The possibility of using quantitative kinematic traits as indirect selection criteria for sport performance could be beneficial to perform an early genetic evaluation of the animals. The genetic parameters for objectively measured kinematic traits under field conditions have been estimated for the first time, in order to potentially use these traits as indicators of gait quality in future selection of the Lusitano breed. The repeatability within three different types of training (dressage, bullfighting and untrained) was also discussed. A total of 176 males (4 to 14 years old) were recorded at trot in hand using a 3D videographic system. The speed and 10 kinematic traits were studied (one temporal, two linear and seven angular variables). The genetic parameters of the kinematic variables were estimated using VCE software. The heritability estimates were moderate to high (0.18 to 0.53). The stride length and the forelimb angular variables presented the highest heritabilities (0.49 to 0.53), whereas the hindlimb angular variables revealed the lowest values (0.18 to 0.40). More than half of the genetic correlations were moderately to highly positive (mostly 0.20 to 0.70; up to 0.88 between hindlimb traits). The dressage and bullfighting groups presented the highest repeatabilities (over 0.6) in the majority of the traits, maybe because of the acquired gait regularity expected in animals subjected to specific training, and suggesting a greater influence of the individuals over the kinematic traits studied in these two subpopulations than in the untrained subpopulation. The longer swing phase duration and the larger range of motion of the elbow, hock and pelvis joints observed in the dressage group may indicate a better gait quality of this group, according to FEI (International Equestrian Federation) standards. The bullfighting and untrained groups were more similar to each other in terms of kinematic traits. Selection of young horses for characteristics such as stride length and the hindlimbs traits can apparently contribute to further genetic improvement of the performance of Lusitano breed.

For many years, the breeding value estimation for Swedish riding horses has been based on results from Riding Horse Quality Tests (RHQTs) of 4-year-olds only. Traits tested are conformation, gaits and jumping ability. An integrated index including competition results is under development to both get as reliable proofs as possible and increases the credibility of the indexes among breeders, trainers and riders. The objectives of this study were to investigate the suitability of competition data for use in genetic evaluations of horses and to examine how well young horse performance agrees with performance later in life. Competition results in dressage and show jumping for almost 40 000 horses from the beginning of the 1960s until 2006 were available. For RHQT data of 14 000 horses judged between 1988 and 2007 were used. Genetic parameters were estimated for accumulated competition results defined for different age groups (4 to 6 years of age, 4 to 9 years of age and lifetime), and for different birth year groups. Genetic correlations were estimated between results at RHQT and competitions with a multi-trait animal model. Heritabilities were higher for show jumping than dressage and increased with increasing age of the horse and amount of information. For dressage, heritabilities increased from 0.11 for the youngest group to 0.16 for lifetime results. For show jumping corresponding values increased from 0.24 to 0.28. Genetic correlations between competition results for the different age groups were highly positive (0.84 to 1.00), as were those between jumping traits at RHQT and competition results in show jumping (0.87 to 0.89). For dressage-related traits as 4-year-old and dressage competition results the estimated genetic correlations were between 0.47 and 0.77. We suggest that lifetime results from competitions should be integrated into the genetic evaluation system. However, genetic parameters showed that traits had changed during the over 35-year period covered due to the development of the sport, which needs to be considered in future genetic evaluations.

There are many instructional texts for dressage management and training, but little scientific knowledge of these practices. It was hypothesized that there would be an effect of horse competition level on management and training. The objectives were to record rider, horse, management and training information from UK dressage horses and to investigate relationships between the variables. A questionnaire-based study was undertaken in all the UK-registered dressage horses. Data from 2554 respondents showed most riders in groups 7/6 with horses competed at novice medium and trained one level higher. Horses were 164–171 cm and warmbloods were the predominant breed. Age and time in competition increased in training and competing from Preliminary to Prix St George. Horses competed twice per month except at Grand Prix (once per month). Most horses had 15–30 h turnout per week. Ninety-five per cent of horses received non-dressage exercise; hacking was the most common, then lunging and jumping. Non-elite horses were significantly more likely to be hacked/jumped than elite. Horses received dressage training three to four times a week, mean warm-up of 16 min and cool-down 11 min at all levels. Mean training duration was 36 min: 21% walk, 45% trot and 32% canter. Preliminary, Novice, Intermediare I and Grand Prix horses trotted most and cantered least in training. Overall, time spent in transitions (30–39%), specific movements (30–39%), working paces (40–49%), collected (10–19%) and extended (0–9%) varied between levels; elite horses spent more time in collected and less time in working than non-elite. Overall, there were differences in training and management practices between competition levels. This information provides baseline data on which to provide future advice.

