A Reduction in Match-to-match Variability Using Maximal Mean Analyses in Sub-elite Soccer.

Abstract

The match-to-match variability of external loads in National Premier League soccer competition was determined. Global positioning systems (GPS) data were collected from 20 sub-elite soccer players over 2-10 matches from a single season. Match data were collected from during one season. Twenty-six matches were recorded and 10 were utilised within final match-to-match analysis based on stringent data selection criteria. A symmetric moving average algorithm was applied to GPS data over specific time windows (1, 5, 10, 60, 300 and 600 s), and maximal speed and metabolic power values then calculated at each time interval during each match. Match-to-match coefficients of variation (CV) were greatest for sprint-speed running distance (36.3-43.6%) when comparing 2 vs. 10 matches. CVs for maximal mean speed (4.9-7.0%) and metabolic power (4.4-9.6%) ranged from good to moderate. As the variability of absolute high-speed distance values are greater, and therefore less reliable, their use as indicators of performance is reduced, suggesting that maximal mean analyses could be used as an alternative in the assessment of match running performance during competitive matches.