Grades · 2025
Quarterback grades
Three components — opponent-adjusted EPA, success rate, and how often the throw came from the first read. The weights are fitted against next-season performance every time the grade is built, never chosen by hand.
Out of sample the composite predicts next season at r = +0.442, against +0.426 for its best single component. It correlates 0.853 with raw EPA — close enough to be measuring the same thing, far enough to be measuring it differently. If that ever passes 0.95 the grade should be retired, because at that point it is EPA with extra steps.
| # | Quarterback | Tm | Grade | DB | EPA/DB | Success | 1st read | vs EPA |
|---|---|---|---|---|---|---|---|---|
| 1 | D.Maye | NE | 75.4 | 540 | +0.308 | 54.4% | 53.9% | |
| 2 | M.Stafford | LA | 74.2 | 617 | +0.245 | 53.2% | 55.6% | |
| 3 | D.Jones | IND | 73.4 | 409 | +0.158 | 49.6% | 63.7% | +3 |
| 4 | M.Jones | SF | 72.9 | 304 | +0.148 | 52.3% | 58.3% | +4 |
| 5 | S.Darnold | SEA | 71.3 | 502 | +0.149 | 51.8% | 58.9% | +2 |
| 6 | J.Love | GB | 62.5 | 461 | +0.240 | 49.7% | 51.7% | -3 |
| 7 | D.Prescott | DAL | 62.5 | 631 | +0.179 | 47.9% | 56.5% | -3 |
| 8 | J.Goff | DET | 62.0 | 614 | +0.174 | 48.4% | 55.4% | -3 |
| 9 | C.Stroud | HOU | 58.0 | 446 | +0.107 | 45.5% | 58.8% | +2 |
| 10 | T.Lawrence | JAX | 55.4 | 600 | +0.064 | 46.2% | 57.2% | +5 |
| 11 | P.Mahomes | KC | 52.0 | 537 | +0.127 | 47.9% | 47.9% | -1 |
| 12 | J.Hurts | PHI | 49.7 | 486 | +0.071 | 43.2% | 56.8% | +2 |
| 13 | T.Shough | NO | 46.2 | 356 | +0.007 | 43.8% | 56.5% | +8 |
| 14 | J.Allen | BUF | 45.8 | 506 | +0.141 | 46.6% | 45.2% | -5 |
| 15 | J.Herbert | LAC | 45.2 | 566 | +0.029 | 45.2% | 50.9% | +4 |
| 16 | B.Mayfield | TB | 42.4 | 584 | +0.001 | 43.1% | 52.0% | +6 |
| 17 | T.Tagovailoa | MIA | 42.3 | 415 | +0.034 | 44.8% | 48.9% | +1 |
| 18 | J.Brissett | ARI | 39.8 | 530 | -0.007 | 44.7% | 48.0% | +5 |
| 19 | C.Williams | CHI | 39.5 | 594 | +0.071 | 43.3% | 47.6% | -6 |
| 20 | B.Young | CAR | 39.0 | 506 | -0.043 | 42.7% | 52.9% | +4 |
| 21 | A.Rodgers | PIT | 38.3 | 526 | +0.054 | 43.2% | 47.1% | -4 |
| 22 | L.Jackson | BAL | 38.0 | 338 | +0.055 | 45.6% | 42.9% | -6 |
| 23 | B.Nix | DEN | 38.0 | 638 | +0.098 | 44.2% | 44.3% | -11 |
| 24 | J.Flacco | CLE | 37.0 | 432 | -0.095 | 41.0% | 56.6% | +1 |
| 25 | J.Dart | NYG | 36.7 | 375 | +0.010 | 41.6% | 49.7% | -5 |
| 26 | G.Smith | LV | 34.6 | 503 | -0.146 | 42.1% | 51.7% | |
| 27 | C.Ward | TEN | 17.9 | 592 | -0.185 | 36.8% | 46.6% |
Where the grade and raw EPA disagree most, the grade was right on Trevor Lawrence 2023 and Geno Smith 2022, and wrong on Mahomes 2023 and Flacco 2024. Roughly even. Two components that seemed obvious were tested and dropped: sack blame is almost pure noise year to year (stability r = +0.013), and a six-component version scored worse than success rate on its own.
FTN charting begins in 2022. The quarterback grade rests on four seasons and 74 quarterback-seasons across three transitions — thin, and the honest reason to treat all of it as directional rather than settled.