the indices of match-running metrics in the U14s and U15/16s cohorts, model 3 provided the best fit. That is, model 4, which allowed the slopes to vary randomly for position and spine/lateral, did not result in improvements in model fit. Discussion The purpose of the present study was to investigate