Why We Publish Our Track Record — and Never Edit It

Here is an uncomfortable fact about research tools: almost none of them show you their history.

You can find no shortage of products that will tell you, today, what looks interesting. Very few will show you what they said six months ago, next to what happened afterward, in a form they can't quietly revise. The incentive runs the other way — a tool's past outputs are a liability to be managed, trimmed, or simply never mentioned.

We decided to do the opposite, and to make it structural.

What the record is

Every month-end, AIR locks in a public entry on our track record: the five widest Discovery Gaps our engine identified at that moment — the companies where, by our methodology, the distance between what was disclosed and what was noticed looked greatest. Each entry is timestamped when it's made.

Then comes the important part: we never edit it. Not the companies, not the scores, not the dates. The record only grows. Where the engine's reading turns out to have aged badly, that stays visible too, permanently, next to everything else.

Why we built it this way

Because a methodology can be copied, but a record can only be earned. Everything else about AIR could, in principle, be replicated by a competitor with enough engineering: the nightly reading, the classification, the scoring framework. What cannot be replicated is months of timestamped, unedited entries — those exist only at the pace of one per month, earned in public. We consider the record, not the engine, our most defensible asset. That's only true if the record is beyond suspicion.

Because unfalsifiable claims are worthless — including ours. Any tool can claim its method works. The claim means nothing unless it's checkable by someone who doesn't trust us. A timestamped, append-only public record is checkable by design: you don't need to take our word for anything, you can go and look.

Because it keeps us honest with ourselves. An unedited public record changes how you build. Every scoring decision our engine makes may end up in an entry we can never take back. That's exactly the pressure a research methodology should operate under.

What a track record is — and isn't

Honesty cuts both ways, so let's be precise about what the record proves. It is evidence of process: that we said what we said, when we said it, and let the results stand. It is not a promise about the future, and a stretch of months — good or bad — is a small sample from which to conclude much of anything.

What it gives you is the ability to judge for yourself, on evidence we can't retouch, whether our way of reading the market's least-read layer deserves a place in your research. That's the standard we'd want from any tool we relied on, so it's the standard we hold ourselves to.

The record is public, free, and grows by one entry every month. It will still be there — all of it — whenever you want to check.

AIR is a research tool, not investment advice. We surface and score public information; decisions are yours.