Improving longitudinal performance assessment of youth soccer players: 10 m sprint percentile curves adapted to biological age

Monitoring athletic development in youth soccer is crucial for player evaluation, identifying training needs, and determining long-term progression. However, standard percentile assessments based on chronological age (CA) do not account for biological maturity or developmental variability. This study aimed to improve 10 m sprint performance assessment in youth soccer by integrating biological age (BA) into percentile modeling and applying linear mixed models (LMM) to capture individual development. The analysis was based on 10 m sprint data collected within the Swiss Football Association's talent development program between 2017 and 2024, comprising 9476 observations for the Lambda Mu Sigma (LMS) method and 3983 for LMMs. BA was calculated using the Mirwald method as an estimation for peak height velocity. Empirical percentile curves (LMS) were generated for both CA and BA, while LMMs established longitudinal reference curves and enabled individual performance predictions using bootstrap resampling. In males, BA explained more variance in sprint performance than CA (R² = 0.22 vs. 0.18), whereas no significant predictors were identified for females. Percentile curves based on BA elevated rankings of late-maturing players and lowered those of early-maturing players, suggesting better consideration of developmental differences. LMMs provided a more comprehensive modeling framework than LMS, by incorporating repeated measures and individual developmental trajectories. Integrating BA and LMMs longitudinal modeling could enhance the fair evaluation of youth soccer players. Findings support individualized, maturity-adjusted monitoring, offering practical value for longer-term performance diagnostic and evaluation. This statistical approach, applied to a large practice-oriented dataset, enables targeted and sustainable improvement of youth player development.