Schrurs, Charlotte B.
(2024)
Using data to track racehorse physiology during training and racing.
PhD thesis, University of Nottingham.
Abstract
This thesis investigates multiple aspects of racehorse training, physiology, and performance. The overarching objective of the work is to explore the various factors that may influence the development of racehorses in-training, where a dearth of information exists, historically due to technological limitations. Here, by using a validated fitness tracker, the 'Equimetre™' then multiple aspects of racehorse physiology (heart rate, heart rate recovery), speed (GPS tracked), stride and on-course performance (results-based) could be analysed in large cohorts of racehorses. All studies were retrospective and observational.
In the first study I investigated how training intensity (slow canter to hard gallop) and training surface (sand, fibre, or turf) impacts upon parameters like heart rate, heart rate recovery, stride length/frequency in the racehorses in-training. Additionally, I also explored the influence of horse age and sex on these parameters. The dataset included 509 Thoroughbred racehorses from Australia and France, of varying age, sex (including geldings), and training conditions. In summary, the precision of the Equimetre was excellent (coefficient of variation from 1-5% for locomotory parameters, 3-6% for speed and 5-15% for cardiovascular parameters). The data revealed the marked effect of training surface – sand being mainly used for canters, but markedly shortening stride relative to turf – and how racehorses reach maximum heart rate even at relatively slow speeds (e.g., often at only ‘hard canter’). Heart rate recovery was mostly influenced by training intensity and did not appear to ‘improve’ through the season, as a marker of ‘fitness’. There was clearly a zone of maximal cardiac flexibility, at hard canter-to-slow gallop, where responses to training were greatest, perhaps offering the opportunity to use as a basis for setting training zones, as opposed to maximal heart rate, which offered little in terms of between or within horse performance difference. Establishing these data as a reference point, I could then begin to look at training data in a new light; for example, with information on the sex of each training rider, and their experience (professional or non-professional) I could examine whether sex of the rider had any overt influence on racehorse physiology in training or performance on track.
Thus, my second study explored the effect of the sex of rider on racehorse cardiovascular (heart rate, heart rate recovery) and locomotory (stride length and frequency) parameters in training (slow canter to hard gallop) and, ultimately, on the track as race performance in both Australia and the United Kingdom. The dataset consisted of 530 Thoroughbred racehorses that were ridden by randomly allocated work riders of varying sex (male, n= 66; female, n=37) and experience (including registered professionals, n= 43). In training, female riders, on average, participated in more training sessions, but usually at slower intensities (e.g., canter rather than gallop) than their male counterparts. Racehorse speed or the average time taken to cover a furlong (200m) did not differ according to the sex of the rider onboard, nor did stride length adjusted according to training surface (turf or sand). While heart rate and peak heart rate increased with training intensity, as expected, this was not influenced by the sex of the rider. Interestingly, the rate of recovery of heart rate in the horses (from peak to 15 minutes post exercise), appeared to be affected by the sex of the rider, but this depended on the anticipated training intensity vs. the actual training surface: horses recovered slower on turf after slow canters, but faster on sand after fast gallops with male riders, suggesting that the riders either transmitted anticipatory information to the expected training intensity to the horse or the racehorses responded differently according to sex of the rider (i.e. less anticipating a hard gallop on turf with a female jockey). In regard to race data, significantly more male jockeys are represented in the starting gates on race-day (in complete contrast to data during training). However, race-day performance (win percentage) between male and female jockeys was not different. With these aspects considered, I then considered what evidence there was for the race specificities of each racehorse; for example, is a racehorse that only participates in short-distance races, a pre-designated sprinter that can be identified early in its training sessions (e.g., by having a shorter stride) relative to a middle- or long-distance racehorse. With three-generation pedigree information for 421 racehorses, I could then simultaneously explore to what extent the unique characteristics of racehorses are pre-determined by genetics or are developed through the early training sessions.
Therefore, my third study delved into understanding racehorse stride patterns (peak, length, and frequency) by comparing differing distance profiles (‘sprinters, milers, or stayers; according to known race distance). I also investigated the heritability of stride by extracting a comprehensive pedigree database and finally also instigated the correlation of stride with performance prediction through previously collected race-results. The dataset comprised of 421 Thoroughbreds, from a single racing yard in Australia, of varying age, sex (including geldings), and training conditions. Stayers, although fewer and mostly older horses, competed in fewer races but were more successful than sprinters. For the race-pace gallops held on turf, then differences in stride were marked: sprinters presented shorter peak stride length but of higher frequency than stayers. When replicating race-day speeds and conditions from starting gates on turf (‘jumpouts’) over consecutive furlongs, then sprinters were significantly faster than stayers. However, no substantial evolution in stride was reported over the course of continuous training sessions for any individual horse. With everything considered, then colts generally had greater win percentages. Although the effect was relatively small, longer stride length increased the chance of a horse winning and/or finishing in the top three of a race. Peak stride length and frequency were considered moderately heritable, which could theoretically, when coupled to experience and objective data, aid trainer decision-making for the selection of horses. Finally, I explored the extent to which in-training speed and heart rate recovery could differentiate successful racehorses (those that won races, including black-type) from horses that did not.
For this final study, I accumulated sufficiently large training and race data from the same 485 racehorses in Australia in which the training sessions were restricted to race-pace workouts on turf only (i.e., hard gallop and jumpouts) conducted close to and before (within two months) races. Speed, heart rate peak and recovery at maximal running intensities were similar to previously described data in the first chapters. During standard gallops, then the greatest recovery of racehorse heart rate was identified within the first minute following exercise, designated as the early phase of heart rate recovery. Horses completing jumpout sessions, took longer to recover than others during the first three minutes after exercise, as expected for a race-pace simulation exercise. Interestingly, horses appeared to recover better (+/- 5-6 bpm), at any intensity during warmer workouts, relative to cold, with no effect or interaction with humidity. Racehorse speed over the final 600m in-training when designated as ‘fast’ versus ‘slow’ was marginally predictive of race performance. Again, as previously described, colts and stayers also tended to win more racehorses in our dataset relative to mares/fillies and sprinters, respectively. I observed very little predictive ability of heart rate measures on race performance. Hence, in-training physiological data is useful for the trainer and/or sports physiologist to monitor the general health and well-being of racehorses and their general condition but does not evidently predict with certainty the chances of a given racehorse winning a selected race.
Overall, these comprehensive, large-scale studies lay the foundation for future research on in-field racehorse exercise physiology. They serve as a valuable resource for trainers and contribute to informed decision-making in the horseracing industry. The findings have the potential to enhance racehorse health and safety, further empowering the future and sustainability of the sport.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
