Built by someone who wanted better tools

I spent too many Saturday mornings scrolling through "locks" and "hammer plays" from people who never publish their actual win rate. The question I wanted answered — "when this system says 60%, how often does it hit 60%?" — never had a real answer. So I built a tool that answers that question for its own projections first, and everything else second.

Every projection you see in TrueProps has its math in public. Where a sportsbook has priced the same claim you get the model probability, the implied probability off the book, and the gap between them; where one hasn't, you get the projection on its own and we say so. The Results page is open without an account: every resolved projection, graded per market against the base rate, day by day. When a week goes badly it is in there too.

The product started with MLB because MLB has the densest, cleanest, cheapest public data. NFL followed for the 2026 season, with NBA and NHL on the roadmap behind it — the scoring engine is sport-agnostic by design, so adding a sport is a spec file, not a rewrite.

No number on the board is set by hand either. A model that is already serving a market is replaced only by a challenger scored beside it, graded on the same resolved outcomes, and only when the graded record says it should be. A version that does not clear that bar is recorded, and then either dropped or served without an edge claim — which is exactly what the NFL point spread and game total are.

If you bet recreationally, the goal is to help you understand the markets you are looking at and why. If you bet seriously, the goal is to be a trustworthy second opinion — the tool that's showing you what the math says, and telling you when the math is less confident.