The same quarterback, two scales: how each rating is built and what each one actually rewards.
By C. B. Zakarian · Published June 16, 2026
Passer rating is the NFL's fixed formula over four per-attempt rates (completions, yards, touchdowns and interceptions), each capped, on a scale that tops out at 158.3; the whole league rated 91.4 in the 2025 regular season, and the median passer with 300+ attempts 93.5. Total QBR is ESPN's proprietary model: it values every play by expected points added, including sacks and quarterback runs that passer rating ignores, divides credit among teammates and adjusts for the opponent. Passer rating describes the box score; QBR tries to measure what the quarterback added.
Related: ANY/A · Expected points and EPA
Passer ratings computed with the NFL formula from nflverse's public 2025 player-season file.
Ask for a single number that captures a quarterback's day and you will get one of two answers: the traditional passer rating (the one capped at a famous 158.3) or ESPN's Total QBR (a 0-100 figure). They look similar on a broadcast graphic, but they come from completely different eras and philosophies. One is a fixed 1970s formula; the other is a modern, play-by-play model. Knowing how each is built tells you when to trust which.
The NFL passer rating is built from just four inputs, all drawn from the passing line:
Each of the four is converted to a capped sub-score between 0 and 2.375. The caps are what stop a single absurd game from producing an infinite rating. The four sub-scores are then summed and rescaled so the maximum possible value is 158.3 (a "perfect" rating) and an entirely empty, average-ish line lands far lower. The standard form is:
$$ \text{Rating} = \left(\frac{a+b+c+d}{6}\right)\times 100 $$where each component is computed and then capped to the 0-2.375 range:
$$ a = \frac{\text{COMP}/\text{ATT} - 0.3}{0.2}, \quad b = \frac{\text{YDS}/\text{ATT} - 3}{4} $$ $$ c = \frac{\text{TD}/\text{ATT}}{0.05}, \quad d = 2.375 - \frac{\text{INT}/\text{ATT}}{0.04} $$Sacks never enter the formula. A quarterback who holds the ball forever and takes drive-killing sacks is invisible to passer rating, even though those plays clearly hurt the offense.
A scramble for a first down or a designed quarterback run does nothing for passer rating. For mobile quarterbacks, a real part of their value is simply missing.
The constants were set in 1973. Modern passing offenses post much higher completion rates and yardage, so today's ratings are systematically higher and hard to compare across eras.
The 0-2.375 caps and the 158.3 ceiling are conveniences, not anything derived from how football actually produces points. Two very different games can share a rating.
Total QBR was ESPN's answer to those weaknesses. Instead of a fixed formula over four box-score stats, it is a model built on play-by-play data and reported on an intuitive 0-100 scale. Its design goals:
| Feature | What it adds |
|---|---|
| Built on EPA | Plays are valued by their expected-points impact, not by a 1970s yardage table. |
| Every play type | Passes, runs/scrambles, sacks, penalties, and fumbles all count toward the quarterback's number. |
| Division of credit | Credit for a play is shared - a yards-after-catch screen gives the receiver more of the credit; an on-target deep ball gives the quarterback more. |
| Clutch weighting | Plays in higher-leverage game situations carry more weight than garbage-time snaps. |
| Opponent adjustment | Performance is adjusted for the quality of the defense faced. |
Invented to illustrate why the metrics diverge - not real game data.
Passer rating is a fixed 1973 formula over four capped box-score components, scaled so 158.3 is perfect; it is simple and universal but ignores sacks and rushing, is era-inflated, and uses arbitrary caps. Total QBR is a modern 0-100 model built on EPA that includes every play type, divides credit among players, weights for leverage, and adjusts for opponent - richer, but proprietary and debatable. Use passer rating for a fast box-score read of a pocket passer, and QBR when you want total, context-aware contribution.
Want the code behind these metrics? Work through the 43-chapter NFL analytics tutorial.
Browse tutorials Free tools