2021). Predicting play calls in the national football league using hidden markov models. IMA Journal of Management Mathematics, 32(4):535–545.