Squawka Signal is our in-house NFL model. It gives every team a probability — to make the playoffs, win its conference and win the Super Bowl — and every game a win probability and a projected margin. The numbers here are the model’s own read, updated as games are played. Here is how it works, in plain terms.
Team power ratings
At the core is a single strength rating for every team, built with the Elo system — the same method chess uses and FiveThirtyEight popularised for the NFL. Win and you take rating points off the team you beat; win by more, or beat a stronger side, and you take more. The ratings run on every NFL game since 2021 and update after each week. Home advantage is built in, and at the start of each season every team is regressed part of the way back towards the league average — last year’s form only tells you so much about this year’s roster.
The quarterback adjustment
The quarterback is the biggest single lever in an NFL game, and a pure team rating misses it: a strong roster can start a backup, a middling one can have an elite passer. So on top of the team rating we score every starting quarterback’s play from the box score — passing and rushing production, touchdowns, interceptions, sacks — into one value figure, and track a rolling measure of each quarterback’s level. A team is then rated on who is actually under centre, not on last year’s depth chart.
Two things keep that honest. When a quarterback changes team in the off-season, last season’s snaps would credit him to the wrong side, so we correct it from the live roster: where last year’s starter has moved on, the model promotes the best-established quarterback who remains. And a quarterback’s rolling value is regressed towards his own career level between seasons, so a hot or cold finish — often on a team going nowhere — doesn’t carry undamped into the following year before a ball is kicked.
Simulating the season
From those ratings we simulate the rest of the season tens of thousands of times — every remaining game, the playoff bracket, the byes and reseeding, and a neutral-site Super Bowl. Count how often each team reaches each stage and you have its playoff, conference and Super Bowl probabilities. It all refreshes as results land, so the numbers move with wins, injuries and quarterback changes.
What we publish
For every game the model gives a win probability and a projected margin; for the season it gives each team’s playoff, conference and Super Bowl odds. These are the model’s numbers — a data-led read, not a bookmaker’s price. Where you want to bet, the sportsbook links on our pages price the same markets.
What it does not do
A word on what the model is not. It is not a tip service, and it does not claim to know more than the market — on plenty of teams it will simply agree with the price. Its edge, where it has one, comes from rating quarterbacks cleanly and from simulating the whole season rather than reacting to a single result. Treat it as one informed input, not a guarantee.