Skip to main content
How Our 2017 NFL Predictions Work

A good deal of FiveThirtyEight’s NFL coverage this season will use Elo ratings, a simple system that estimates each team’s skill level using only the final scores and locations of each game. As we’ve done in the past few years, we’re using those Elo ratings to calculate NFL predictions that include win probabilities and point spreads for every game in this season’s NFL schedule, as well as continually updating team rankings.

How do our Elo ratings work? FiveThirtyEight editor-in-chief Nate Silver wrote a detailed FAQ about the formula a few years ago, and almost all of it still applies. Teams gain and lose ground based on the final score of each game and how unexpected the result was in the eyes of the pregame ratings. Under Elo, teams pick up where they left off: The initial team ratings for 2017 are by definition the same as last season’s end-of-year ratings, only more compressed because of regression toward the mean. (Specifically, we regress each team’s rating to the mean by one-third, with the league average team clocking in slightly above 1500.)

This means that going into Week 1, the Atlanta Falcons and New England Patriots are again the NFL’s highest-rated teams, albeit with lower Elo ratings than when they faced off last season in Houston for Super Bowl LI. Why? Like other well-designed predictive rating systems, including ESPN’s Football Power Index, Elo is appropriately cautious early in the season; a team must prove itself to warrant a very high or very low rating. Combine that with the luck inherent in the NFL — the best teams don’t always win — and even the Patriots have just an 18 percent chance of winning the Super Bowl. That forecast is actually quite high by Elo standards; indeed, it’s the highest preseason Super Bowl win probability we’ve had since we began publishing these numbers. But at the same time, 18 percent will seem conservative to the NFL fans who assume the Patriots have a 95.5 percent chance of winning the Super Bowl until Tom Brady retires or Bill Belichick resigns.

It’s important to note that Elo does not reflect injuries or personnel changes, which helps explain why the forecast for the New York Jets, who spent the summer getting rid of all their players with talent, may seem rosy. We currently have the Jets with a 17 percent chance of making the playoffs; that is very optimistic compared with the long-shot odds that a Las Vegas sportsbook is giving the Jets to win the Super Bowl (1,000-to-1!). Don’t get too excited, Jets fans: Their Elo rating will catch up with them in no time if they are as bad as billed.

Neil Paine is a senior sportswriter for FiveThirtyEight.