Fair value is the end of a four-step pipeline, not a guess. Power ratings feed a season simulation, the simulation outputs a raw title probability, and that probability is calibrated and de-vigged across venues into a single consensus number. On the current national-title board, that process puts Georgia on top at 14.7%, with Miami at 14.4% and Ohio State and Notre Dame level at 13%.
Each stage removes a different kind of noise: ratings strip out record-flattering schedules, the simulation accounts for bracket paths, and the de-vig removes the house margin baked into quoted prices. What survives is a probability that can be compared directly to the cheapest contract on the board.
What is a power rating, and why start there?
A power rating estimates team strength on a single scale, independent of win-loss record. Two teams at 8-1 can carry very different ratings if one earned it against ranked opponents and the other against a soft slate. The rating is the input that keeps the rest of the pipeline honest.
Ratings move on performance and roster signal, not on narrative. That matters because the market often overreacts to a marquee result and underreacts to quiet, repeatable dominance. Starting from ratings rather than records is what lets the model disagree with price in the first place.
How does the model turn ratings into a title probability?
Ratings alone do not win titles; paths do. The model plays the remaining schedule and the projected bracket thousands of times, letting rating gaps set each game's win probability and letting upsets compound the way they do in a real postseason.
The share of simulations in which a team lifts the trophy is its raw title probability. A strong rating with a brutal path can land below a slightly weaker team with a cleaner route, which is why the board does not simply mirror the rating order. The raw number is then calibrated against historical outcomes so the percentages mean what they say.
Where does fair value come from, and why de-vig?
Quoted contract prices are not probabilities. Across a full market, the quotes sum to more than 100% because every venue bakes in a margin. Reading a single 15c contract as a flat 15% chance double-counts that margin.
The model de-vigs each venue so its implied probabilities sum to 100%, then blends the venues into one consensus fair value. The result is comparable across Kalshi and Polymarket, which is what makes a true cheapest-venue read possible. Codes like Polymarket's TGSWC or Kalshi's FADE change the cost of entry, not the underlying fair value.
What does the current title board show?
The top of the board is tight. Georgia at 14.7% and Miami at 14.4% are inside a point of each other, with Ohio State and Notre Dame knotted at 13%. Indiana opens the second tier at 7.3%, trailed by LSU at 5.9%, Texas at 5.8% and Alabama at 4.6%.
The compression at the top reflects a field without a clear favorite: four teams priced within roughly two points on probability. That is the kind of board where small rating moves and single results swing the order week to week.
Where does the model disagree with the market?
Once fair value is set, the comparison to best price is mechanical. Georgia's cheapest contract is 14c on Polymarket against a 14.7% fair value, so price trails the model. Indiana shows the same shape, 7c on Polymarket against 7.3%. Those are the board's two clearest discounts.
Miami runs the other way: 15c on Kalshi against 14.4% fair value means the market is paying slightly above the model. Ohio State and Notre Dame both print 13c against 13% fair value, as close to flat as the board gets. Prices and the model can both be wrong, and none of this is financial advice, but the gaps are where the model is pointing.
