How reliable are the tactical measures obtained in soccer small-sided games? A test-retest analysis of observational instruments and GPS-based variables

This study investigated the test-retest reliability in both within and between-session reliability of commonly adopted tactical measures in small-sided games. GPS-based and observational tactical variables from three instruments were collected from a 4 versus 4 small-sided game played by 16 U-17 elite players. The games were played three times in the same training session (within-session reliability) and repeated after 1 week (between-session reliability). The reliability of the data was analyzed by calculating the Intraclass Correlation Coefficient (ICC) and the Standard Error of Measurement (SEM) for all dependent variables. Results showed good-to-excellent reliability and a low error for GPS-based tactical variables, mainly the width and length positioning (ICC > 0.9, SEM < 10%, p < 0.05). Observational variables showed poor-to-moderate reliability, except the offensive performance measured by the FUT-SAT in the between-session analysis. It is concluded that GPS-based tactical variables show acceptable reliability and might be used for talent identification and training outputs assessment in soccer. The variability inherent to small-sided games is associated with both the open nature of the task and poor measurement quality. Increasing the quality of the instruments will likely provide a better view of the actual variability of this training tool.