Agreement between official race data and a high-end sports tracker for a Paralympic rower

Abstract Race data containing speed and stroke rate per 50m, provided by World Rowing, were compared with measurements using an ASI AdMos sensor, which is a high-end sports tracker, to assess the agreement between the two measurement systems. Twelve races in 2022, 2023, and 2024 were considered. All measurements were of the same rower competing in the Paralympic PR1 W1x class. The race data speed was compared with speed calculated from the GNSS data of the AdMos sensor, while stroke rate was compared with the stroke rate calculated from the surge acceleration of the IMU data of the AdMos sensor. Comparisons were made for each 50m segment. Additionally, official 500m split times were used to assess the accuracy of both the race data and the AdMos measurements. The absolute difference between the race data and the AdMos measurements was within rounding tolerance in 312 of 480 segments (65%) for speed and 410 of 480 segments (85%) for stroke rate. The comparison of mean speed per 50m segment revealed a mean absolute difference (MAD) of 0. 05ms ^-1 (mean absolute relative difference: 1. 6%). The speed difference was largest in the first 50m of the race, with a MAD equal to 0. 39ms ^-1 (12%). If comparing the race data with the AdMos speed of the last stroke before each segment end, the difference in the first 50m reduced to 0. 12ms ^-1 (3. 3%). The MAD in stroke rate was 0. 3 strokes per minute (1. 0%). Compared to the official 500m split times, the MAD of the AdMos was 0. 3s (0. 2%), whereas calculation of the 500m splits from the race data yielded a MAD equal to 1. 0s (0. 7%). World Rowing’s official 50m race data showed good agreement with the high-end sports tracker in most segments, but the systems diverged considerably for speed in the first segment. Practitioners should verify data characteristics before using official race data for high-precision biomechanical or pacing analyses.