The inter-device reliability of global navigation satellite systems during team sport movement across multiple days

Objectives
(1) Determine the inter-device and inter-manufacturer reliability; and (2) investigate the variation in reliability over time for common global navigation satellite systems.

Design
Repeated measures.

Methods
A total of twenty 10-Hz devices manufactured by StatSports (n = 10, Apex Pro; StatSports, Newry, Ireland) and Catapult Sports (n = 10, Vector S7; Catapult Sports, Melbourne, Australia) were towed on a sprint sled during 8 × 40-minute team sport movement protocol over a 4-week period. The coefficient of variations for distance, velocity and acceleration/deceleration metrics were calculated to show dispersion of the data relative to the mean or median for each manufacturer and interpreted as good, ≤5%; moderate, <10%; and poor, coefficient of variation ≥10%. The coefficient of variation range described the variation in reliability and was interpreted as small, ≤5%; moderate, <10% and large, ≥10%. Inter-manufacturer agreement was represented as a Cohen d (±95% confidence interval) standardised effect size.

Results
Inter-device reliability for distance, peak velocity and average acceleration was good (coefficient of variation = 0.1 to 3.9%) for both manufacturers, with small variation across sessions. For most threshold-based acceleration and deceleration counts, StatSports devices showed good to moderate reliability, with moderate variation across sessions; Catapult showed good to poor reliability, with large variation across sessions. Inter-manufacturer agreement demonstrated moderate to very large effect sizes reported for most metrics.

Conclusions
Reliability was suitable and consistent for measures of distance, velocity, and average acceleration. StatSports devices generally possessed suitable reliability and consistency for threshold-based accelerations and decelerations, though Catapult devices did not. Most metrics should not be compared between manufacturers.