How it works: Chance keeps a power score on every team — rebuilt from every final score of the season, weighted by how big a win was (a 3-point squeaker moves more than a 40-point stomping). Add home field and each team's injury-report load, and the model prints the fair line it believes the game should sit at — published right next to the bookmaker's actual line. Where they disagree, that's a question worth studying. Not a promise.
Free, no account, no wagering — the numbers and the reasoning stay free forever.
warming up the desk…
Full stat sheets, side by side — offense and the points they allow — plus the model's verdict on the duel. The sheets are the star; the number is the garnish.
POWER — the Elo-style rating (higher = stronger season-form).
FAIR LINE — the margin + home field + injury nudge the model believes the game actually is.
BOOK — the real spread a bookmaker has posted right now.
DISSENT — how far the book sits from the model. It measures disagreement, not advantage:
books also charge you a margin to enter (the vig), price in injuries and weather faster than any
free feed, and are usually right. Treat big dissent as "look closer", never as free money.
The streak column is story, not signal. Hot teams cool. The injury nudge is deliberately small — star status matters more than raw counts, and the model can't see that. This desk is fan math for watching games better.
Research and entertainment only — never a wagering service, never betting advice, no tips, no paid picks. If you choose to bet anywhere, that's a private adult decision: know your local law, know the house always charges for the privilege of playing, and never stake rent money. A friend's winning screenshot is a lottery ticket, not a strategy. Prediction markets are a different building with their own tolls — see our plain-English scout report on lookalike apps and fee math. 18+. Model source: public ESPN scoreboards, team schedules, injury reports. Model: AOI Elo v0.2 (margin-weighted, K=20/25).