Why we timestamp our own prices
It is the least glamorous thing we do and the most expensive. Every price we use was written down before the race was run, with the time attached, and never edited afterwards.
A price is not a fact, it is a moment
People talk about the price of a horse as though it were a single number sitting in a box somewhere. It is not. A market is a thing that moves continuously right up to the off, and any number you take from it is a photograph, not a portrait.
Which means the honest version of a price always has two parts. What it was, and when you looked. Drop the second part and you have something that sounds precise and is not.
That is fine until you want to use the price as a control. And a control is exactly what we use it for. Nearly every study we run works by holding the price constant and asking whether some other piece of information still adds anything. If the price is soft, the control is soft, and the finding is worth nothing.
What contamination actually looks like
This is where people expect an accusation, and there is not one to make. Almost nobody is faking prices. Contamination is nearly always innocent, which is what makes it dangerous.
A record that is assembled after the race has been run has been through hands that knew the result. Gaps get filled with the version that makes sense in hindsight. A non runner is withdrawn and the whole book reprices, and the archive keeps the tidy version rather than the messy one that existed at the time you claim to have looked. Somewhere along the line the number that got kept is the one that agrees with what happened.
None of that requires bad faith. It only requires a human being deciding, after the fact, which of two plausible numbers to keep. Hindsight does not feel like bias from the inside. It feels like tidying up.
The tell is a study where the price always seems to have been read at the perfect moment. If a result is sensitive to when the price was taken, and the price was taken afterwards, no amount of statistics will patch the hole.
What it costs
The discipline itself is embarrassingly simple. Record the market at a fixed clock time before the race. Stamp it. Store it. Never rewrite it, even when it turns out to be inconvenient.
The cost is that you cannot go back. A checkpoint you missed is missed permanently. There is no clever code that recovers a price nobody wrote down, and there is no version of the archive that will honestly sell it to you later. Miss a race and that race is not in the study, forever.
Worse, the failure is silent. Nothing breaks. No alarm goes off. The system carries on looking healthy and the row simply is not there, which is why we treat a missed snapshot as a defect to be recorded rather than a gap to be shrugged at.
And it means the dataset grows at the speed of the racing calendar and not one day faster. You cannot buy history. You cannot download it. You have to turn up and wait, which for a research company is a strange and humbling business model.
Why almost nobody bothers
Because reconstructing is free, instant, and looks identical in the finished article. Both approaches produce a column of numbers. Nobody can tell which column is which by looking at it, and the reconstructed one arrives this afternoon rather than in a year.
There is a quieter reason too. If you record prices live, you are making a promise you can be held to. The claim becomes checkable, and checkable claims are more likely to come back and embarrass you. Reconstruction is not just cheaper. It is safer.
Which is an odd sort of argument for doing it the hard way, but it is the honest one. We would rather be wrong in public on a record we cannot quietly adjust than right on a record nobody could ever audit.
What it buys
One thing, and it is the only thing that matters. It lets us say a comparison is fair.
When we compared horses that had shortened against horses that had drifted at the same final price, and the advantage disappeared, that result is only worth reading because both sides were recorded in the same way, at the same clock time, before anybody knew anything. Take away the discipline and the finding becomes an opinion with numbers attached.
The same goes for the number we quote most often. Across 60,671 races with one clear favourite, the favourite won 35% and did not win 65%, settled against starting prices. That figure is unremarkable on its own. What makes it usable is that the prices behind it were not chosen after the result was known.
None of this makes us right. It makes us checkable, which is the most any research company should be claiming before it has something worth defending.
Related research
- Why we publish the ideas that failedThree of the first five beliefs we tested died. Keeping the dead ones in public is the only honest way to ask anyone to trust the ones still standing.
- The race that changed the questionA friend's horse had everything in its favour at Newmarket. The race went somewhere else entirely, and the question I left with was not about the horse.
- What two months of testing actually taught usThree of the first five beliefs we tested did not survive. Here is what lived, what died, and what changed about the question we were asking.
Elite Pass publishes research, not guaranteed outcomes. Findings are measured against the market and remain subject to replication and sample size. The evidence room and the public ledger carry the full record.