Warm-up of 267 competitors at British Dressage affiliated competitions was observed, including competitors at novice (N) (n = 104), medium (M) (n = 65), Prix St Georges (PSG) (n = 60) and Grand Prix (GP) (n = 38) levels. Competitions were classified as local (n = 103), regional (n = 57) and national championship (n = 107) events. Overall, the mean warm-up duration for competitors at dressage competitions was 29 min 53 s. Total warm-up duration was 25 min 23 s ± 10 min 2 s (mean ± SD) at N level; 31 min 32 s ± 11 min 32 s at M level; 32 min 53 s ± 11 min 19 s at PSG and 34 min 34 s ± 10 min 10 s at GP. Mean proportion of walk, trot and canter at each level was N: walk = 39.26%, trot = 40.31%, canter = 20.43%; M: walk = 43.77%, trot = 32.54%, canter = 23.69%; PSG: walk = 38.53%, trot = 31.03%, canter = 30.43% and GP: walk = 38.79%, trot = 33.26%, canter = 27.95%. There was no effect of rider experience, but level and type of competition affected the proportion of time spent in different paces and total time of warm-up, which was increased at higher levels and championships. Increased warm-up time and specific warm-up design were positively associated with final score at novice and Prix St Georges levels.

The objective of the study has been to examine whether the freestyle dressage classes are judged reliably and to elaborate a method assessing the consistency of judging. The data contained 13 000 marks of the ten best-ranked horses at nine Grand Prix classes and four Intermediate I classes from ten international competitions. The marks concerned 119 horses in total and were given by 37 judges. A method for evaluating the consistency of judging has been proposed. The index of disagreement (ID) assesses the disagreement of ranking by an individual judge relative to the general ranking based on the sum of marks awarded by five judges. The results show that the mean ID of individual judges is highly differentiated. The ID is influenced by the judge's position in the arena. The consistency of judging was lower in Intermediate I Freestyle Tests than in Grand Prix Freestyle Tests and it varied in different competitions. The conclusion of the study is that the results of the freestyle dressage classes are often biased. The present system of judging should be permanently checked. The offered method of evaluating agreement between judges' rankings may help to improve the consistency of judging and the reliability of the horses' scores.

Subjective bias by judges within the dressage phase of eventing competition is problematic if correct evaluation of horse and rider performance is to be undertaken. The present study examines dressage penalty scoring (penalty scores are awarded for completion of a pre-set series of dressage movements) within a population of novice event horses (n = 2471). Between May and June 2003, 22 novice events within the UK were analysed, and at each event up to six competitions or sections (at the same competitive level) were run. The whole population data structure was normal in distribution, Pearson's skewness 0.314 (Kolmogorov–Smirnov Z = 1.855, P>0.05 from zero) and kurtosis at 0.425. The mean score recorded for the whole population was 37.82 penalty points (standard deviation = 5.65 and standard error = 0.11). A number of effects were observed within the population. Between-events effects were observed in three of the 22 events studied (P<0.01). Of those events running three or more sections (n = 16), within-event effects were observed for 10 events as differences (P<0.01) in mean scoring patterns between sections. All events running two sections (n = 4) were observed to exhibit differences (P<0.02) in mean scores. Differences (P>0.01) between the dressage test used at events were determined using post hoc Bonferroni tests. However, these differences were not found to have a confounding effect on the between-event observations. The study indicates that additional methodologies need to be implemented to ensure that accurate and impartial evaluation of event horses is conducted.