The model turns power ratings into fair value in three steps: rate every team, simulate the paths to a national title, and reconcile the resulting probability against de-vigged market prices. The output is a single consensus fair value per team, expressed as a percentage, that can be compared directly against the cheapest contract on the board. For 2026, that process puts Texas on top at 9.1% fair value, with a best price of 11c on Kalshi.
How does the model convert a power rating into a probability?
A power rating is a strength estimate, not a probability. On its own it says one team is better than another; it does not say how often that edge survives a full season and a bracket. The model bridges that gap by running each team's schedule and the title path repeatedly, letting stronger ratings win more often without ever winning every time.
The result is a raw win probability for each contender. Texas at 9.1%, Notre Dame at 8.6% and Oregon at 7.8% show how compressed the top of the board is: small rating gaps translate into fractions of a percentage point once uncertainty is layered in. That compression is the point. A title requires surviving variance, and the model prices that variance rather than rewarding the highest rating outright.
Why de-vig market prices into fair value?
Raw market prices overstate probability because every venue bakes in a margin. Summed across all teams, listed prices add up to more than 100%. The model strips that overround out, normalizing the field so the implied probabilities sum to a coherent whole. That cleaned number is the consensus fair value.
Blending the model's simulated probability with the de-vigged market keeps the estimate honest in both directions. Where the simulation and the market agree, confidence is high. Where they diverge, the gap is the signal worth examining. Ohio State and Indiana both land at 7% fair value, a case where distinct profiles converge on the same number once vig is removed.
What does the fair value board look like for 2026?
The national title board separates into tiers. Texas, Notre Dame and Oregon form the lead group between 7.8% and 9.1%. Ohio State and Indiana sit at 7%, with Miami at 5.8%, Georgia at 5.3% and LSU at 4.9% filling the next band. Below that, Texas A&M at 2.9% and a cluster of Texas Tech, Oklahoma and Ole Miss at 2.1% round out the listed field.
Reading the board this way makes relative value obvious. The difference between Texas and Georgia is not a coin flip; it is roughly four points of fair value, and the prices should reflect that spacing.
Fair value versus best price across the board
Fair value only matters next to what the contract costs. Each team's best price, quoted in cents, is the cheapest available entry across venues. When that price sits above fair value, the market is charging a premium over the model's estimate; when it sits at or below, the contract is closer to fairly priced.
On the current board, the best prices live on Kalshi, from 11c on Texas down to 3c on the longshots. The chart below lines up fair value against best price for the lead group, where the gaps are tight enough that venue selection and entry price do real work.
How to read the model without overtrusting it
The model is a lens, not a verdict. Power ratings shift as rosters and results change, de-vigging assumes the market is roughly efficient, and a single venue quoting the best price means liquidity can be thin. Fair value is an estimate with error bars, and both the model and the market can be wrong at the same time.
Used carefully, the framework still does the core job: it converts ratings into comparable probabilities and flags where price diverges from estimate. Venues such as Kalshi, Polymarket and ProphetX (promo code VAULT) quote these contracts continuously, so the board can be re-checked as ratings and prices move. None of this is financial advice; it is a structured way to see where the model and the market disagree.
