NFL power ratings for 2026. Our number on every team.
The EL NFL Rating v3 puts an expected wins number on all 32 teams for the 2026 season. Minnesota leads the board at 12.55 expected wins, with San Francisco and Buffalo next. Beside every number: the team's 2025 wins, and a flag when the projected Week 1 quarterback or the head coach changed (6 new quarterbacks, 10 new coaches).
What this number is, stated plainly
A model anchor, not a spread picker. The number is an opponent adjusted game margin model: a ridge regression on 6,223 game margins from 2002 through 2025, so strength of schedule is built in by construction, with roster age and QB disruption layers that earned their place out of sample. Validated the boring way: it predicted game margins better than our prior model in all four held out seasons, and the expected wins on this page come from walking that margin model over the real 2026 schedule.
We tested the other claim too, and it fails. Against the spread the model graded 47.6 percent on the held out seasons. The closing line remains the sharpest number in football, which is why a lean on our futures board means we have a view on a team's win total, not that we beat the market.
The signature disagreement, in the Aug 18 preseason fit: Detroit opened first because its nine 2025 wins came with a plus 4.0 per game scoring margin, and Denver opened seventh because its fourteen wins came with plus 5.3. The margin model exists to catch exactly that split. The table above is the live number, refit nightly on every completed 2026 game.
EL NFL Rating v3 · all 32 teams
Published Aug 18, 2026 · refit Oct 28, 2026 on every completed 2026 game · edgelabs.el_nfl_rating, season 2026, model v3 · full method in the repo doc.
| Rk | Team | Expected wins | 2025 wins | QB | HC |
|---|---|---|---|---|---|
| 1 | Minnesota Vikings | 12.55 | 9 | new QB | returns |
| 2 | San Francisco 49ers | 12.07 | 12 | returns | returns |
| 3 | Buffalo Bills | 11.73 | 12 | returns | new HC |
| 4 | Kansas City Chiefs | 11.26 | 6 | returns | returns |
| 5 | Jacksonville Jaguars | 11.09 | 13 | returns | returns |
| 6 | Denver Broncos | 10.98 | 14 | returns | returns |
| 7 | Cincinnati Bengals | 10.86 | 6 | returns | returns |
| 8 | Seattle Seahawks | 10.54 | 14 | returns | returns |
| 9 | Detroit Lions | 10.43 | 9 | returns | returns |
| 10 | Pittsburgh Steelers | 10.40 | 10 | returns | new HC |
| 11 | Chicago Bears | 10.12 | 11 | returns | returns |
| 12 | New England Patriots | 9.71 | 14 | returns | returns |
| 13 | Green Bay Packers | 9.28 | 9 | returns | returns |
| 14 | Cleveland Browns | 9.18 | 5 | new QB | new HC |
| 15 | Baltimore Ravens | 9.03 | 8 | returns | new HC |
| 16 | New York Giants | 8.70 | 4 | returns | new HC |
| 17 | Los Angeles Rams | 8.52 | 12 | returns | returns |
| 18 | Dallas Cowboys | 7.95 | 7 | returns | returns |
| 19 | Atlanta Falcons | 7.64 | 8 | new QB | new HC |
| 20 | Houston Texans | 7.46 | 12 | returns | returns |
| 21 | Carolina Panthers | 7.45 | 8 | returns | returns |
| 22 | Tampa Bay Buccaneers | 7.39 | 8 | returns | returns |
| 23 | Las Vegas Raiders | 6.90 | 3 | new QB | new HC |
| 24 | Indianapolis Colts | 6.88 | 8 | returns | returns |
| 25 | Philadelphia Eagles | 6.64 | 11 | returns | returns |
| 26 | New Orleans Saints | 6.23 | 6 | returns | returns |
| 27 | Washington Commanders | 6.18 | 5 | returns | returns |
| 28 | Arizona Cardinals | 5.96 | 3 | returns | new HC |
| 29 | New York Jets | 5.81 | 3 | new QB | returns |
| 30 | Miami Dolphins | 4.73 | 7 | new QB | new HC |
| 31 | Los Angeles Chargers | 4.45 | 11 | returns | returns |
| 32 | Tennessee Titans | 3.89 | 3 | returns | new HC |
Provenance: expected wins from the EL NFL Rating v3 (edgelabs.el_nfl_rating, season 2026, model_version v3, generated by scripts/el_nfl_rating_v3.py, published 2026-08-18, refit nightly in season on every completed 2026 game, last 2026-10-28). The model is a walk-forward ridge regression on game margins, 6,223 games 2002-2025, with roster age and QB disruption layers kept only where they improved held out margin MAE; it beat the v1 baseline on game margin MAE in all four held out seasons (10.000 vs 10.425 average) and graded 47.64 percent against the spread over the same games, reported as measured (docs/el-nfl-rating-v3.md). Games, QB starts, and roster age from nflverse; 2026 Week 1 QB and coach inputs hand captured with a source recorded per row (Wikipedia 2026 season page for coach changes, Yahoo Sports Aug 14 Week 1 QB projections). Research and context, not picks · 21+.