07.09.2026 · issue 1

You Can't Ride on Paper. You Can't Ignore It Either.

There's a line you hear in every breeding conversation in this sport. You can't ride on paper.

It's half right, which is the interesting half.

Paper does tell you something. Certain damlines throw certain horses. Certain stallions produce scope and certain ones produce rideability, and anyone who has bought a lot of young horses has a list in their head of which is which. That list is real. It is built on hundreds of horses seen over years, and it works often enough to be worth having.

What it isn't is a guarantee. Ten horses out of a good line are not ten good horses. Two might be, and one of those might be exceptional, and the rest go on with their lives at a level nobody writes about.

So the saying is really this: paper moves the odds. It does not decide the horse.

Results data does the same job. Same strengths, same ceiling.

Both ways of being wrong

The first mistake is the old one, and everybody already knows it. A horse has a brilliant record and you buy the record. Then it arrives and it isn't the horse you thought, because a record is a summary and a summary leaves things out.

The second mistake is newer, and it is the one I see growing.

People now look at a horse's results and believe they have read the whole horse. A certain rider had it for a season and it didn't produce, so the horse is finished — nobody needs to look at it again. A clear rate looks low, so it's not a 1,50m horse. Twelve rounds somewhere with a bad run in them, and the horse is written off before anyone has seen it jump.

That isn't data being used. That's a shortcut wearing data's clothes.

A record like that can mean the horse is limited. It can also mean the horse was over-faced, or under-campaigned, or in a programme that didn't suit it, or ridden by somebody who rides a different type. The number is the same in all four cases. Only one of them is about the horse.

Here's the falsifiable version, so you can hold me to it: I think the "so-and-so has already had it" rule costs this sport more good horses than it saves buyers money. I'll test that properly and publish whichever way it lands.

The cost of being wrong went up

I grew up in a dealing stable that competed at the top of this sport, where the job was to find horses, build them and sell them. That shaped how I see all of this.

Back then most of the sport happened close to home: national shows almost every weekend, the best riders in the country turning up at them, an amateur entering the same class as a professional.

Now the top of the sport moves every weekend, to a different country and often a different continent, with the horses flying in behind it. Prize money went from hundreds of thousands into the millions, and every round a horse jumps is streamed, filmed and public.

My honest position on all of it: I don't think the sport got worse. I think it got more expensive.

More money brought real things with it. Better footing, better veterinary care, better prize money, more eyes on the sport. Most of that is good.

But good young horses were never cheap, and prices have moved into a range where getting one wrong costs what a whole programme used to cost. That is the only reason any of this matters. When mistakes were cheap, you could afford to buy on instinct and be wrong a third of the time. Now you can't.

Twelve out of how many?

A horse goes clear a lot. What does "a lot" mean? Thirty percent? Fifty? Eighty?

At what height? Under which rider?

You can't answer that from the number alone, and that's the whole problem. It has a name.

The denominator. In horse terms: a stallion's record is meaningless until you know how many he sired. A dozen offspring jumping 1,60m sounds impressive. Twelve out of fifty is remarkable. Twelve out of a thousand is ordinary. The number on the page never changes — what it's divided by changes everything. The limit: a denominator tells you how often, never how good. Twelve 1,60m horses out of fifty could still be one exceptional mare line doing the work.

Same blank spot, every time. Thirty percent clear under one of the best riders in the world might be a horse underperforming badly. The same thirty percent under a green rider might be a very good horse being held back.

Same number. Opposite meaning.

We're not short of numbers. There are apps, platforms and rating systems, and EquiRatings has done genuinely good work bringing analytics into this sport. What we're short of is the thing you divide by.

Other sports already ran this experiment

Baseball was one of the first.

Long before anyone was talking about statistics, some teams built farming systems and developed their own players instead of only buying finished ones. It worked, and for a while it was a real edge.

Then everybody did it, and it stopped being an edge. It became the price of entry. Which is where our sport is standing right now.

We have our own version. Big breeders and top stables own a lot of horses spread across different riders and yards, and some riders employ riders themselves just to keep a roster behind them. That's a farming system, and it's a good one. It's also becoming normal, and normal is a smaller edge.

So baseball went looking at the game itself, hunting for what the market was pricing wrong. Then it produced the one number that makes the case better than any argument about data ever could.

In 2002 the Oakland Athletics and the New York Yankees each won 103 games. Oakland's payroll was about 40 million dollars. New York's was about 126 million.

Same wins. Roughly a third of the price.

You might not care about statistics. You will care about that gap. The distance between what something costs and what it's actually worth is the only thing I'm looking for in this sport.

What this newsletter actually is

It's me correcting my own biases in public.

I've been buying and selling horses long enough to have a full set of them. Rules of thumb I inherited, patterns I was sure about, numbers I quoted for years without ever checking what they were divided by. Some of them held up when I finally tested them against the results data. Several didn't.

The clear rate is the cleanest example. I had a figure in my head for what a good one was. Everyone in this sport has a figure in their head. Ask five people this week and you'll get five answers and no source. When I actually ran it by height, the real answer was not what I'd been carrying around — and it's different enough that it changes which horses look interesting.

That's the whole format. One belief per issue, tested properly, published whichever way it comes out. When I'm wrong I'll say so, because a newsletter that's never wrong isn't testing anything.

And I'm giving the corrections away rather than keeping them. Not out of generosity — a sport where everybody prices horses better is a sport where the good ones stop hiding in plain sight, and that's a better market to work in than this one.

Data will not replace an eye. It never has in any sport that adopted it. What it does is tell you where to point the eye, and stop you paying a premium for something the market has already worked out.

Next issue: what a good clear rate actually is at every height, and why the answer stops making sense above 1,35m.

That is the full scope. See you in two weeks.

Florian

P.S. Free account on The Show Jumper. Take one horse you're sure about and check what it's actually done — see how sure you still are.