A1182
Title: Borrowing strength in the era of tracking data: Statistical challenges and opportunities in sport
Authors: Stephanie Kovalchik - Teamworks (Germany) [presenting]
Abstract: Borrowing strength, the use of information from related units or populations to improve estimation for a target group, is a central idea in modern statistical modeling. This principle has found particularly influential applications in sports research, where methods such as Efron and Morris empirical Bayes shrinkage and PECOTA-style projection systems stabilize player-level estimates by pooling across similar athletes. The purpose is to examine how emerging trends in data collection in both amateur and professional sport, particularly the increasing ubiquity of comparable tracking technologies and fitness monitoring tools, are generating new opportunities to extend this framework. Applications include joint modeling of training and competition data, hierarchical pooling across levels of play, and transferring information across sports with shared performance features. The major statistical and computational challenges that must be addressed are also discussed, including the need for harmonization across measurement systems, contextualization, and scalable inference methods.
