The objective of the study was to devise methods that assess judging of horses in the dressage discipline. Classes run at the Olympic Games were analyzed. The data contained 30,055 marks awarded for single movements. The accuracy of the judging was measured with the Index of Disagreement that compares the horse ranking in a movement determined by a judge, to the ranking resulted from the marks awarded by other members of the jury. The authors proposed Horse Mean Placing Standard Deviation to test the consistency in judging particular horses. The study showed that the horse's scores were not sufficiently reliable and should be processed by checking the quality of judging before using the results both in breeding or sports and in further scientific analyses.