How hard is it to find a horse's data from public sources? What nearly 9,000 profiles reveal

Try to reconstruct the full picture of a horse — identity, studbook, pedigree, track record — using only free tools and public sources. It is harder than it looks. We tested it in practice, mapping 8,861 profiles of show jumping horses with results in Italy. The story is not about a single horse: it is that today, publicly, obtaining a reliable picture of a horse is laborious and incomplete — and that is a problem for the whole sector.

What we collected (and what it is not)

The 8,861 profiles come from equiresults.com, a public aggregator of competition results whose competitions originate from the FISE circuit, and power the information pages of our site. They are not a representative sample of the FISE population, and they are not the engine that generates the reports: they exist to build reference pages. The percentages measure how much of each piece of information is resolvable from public sources, not how much exists in closed archives.

Basic identity is findable. The rest, barely.

Name, sex and year of birth are almost always there. But everything that really matters to assess a horse — studbook, pedigree, the maternal line, even an international FEI identifier — is available only for a fraction, and coverage drops as the data becomes more valuable.

The critical gap is genealogy — the data that matters most to assess a horse. From public sources it can be reconstructed for only just over half of the horses: the sire is missing for 41%, the damsire for more than half (52%). And only about one horse in three (36%) has an international FEI identifier — the rest compete only nationally and are far less traceable.

The real obstacle: the data is fragmented

There is no single free source that gives the full picture. Results sit in one place, FEI identity in another, pedigree in studbooks or paywalled databases (such as HorseTelex), each covering only a slice. To build a serious profile you have to manually cross-reference several systems, each partial — and what you find is often inconsistent:

• Duplicate studbook codes: the same population under different labels (e.g. SF and SEFR for Selle Français, HOLST and HOLS for Holstein), distorting counts unless normalised.
• Entries that are not horses: competition statuses and placeholders ("A piedi" = on foot, "Cavallo al seguito" = led horse, "Temporanea" = provisional, "X") to recognise and filter out.
• Unequal documentation: imported horses have more traceable pedigrees than Italian-bred ones, so even what you do find is skewed.

Why it is a problem for the sector

Buyers, breeders, riders and agents make important — often financially significant — decisions about a horse. If the public picture is incomplete, scattered and partly unreliable, those decisions rest on fragile ground. Data opacity helps no one: not the buyer, not the serious breeder, not the transparency of the market.

This is why HorseReport makes sense

HorseReport.ai exists precisely to close this gap. Instead of leaving the user to chase different sources, at request time it makes real-time calls to multiple sources, aggregates, normalises and verifies them, and produces a reasoned analysis of the individual horse. What free tools turn into hours of cross-referencing — and still leave incomplete — becomes a readable, reliable report. See how we work, how we ensure source reliability, or browse the horses.

Methodology

Data drawn from 8,861 profiles of show jumping horses, collected from equiresults.com — a public aggregator of competition results whose competitions originate from the FISE circuit — to build the information pages of the HorseReport.ai site, extracted in June 2026. Non-horse entries (competition statuses and provisional registrations) were excluded from the raw total. The FEI identifier is counted only in its canonical format (3 digits + 2 letters + 2 digits, e.g. 103MK84): placeholder values the source uses for horses without an FEI ID were discarded. This is a static collection for reference, not the report-generation engine, which operates in real time across multiple sources at request time. Profiles were selected by availability in the public source, not by probability sampling: the figures should not be read as estimates of the federation population, but as a measure of data accessibility. "Resolved" means a non-empty, non-placeholder value after cleaning. The data may be cited with attribution to HorseReport.ai.