Why I built this

Tennis betting content has a culture problem. Touts post their winners, delete their losers, and sell records nobody can check. The incentive is to look right, not to be right — and the two come apart fast when nobody is keeping score.

I spent years on the other side of that: building a model, testing it against the sharpest closing lines, and watching most of my own clever ideas die under honest evaluation. What survived was narrower than I'd hoped and more useful than I expected — a calibrated fair line, and a repeatable way to find the books that stray from it.

The conviction behind Baseline is simple: a model's claims should be checkable by strangers. Every pick logged before the match, graded against the close, kept forever — the record is the product, and I wanted the version of this site that I could never find as a customer. So I built it.

We don't beat Pinnacle — nobody reliably does. We find the books that disagree with it, and we grade every pick against the close, publicly — the receipts, including what we got wrong. The one quantified claim we make is scoped and confidence-intervaled, and a pre-committed kill-gate retires it automatically if the forward record goes flat.