Abstract: Data from horse-riding competitions recorded in Germany in 1976 and 1977 have been analysed to estimate genetic parameters for performance traits of riding horses measured in dressage, jumping competitions and trials. The performance traits analysed were logarithmic earnings per start, relative place number, and place value. The results are the following. 1. (1) Heritability and repeatability estimates for performance in dressage shows are 0.2 and 0.4 respectively. Corresponding estimates for performance in jumping competitions are 20% less. No genetic differences are found between stallions for performance in trials.2. (2) A selection index for estimating the breeding value of stallions was constructed by using the repeated performances of their offspring in dressage and jumping shows. For this purpose, performance data for at least ten progeny should be available. The correlation between the breeding values estimated from the dressage and jumping performances of the same stallions was approximately zero.3. (3) Reliable progeny-testing requires that the assumptions of mating stallions at random, selecting progeny randomly, and distributing them equally across environmental effects be fulfilled.4. (4) The genetic use of breeding values of stallions estimated from the performance of their progeny is opposed by the prolongation of the generation interval. This can be partly overcome by sampling young stallions and making use of the test results for young progeny only.