By Katie Gerbasich
Brook Russell, Ph.D., studies athletic performance beyond the average. As an associate professor in the Clemson University School of Mathematical and Statistical Sciences, he examines what happens at the far edges of a data set: human performance at the extremes.
Russell presented an invited talk in May at the American Statistical Association’s Virtual Sports Analytics Conference about two collaborative research projects examining speed and athleticism among NFL players.
The work grew from Russell’s collaboration with Paul Hogan, Clemson’s senior assistant football strength and conditioning coach. Russell conducted the quantitative analysis, while Hogan helped interpret the findings and place them in the context of football and human performance.
“It’s kind of a unique collaboration, to be honest,” Russell said. “There’s probably not a lot of football strength and conditioning coaches working on papers with people in the math department.”
The researchers use extreme value theory, a statistical framework traditionally applied in areas such as hydrology and finance to study rare events. Because elite NFL athletes are statistical outliers, this approach allows Russell and Hogan to examine performances at the far end of human ability and find characteristics that might be obscured when the data is viewed as a whole.
Their first project, published in 2018 in the Journal of Quantitative Analysis in Sports, examined relationships among NFL Combine event performances using both extreme-value and non-extremes-based approaches. The researchers explored how those different approaches could inform the evaluation of NFL prospects.
Their second project focused on differences between position groups, comparing cornerbacks and wide receivers across three NFL Combine events: the 40-yard dash, vertical jump and broad jump. The analysis suggested that wide receivers held a slight advantage in the 40-yard dash and vertical jump performances, while cornerbacks appeared to have a small edge in the broad jump.
In separate work, a graduate student examined how different Combine events relate to one another across typical and elite performers. The findings could help teams think differently about talent evaluation. If certain events provide overlapping information, teams may not need to rely on every available measurement when assessing a prospect.
Connecting statistical analysis to practical application is one of the central challenges of sports analytics.
“It’s easy to take in a lot of data and really interesting data sets and do interesting models,” Russell said. “But actually tying it back to some sort of decision-making process in practices or games can be a challenge.”
As sports analytics has grown, the field has drawn students and researchers from communication, marketing, economics and other disciplines. That interdisciplinary environment mirrors Russell’s work, which extends beyond athletic performance to the relationship between sports and business.
In a separate collaborative project with Angeline Scheinbaum, Ph.D., Dan Duncan Professor of Sports Marketing at Clemson, Russell supported a graduate student studying how fans’ perceptions of corporate social responsibility influence their willingness to use a company’s products or services. The researchers are preparing the project for publication.
Whether he is studying a 40-yard dash or the relationship between corporations and their communities, Russell uses data to better understand human performance and sports business.
