Breeding Values and Performance Data

Breeding values are statistical estimates of the genetics a stallion or mare passes on — the KWPN’s fokwaarden and the German Zuchtwerte the systems buyers meet — expressed as indexes against a population average (typically centred on 100) with a reliability percentage attached, and computed from the measured performance of relatives and offspring. They are genuinely informative about populations of future foals and only weakly about the individual horse standing in front of you — which fixes the buyer’s weighting exactly: real weight on breeding purchases and young stock, context on everything else, and never a substitute for the horse’s own evidence.

This page decodes the quantitative layer for buyers; the qualitative use of pedigree is the pillar’s subject, and the titles that summarise selections into words are the predicate decoder’s.

What the numbers are

The studbooks run genetic-evaluation programmes on the data their systems generate — inspection scores, performance tests, sport results across thousands of horses — and publish, for breeding animals, index estimates of transmissible quality: overall dressage indexes and component traits (gaits, sometimes conformation and rideability dimensions) depending on the book’s model. Two conventions matter for reading them:

The scale. Indexes centre on the population average (100 in the common convention), with points expressing standard deviations of estimated genetic merit — a stallion indexing well above the mean is estimated to sire above-average foals for that trait, in that population, on average.

Reliability. The percentage beside the index is the estimate’s statistical confidence, driven by data volume: a young stallion’s early index rests on his own results and relatives (modest reliability, large future swings); a proven sire’s rests on hundreds of measured offspring (high reliability, stable number). The reliability figure is not decoration — an impressive index at low reliability is a forecast, and the trade’s history of young-stallion indexes deflating as offspring data arrives is exactly why the convention exists.

The adjacent numbers buyers meet

Young-horse and studbook test scores — IBOP, stallion performance tests, mare tests: measured performances of the individual, on a day, in a protocol. More directly informative about that horse than any index, and still a snapshot — the gaits page’s reading rules (weight walk and canter scores; a score rates the day, not the career) apply verbatim.

Young-horse class results — Bundeschampionat and world-championship placings: expert evaluations under competition conditions, the market’s premium evidence on youngsters, carrying the same maturation caveat the Vivaldi and Sandro Hit profiles attach to everything spectacular and young.

WBFSH sire and studbook rankings — computed from offspring results in international sport: the conversion scoreboard, informative about lines and books at the elite tail (the breeds pillar’s calibration), and mute about individuals.

The taxonomy worth internalising: indexes estimate genes, test scores measure the individual once, sport records document the individual repeatedly — and evidential weight for a purchase runs in exactly that ascending order.

How much should a buyer weight them?

By purchase type, per the pillar’s tier rule with numbers attached:

Breeding purchases — full weight. Buying a broodmare or breeding to a stallion is buying the indexes’ subject matter: here the fokwaarden/Zuchtwerte, read with their reliabilities, are primary evidence, alongside the predicates and WFFS status.

Foals and unbacked youngsters — meaningful weight. The parents’ indexes are a legitimate probability input where the horse has shown little — one input beside the dam herself, the damline and the loose evaluation.

Riding horses — context only. By the time a horse works under saddle, its own gaits, test scores and record outrank any estimate of its parents’ genes; a seller leading with sire indexes on a six-year-old is leading with the weakest available evidence, which is itself information.

Never — as rideability-for-you or soundness predictions. The models estimate what they measure; the match with your riding and the state of this horse’s legs are measured by the trial ride and the PPE, full stop.

Reading them without being read to

Three practitioner’s habits: compare within systems only — a Dutch fokwaarde and a German Zuchtwert are different models on different data and do not convert; check the reliability before the index — the percentage decides how seriously the number deserves taking; and source from the registries — the books publish the current values, adverts quote the flattering vintage, and the database habit takes minutes as always.