Data — XoverO
data.xovero.ai
Almost everything being asserted about college football spending and roster value is untestable with public data. This is where the evidence lives.
STATUS — in build.
What's established
- Roster turnover broke in 2021, not 2019. It was never the portal. It was the money attached to it.
- Group of Five turnover held near 36% for a decade and now sits at 49%.
- Spending does not predict production. Nine seasons, 131 FBS programs, null throughout.
- Close-game records are near-random.
What this is not
- Not a valuation product. The public data won't carry one. That's the finding.
- Not leaked NIL figures. Public filings and public play-by-play only.
- Not player-level compensation. No individual, no program identified.
- Not licensed data. Nothing behind a commercial-use restriction is republished here.
The tier-level breakdown of the turnover finding publishes on October 22.
Sources
- EADA
- Federal Equity in Athletics filings.
- CFBD
- Play-level and season-level efficiency, FBS and FCS.
- Open records
- State public-records requests.
Method goes up alongside each result so the number can be argued with.