To investigate the factors affecting the utilization of retired Thoroughbred racehorses in equestrian disciplines, Bayesian linear mixed models were separately fitted using rank, round time, and obstacle faults from show-jumping competitions restricted to retired Thoroughbred racehorses as dependent variables, with the interaction between horse sex and the interval from race retirement to competition (as a proxy for transition training to show-jumping) as a fixed effect. When the interval was short (≤1 year), the estimated marginal mean of rank was statistically significantly lower in stallions (0.26) than in mares (0.41) and geldings (0.39). However, ranking improved with longer intervals in all sexes, with the greatest improvement observed in stallions, and the significant sex-related differences disappeared at the 3-year interval, suggesting an effect of transition training on ranking. Round time improved significantly with longer intervals in all sexes, consistent with the ranking pattern; significant improvement in obstacle faults was observed only in stallions and geldings. The explanatory power of the models, including major random effects, rider, horse ability, sire and affiliation after retirement, was moderate (conditional R2: 0.40–0.65), whereas that of the fixed effects was small (marginal R2: 0.02–0.07), indicating the multifactorial nature of success in competition.
Naito et al. (Wed,) studied this question.