Investigation of the convergent validity and reliability of unit position differences of Catapult S5 GPS units in field conditions

Abstract

This research aimed to examine the validity and reliability of GPS units located in different positions. Nine recreational soccer players (age: 23.18 ± 2.21 years; height: 176 ± 7.65 cm; and body mass: 71.13 ± 4.67 kg) participated voluntarily in the current study. Athletes were tested through the team sports simulation cycle (TSSC) protocol. This protocol consisted of a total of 1200 m. Each lap consisted of a distance of 150 m, and the athletes were asked to perform eight laps. Two GPS units (OptimEye S5; Catapult Innovations, Scoresby, Victoria) were used for each athlete during the TSSC protocol. The first unit was positioned in the scapula location, and the other GPS unit was positioned in the center of mass (COM) location, and simultaneous data were recorded. A paired-samples t-test was used to determine the difference between the meter values measured in the field and the devices. The main finding of this research was that the player load parameters, which are derived from the accelerometer in GPS units, changes with the player’s position (total player load scapula − total player Load COM p≤ 0.001, Cohen’ d−2.449). There was no statistical difference between the other parameters (total distance covered, max velocity, deceleration max and acceleration max) examined in the study. CV% and SWC values showing the reliability of total distance covered scapula (CV% = 1.41; SWC = 0.28), total distance covered COM (CV% = 3.64; SWC = 0.73), total player load scapula (CV% = 2.29; SWC = 0.46), total player load COM (CV% = 1.83; SWC = 0.37), deceleration max scapula (CV% = 3.51; SWC = 0.70), deceleration max COM (CV% = 2.78; SWC = 0.56), Acceleration max scapula (CV% = 3.85; SWC = 0.77), and acceleration max COM (CV% = 2.74; SWC = 0.55) were within acceptable limits (CV% 5). The reliability of GPS units in different locations was investigated by CV% SWC analysis. It was found that all values in the scapula and COM locations were measured validly and reliably, but the total player load measurements were statistically different in the scapula and COM.