College football analytics, honest about what it can and can’t tell you.
Fourth Down Data is a research platform for analytical CFB fans. Queryable tools, transparent methodology, no predictions, no picks. We surface what’s true and let you decide what it means.
What makes this different
Most sports analytics products give you their model’s answer. We don’t.
We built queryable interfaces so you can ask the data your own questions instead of accepting someone else’s framing. The Query Builder runs your filters across every FBS game since 2017. The Archive surfaces historical games most analytically similar to any matchup you pick. The Read assembles each week’s most interesting analytical angles automatically.
When something doesn’t work, we say so. Our Reality Check tool exists because we tested whether divergence between record and model predicts future games — it doesn’t. We published the empirical null finding rather than hiding it. That kind of honesty is rare in sports analytics, and it’s what we’re building the rest of the product around.
We’re not a betting site. We’re not a picks service. We don’t tell you what’s going to happen Saturday. We give you the tools and the data to think it through yourself.
What’s in the platform
We document our methods because we want them scrutinized
Stat headers across the rankings, leaderboard, and betting tables link to glossary entries explaining how each metric is computed, why we chose that definition, and what the limitations are. Our defensive metric conventions are documented — there are three of them, with different sign conventions, and we surface that publicly so users don’t conflate them.
We’ve corrected published bugs in our own analytical layer mid-development. We’ve deprecated a 46-feature spread prediction model because the empirical evidence didn’t support its claims. When our methodology changes, we say so.
Built for serious analytical fans, content creators, and sports markets analysts who want tools that match their thinking — not someone else’s predictions.