Use of a global positioning system to assess variation in velocity during track work in 3-year-old Thoroughbred horses

Abstract

Training is an important variable for determining athletic success. Nonetheless, there has been limited scientific evaluation of racehorse training programmes. This study aimed to determine the exact velocity during training, using a global positioning system (GPS) and see if there’s a significant difference between the average GPS velocity and the average velocity recorded by the trainer. Nineteen Thoroughbred racehorses age 3 years were followed through a traditional training and racing programme over a 5 month period. Daily GPS data together with the trainer’s timing and distance data were collected while the horses were trained. Data of all horses in training were analysed and compared for the months August and December. Parametric data were examined using a general linear model. Acceleration curves were smoothed using an autoregressive moving average algorithm. For canter workouts, the mean velocity timed by the trainer was significantly lower the mean GPS velocity(P=0.001) and mean GPS20 velocity (P=0.001). There was no significant difference between the true gallop work (mean GPS40 velocity) and the mean trainer’s velocity(P=0.169). The mean acceleration during canter was significantly lower than during gallop (P=0.001). There was a significant difference in variation in velocity between good and poor performing horses in both August (P=0.003) and December (P=0.002). A significant decrease of variation over the period of training was seen within the good performing horses (P=0.001), but not in the poor performing horses (P=0.059). The results demonstrated that there is no constant velocity during canter, but a constant acceleration. Therefore the velocity did not fluctuate around the mean velocity as we expected, but instead the acceleration fluctuates around the mean acceleration. This means that racehorses are trained in a different way than the trainer intended. The current method for measuring velocity during training does not reflect actual velocity for canter, but does reflect actual velocity for gallop work. During gallop work, the horses only reach racing velocity for a couple of seconds. Good performing horses have less variation in velocity compared to poor performing horses. They also have a decreasing variation in velocity over the period of training.