University of Technology Sydney mathematician, Steve Woodcock, describes the controversial alternative statistical inference culture within the field of sports science.

Elite sport is a multi-billion dollar industry and data analyses involving sports decisions are prone to abuse. This is one of the main messages from Dr Steve Woodcock's talk to the New South Wales branch of the society on 25th November 2020. Steve is a senior lecturer within the mathematics group at University of Technology Sydney.

The talk started by pointing out that out of Australia-based researchers who describe themselves as a statistician within the Google Scholar system, one that has among the highest number of citations is Victoria University employee Will Hopkins. His work is almost exclusively applied in sports science and he pioneered a concept known as magnitude-based decisions. According to the talk, Hopkins is best known for, in his words, "rejecting statistical inference and replacing it with a clinically and practically more relevant method of inference based on uncertainty in the magnitude of effects". It was then pointed out that several prominent statisticians - including big contributors to our society such as Adrian Barnett and Alan Welsh – have strongly criticised and debunked the magnitude-based decision approach.

The speaker gave us the following quote from former society president Barnett ``If I was ever to peer review a paper using magnitude-based inference then I would reject it and tell the authors to completely redo their analysis''.

Speaker Woodcock expressed at least some sympathy for magnitude-based decisions and said that in the context of small sample studies and potentially skewed risk-reward payoffs - which he defined loosely as "this won't do any harm and may be beneficial" - the approach may be a useful decision-making tool.

Moving away from Hopkins and the magnitude-based decisions controversy, Steve then described some dodgy sports science analyses and claims - such as one involving concussion evaluation in the Australian Football League. Another one concerned the claim of a sweet spot for reduced industry risk which relied on a dubious quadratic regression model.

The talk concluded with the question "An unmissable opportunity?" concerning how high quality statistics can permeate into Australian sports decision-making. He concluded by making the observation that the stumbling block is not from an unwillingness to improve statistical practice in the field, but rather a lack of dialogue between practitioners and statisticians about possible improvements to employed methodology.

Matt Wand

University of Technology Sydney