Using the exact rank within a complete competition group leads to a co
nsiderable increase of information compared with using earnings. The m
odel to estimate genetic parameters and breeding values requires to co
nsider the effects of age, sex and competition. Environmental similari
ty of repeated performances of the same horse should be considered wit
h the help of a random, permanent, environmental effect. The estimated
genetic parameters are very low. The ranking traits have a clear high
er heritability and repeatability than the logarithm of earnings. Ther
e was no antagonism between show jumping and dressage but correlated g
enetic progresses are also not to expect. A breeding program should us
e breeding values on the basis of pedigree information, own performanc
es and performances of offsprings.