Determinants of stride parameters in Thoroughbreds racing in Japan.

Overview

• This study investigated which race-related and horse-specific factors influence stride characteristics such as stride frequency and stride length in Thoroughbred racehorses competing in Japan.
• Data from Global Navigation Satellite System (GNSS) sensors attached to 921 horses during 1,189 race starts across 83 races were analyzed to identify determinants of stride parameters at different stages of a race.

Research Objective

• To identify race-level and horse-level factors that affect stride parameters (stride frequency and stride length) during Thoroughbred races held in Japan.
• Stride parameters were examined across multiple race phases and a broad range of variables including speed, race distance, surface type, horse sex, age, body mass, and race conditions.

Methods

• Subjects and Data Collection:
  • 921 Thoroughbred horses participated with 1,189 race starts recorded.
  • Data collected from 83 races held between April and July 2024.
  • Race distances varied from 1,000 to 1,800 meters.
  • GNSS sensors were attached to horses to measure movement parameters during the races.
• Stride Parameter Calculation:
  • Stride frequency (strides per second) and stride length (meters) calculated using speed spectrograms.
  • Measurements taken at three distinct racing phases:
    • Phase 1: 200 meters after gate opened.
    • Phase 2: 10 meters after entering the final straight stretch.
    • Phase 3: 130 meters before the finish line.
• Variables Analyzed: Ten variables included
  • Race distance
  • Surface type (dirt or turf) and surface condition
  • Sex of the horse (gelding, mare, colt, etc.)
  • Age
  • Finishing position in the race
  • Racing class (level of competition)
  • Racecourse location
  • Horse body mass
  • Speed (measured during race phases)
• Statistical Modeling:
  • A multivariable linear mixed model was used to analyze the relationship between these variables and stride parameters.
  • This approach allows for both fixed effects (measurable variables) and random effects (differences between horses) to be accounted for.

Key Results

• Mean Stride Metrics:
  • Average stride frequency: 2.36 ± 0.12 strides per second.
  • Average stride length: 7.30 ± 0.39 meters.
  • Average speed: 17.2 ± 1.15 meters per second.
• Factors Associated with Longer Stride Length:
  • Faster speed during race phases.
  • Gelding horses compared to others.
  • Longer race distances.
  • Greater body mass of the horse.
• Effect of Surface Type:
  • Stride length was significantly shorter (by approximately 0.11 m) on dirt tracks compared to turf during phases 2 and 3.
  • No significant difference in stride length between surfaces during phase 1.
• Model Performance:
  • The final model showed a high level of explained variability with a conditional R² of 0.76 (accounting for fixed and random effects).
  • Marginal R² (only fixed effects) was 0.55, indicating over half the variability could be explained by measured variables alone.
• Between-Horse Variability:
  • Moderate variability in stride parameters was observed among different horses, demonstrating individual differences beyond measured factors.

Conclusions and Implications

• Stride parameters in Thoroughbred racehorses are influenced not only by speed but also by race phase, surface type, race distance, horse sex, and body mass.
• These factors should be considered when analyzing stride data for performance evaluation or injury prediction.
• Understanding how these determinants interact helps trainers, veterinarians, and researchers better interpret stride characteristics to optimize performance and welfare.
• The study provides valuable insight into the biomechanical adaptations horses make in different race conditions, informing strategies for training and race planning.