Statistics and Computational Methods Seminar Series - Fall 2025

Speaker: Jim Griffin (University College London)

Title: "Using statistical modelling of elite athlete performance for clean sport"

Abstract: 

Many anti-doping organisations are tasked with protecting clean sport through drug testing. Doping in elite sport is designed to improve performance. This has led to interest in using competition performances, which are routinely collected, to estimate an athlete's risk of doping for targeting testing. In this talk, I will describe longitudinal models of athletic performance, their use in a risk scoring system and the ability of the risk score to discriminate between doped and clean athletes. The project has raised some interesting statistical problems including the modelling of seasonality, estimation of the population distribution of latent variables, and ROC curves with errors in the labels.

Related papers:

J. Hopker, J. E. Griffin, J. Brookhouse, J. Peters, S. Iljukov, and Y. O. Schumacher (2020): “Performance profiling as an intelligence-led approach to anti-doping in sports,” Drug Testing and Analysis, 12, 402-409.

J. E. Griffin, L. C. Hinoveanu and J. Hopker (2022): “Bayesian modelling of elite sporting performance with large databases,” Journal of Quantitative Analysis in Sports, 18, 253-268.

J. G. Hopker, J. E. Griffin, L. C. Hinoveanu, J. Saugy and R. Faiss (2024): “Competitive performance as a discriminator of doping status in elite athletes,” Drug Testing and Analysis, 16, 473-481.

M.-Z. Spyropoulou, J. G. Hopker and J. E. Griffin (2024): “Modelling between- and within-season trajectories in elite athletic performance data” 

https://arxiv.org/abs/2405.17214