Uncertainty, Heuristics and Injury Prediction
I was asked by Rod Whiteley and Nicol van Dyk to contribute to the Aspetar Journal targeted topic issue that just got released off the press. I tried to combine my knowledge of predictive analytics, machine learning, philosophy of science, heuristics and practical experience as coach & sport scientist into one article. Hopefully I managed to create readable narrative.
I must admit that the paper was highly influenced by Richard McElreath, Nassim Taleb, Gerd Gigerenzer, Scott Page and Leo Breiman, and I should probably add Galit Shmueli. I hope that this paper reaches them somehow since I am more than interested in their critiques.
Some of you might be already familiar with my “critique” of current injury prediction (or should I say retrodiction) models and papers (click HERE). This article provides similar, but broader discussion, mainly focusing on three objectives of modelling: prediction, inference and intervention. I also cover the distinctions between risk and uncertainty and why fast and frugal rules might work better in uncertain and complex world (I also presented one simple model for decision making – heuristic matrix).
Anyway, hopefully you will enjoy the article and make sure to check other great papers in the journal.