Determinants of stride parameters in Thoroughbreds racing in Japan

To identify the race- and horse-level factors affecting stride parameters during Thoroughbred races in Japan.

Global Navigation Satellite System sensors were attached to 921 horses (1,189 starts) participating in 83 races, with distances ranging from 1,000 to 1,800 m, held from April through July 2024. Stride frequency and stride length were calculated from speed spectrograms at 3 racing phases (phase 1, 200 m after gate open; phase 2, 10 m after reaching the final straight stretch; and phase 3, 130 m before the finishing line). Additionally, 10 variables (race distance, surface type and condition, sex, age, finishing position, racing class, racecourse, body mass, and speed) were analyzed using a multivariable linear mixed model.

Mean (± SD) stride frequency, stride length, and speed were 2.36 ± 0.12 Hz (ie, strides/s), 7.30 ± 0.39 m, and 17.2 ± 1.15 m/s across all phases, respectively. Faster speed, geldings, longer race distance, and greater body mass were associated with longer stride length. Stride length was 0.11 m shorter on dirt than turf during phases 2 and 3 (P < .01) but not phase 1. The conditional R2 of the final model was 0.76, and the marginal R2 (ie, only fixed effects considered) was 0.55.

Moderate interhorse variability in stride parameters was found. In particular, racing phase and surface type affect stride parameters.

Racing phase, surface type, race distance, sex, and body mass in addition to speed should be considered when using stride parameters to evaluate performance and predict injury.