What a progeny record actually tells you

The line appears near the bottom of most stud advertisements, and it is usually the last thing anybody argues with.

Forty-one puppies to date. No known health issues.

It reads as a summary of evidence, and it is doing an enormous amount of work in the decision that follows. It is also two quite different claims wearing the same sentence. The first is a fact about registrations. The second is a fact about somebody's inbox.

I have been thinking about this since writing about shipped matings last week, because the further the buyer is from the animal, the more weight the progeny record has to carry — and it is, of everything in the file, the item least equipped to carry it.

A record and a count are not the same object

A progeny count is a number of animals produced. It is generally accurate, because registries hold it and nobody has much incentive to inflate something checkable.

A progeny record is a claim about how those animals turned out. That is a different kind of statement entirely, and there is no registry behind it. It is assembled from whatever came back to the stud owner — a Christmas card, a message about a hip score, a phone call at eleven at night about something that had gone wrong. What the stud owner knows is not what happened to forty-one dogs. It is what forty-one families chose to tell him.

Most of them told him nothing, and that is not evasion. It is ordinary life. People buy a puppy, the puppy is fine, and after the first year the breeder stops being someone they think to update — which is a shame for reasons I have written about elsewhere, and a real problem here.

The denominator nobody publishes

Ask what proportion of a stud's progeny the owner is actually in contact with and you will find that hardly anybody has the number, and that the honest ones say so.

It matters more than almost anything else on the page. "No known health issues across forty-one puppies" means something considerable if the owner is in regular contact with thirty-five of them and has hip and elbow scores for twenty. It means very close to nothing if he has heard from six.

The claim is identical in both cases. The evidence behind it differs by an order of magnitude, and the buyer has no way of telling which one they are being shown.

This is the single most useful question in a stud conversation, and it is not a hostile one: how many of them do you still hear from? A good stud owner has thought about this already and will tell you the number without being pushed. Some will tell you they have made a deliberate effort — a yearly email, a request for scores, a note in the contract — and that answer tells you a great deal about the operation.

Why silence is the weakest evidence we have

There is a specific asymmetry in what gets reported back, and it runs in the direction that flatters the stud.

A dog with a good hip score gets photographed and sent. A dog with a poor one often does not, particularly if the owner half-suspects the breeder will be upset, or feels it reflects on a decision they made. A dog that develops something serious at seven is owned by a family who by then may not have spoken to the breeder in six years and would not think to.

None of this requires anyone to lie. It requires only that good news travels more comfortably than bad, which it does everywhere else too.

So an absence of reported problems is compatible with a genuinely sound line, and it is equally compatible with a line nobody has been asked about. Both look like a blank space. The blank space is being sold as reassurance, and it is the one part of the file that cannot be verified by anyone, because there is nothing there to verify.

I would go further. Of everything a breeder can put in front of a buyer — a hip score, an eye certificate, a DNA panel, a coefficient of inbreeding — the unstructured progeny claim is the weakest, and it is routinely treated as among the strongest, because it sounds like experience rather than like paperwork.

What it can do, which is not nothing

I am not arguing that progeny records are worthless. Used properly they are one of the few things that tells you about an animal's actual output rather than its potential.

A progeny record is at its most informative when it is structured and partial rather than sweeping and clean. Nineteen of forty-one hip scored, with the range given, is far better evidence than forty-one with no known issues. A range that includes some indifferent scores is better still, because it tells you the numbers were not filtered on the way in.

It is informative when it contains something that did not go well. A stud owner who volunteers the litter that had a problem, or the one bitch that did not take, is telling you what their reporting standard is. That is the most valuable single thing they can give you, and it costs them something to give.

And it is informative when it is traceable. Registered names, so you can look at the animals yourself, ask a society, or find somebody who owns a half-sibling. A count without names is an assertion. A list with names is a starting point.

The horse version of the same problem

None of this is a dog-specific failing; it just presents differently.

With stallions, the record is more public — competition results are published, and a stallion with a competing progeny leaves a trail you can follow without asking anyone's permission. That is a real advantage, and it is why stallion progeny data is generally taken more seriously than canine.

But it carries its own distortion, and it is a sharper one. Competition results are a record of the progeny that made it into competition. The foal that did not come sound, the three-year-old that was quietly sold on, the one that never made a competition record at all — none of those appear anywhere. The published trail is a survivor's list, and the animals it omits are precisely the ones you were trying to learn about.

So the horse buyer gets better data about a worse-selected sample. Both species end up with the same problem from opposite directions: the pedigree and the progeny list together still cannot tell you what you actually want to know.

Three questions

If you take nothing else from this, these are the three that turn a number back into information. None of them is confrontational, and a stud owner worth using will not treat them as though they were.

How many of the progeny are you still in contact with? The denominator. Everything else depends on it.

Which of them have been health tested, and can I see the results rather than the summary? Scores, certificates, named animals. A summary is somebody's reading of the evidence; ask for the evidence.

Has anything come back that concerned you? The reporting-standard question. "Nothing at all" from a stud with sixty progeny is not a reassuring answer. It is an answer about their contact rate.

Where we are with this

I will be straight about the limits of what we have built, because this is an area where it would be easy to overclaim.

We can hold the structured part. Named progeny, health test results attached to the animal they belong to rather than to a claim about a line, scores presented as scores. That is most of what a listing on GenoVaq is: documentation a breeder already holds, verified and set out as comparable data rather than as a sentence in an advertisement.

What we cannot do is manufacture the follow-up. If a stud owner has heard from six of forty-one, no platform changes that, and any system that made the gap look tidier than it is would be worse than the advertisement it replaced. The free lifelong health record is our attempt at the actual fix, which is a long one: if the record belongs to the animal and travels with it through every owner and every vet, then in ten or fifteen years the progeny question has a real answer behind it instead of a recollection.

That is a slow thing to build and I am not going to pretend otherwise. In the meantime, the honest position is the one I would want from the other side of the transaction: the number is a count, the reassurance is a recollection, and it is entirely reasonable to ask which is which.

— Rene