This study aimed to analyse the (i) concurrent validity, (ii) inter-unit reliability, and (iii) biological variability of a low-cost device called Pocket radar. Eleven men recreational soccer players performed 6 kicks to a soccer ball, whereas 13 men recreational tennis players conducted 10 shots to a tennis ball. All executions were simultaneously measured by two Pocket units and the Stalker radar (reference criterion). The within-subject variation among the executions was used for the biological variability analysis. The level of agreement and magnitude of errors included the intraclass correlation coefficient (ICC), Pearson's correlation coefficient (r), bias, and the smallest detectable change (SDC). A good agreement (ICC ≥ 0.98, r ≥ 0.98) and very low magnitude of error (SDC ≤ 7.70 km·h-1, bias ≤ 3.19 km·h-1) were found between both Pocket units and the Stalker, in soccer and tennis. Inter-unit analysis found limited technical errors (SDC ≤ 5.49 km·h-1, bias ≤ -0.93 km·h-1) and nearly perfect agreement (ICC = 0.99, r ≥ 0.98) in both sessions. These technical errors were lower than the variations due to the biological variability, in soccer (SDC = 2.47 km·h-1 vs. SDC ≥ 8.6 km·h-1) and tennis (SDC = 5.49 km·h-1 vs. SDC ≥ 21.95 km·h-1). These findings suggest the Pocket radar as a valid and highly sensitive tool for BV measurement.
Keywords: Soccer; sensitivity; speed; tennis; throwing.