Anthropometric and physical performance profiling does not predict professional contracts awarded in an elite Scottish soccer academy over a 10-year period.

AbstractThe purpose of this long-term retrospective analysis was to determine whether anthropometric and physical performance data could predict success in elite youth Scottish soccer players. Stature, body mass, sprint, jump and aerobic performance were collected from 512 players (U10 to U17) across a 10-year period. Players participated in an average of four profiling sessions (range: 1-14) and up to a maximum of three per year (August, December, and May) with standardisation applied to the surface, test order, time and protocols. One hundred players were awarded professional contracts. Associations between variables were quantified with mixed-effects linear models. Prediction was assessed with least absolute shrinkage and selection operator (LASSO) regression developed on a training set (2/3 data) and tested with proportion of error on a leave-out (1/3 data) test set. Confidence intervals were obtained through bootstrap LASSO samples. A strong relative age bias was identified with 50% of successful players born in the first quarter of the year. Successful players were on average taller and performed better in sprint and jump tests (p < 0.05). However, effects were small and even when variables were combined, proportion of errors identified were similar to random guessing (0.45[95%CI:0.41-0.49]). The results indicate that whilst successful players as youths demonstrate on average distinct anthropometric and physical profiles, these differences are unlikely to provide a reliable source to predict success within an already talented group. Practitioners should use data collected to guide exercise prescription but be aware of its substantive limitations in predicting success in isolation.Highlights Using robust statistical procedures, researchers and practitioners within soccer academies that are continually collecting data should assess whether accurate predictions can be made combining data across a holistic range of dimensions including physiological, technical, psychological, tactical skills and expertise from technical coaches.Academies should consider processes such as coach and scout education programmes to reduce the negative impacts of controllable factors such as the RAE.There are limitations of using anthropometric and physical performance profiling data to predict who will become a successful player. The information is still an important part of the talent development process with data being used to assist the creation and individual tailoring of physical training and appropriate recording and monitoring is encouraged.