One column per season. The box covers the middle half of his starts and the thin line inside it is the median. The bold bar is the season average (attempt-weighted, so it matches the official season line). Dots are individual starts, sized by attempts; a hollow dot pinned to the top or bottom edge is a start that falls off the chart (hover it for the value). A tall box means he was inconsistent from game to game. Every number on the chart is adjusted for the opponent’s defense; the raw stats are in the table below and in each start’s tooltip.
What's included
Nix: every start, 2019 Auburn through 2023 Oregon.
Raiola: every Nebraska start in 2024–25, plus 2026 at Oregon from the USC relief appearance on. Portland State (mop-up) is left out.
Manning: his starts only, 2025 season opener on. Pre-2025 relief and mop-up games are left out.
Sellers: starts from 2024 on. 2023 relief snaps are left out, and so is 2025 Vanderbilt (7 attempts before he left the game), to match the 10-attempt rule used for Moore.
Sagapolutele: every start, 2025 and 2026 at Cal.
Stokes: every start, 2026 at Memphis.
Kienholz: 2026 starts at Louisville. His Ohio State years were relief appearances except the 2023 Cotton Bowl, which was his only start that season; one game can’t show a spread, so it’s left out.
Lopez: starts from 2024 on (South Alabama, North Carolina, Wake Forest). His 2023 South Alabama games are left out as a backup season.
Gabriel: every start, 2019–2021 UCF, 2022–23 Oklahoma, 2024 Oregon. 2021 is three games (broken collarbone at Louisville), two once the FCS game drops out.
Atkinson, Jackson and Chiles were pulled straight from CFBD box scores, not Sports-Reference logs. A start is a game where he led his team in pass attempts with 10 or more. Atkinson: 2026 Oregon State only; his 2025 Mercer season was against FCS defenses, which can’t be adjusted to an FBS baseline. Jackson: 2025 and 2026; his 2024 was relief. Chiles: 2024–25 Michigan State and 2026 Northwestern, including 2025 UCLA (he threw one fewer pass than the backup but started); his 2023 Oregon State backup year is left out.
Moore: games with 10+ attempts. That drops 2023 Stanford (5 att) and Arizona (7 att) and all 2024 backup snaps at Oregon.
Passer rating is the NCAA passer efficiency formula: (8.4·yards + 330·TD + 100·completions − 200·INT) / attempts. Source: Sports-Reference career game logs as of Nov 16, 2026.
EPA/play is CollegeFootballData’s expected points added (PPA) per play for the QB, garbage time excluded: his dropbacks (passes and sacks) and his runs. Opponent adjustment: adjusted EPA/play = his EPA/play minus the opponent’s defensive effect, split by play type. The opponent’s effect on dropbacks is its EPA allowed per dropback minus the FBS average; on runs, its EPA allowed per run minus the FBS average. Both are play-weighted totals from its other games that season against FBS offenses, garbage time excluded (CFBD advanced game stats). The two are mixed by his own share of dropbacks and runs that game, with sacks counted as dropbacks. Season averages are weighted by his non-garbage-time plays. This method was checked by an independent review on Nov 17, 2026. Missing: CFBD has no EPA for Gabriel vs Baylor 2022 (a 3-interception game, so his 2022 EPA reads high) or Sagapolutele vs Wake Forest 2026; those columns are marked * in the Starts row. Noise: each opponent’s rating rests on about 8–13 games, so roughly half of any single start’s adjustment is noise; it averages out over a season.
Opponent-adjusted passer rating = the QB’s rating minus how far his opponent’s pass defense sat from the FBS average that season. The opponent’s number is the NCAA passer rating it allowed to FBS offenses in all its other games that season (so the QB’s own game doesn’t count), from CollegeFootballData team box scores. Example: UCLA’s 2026 defense allowed a 95.6 rating against a 134.2 FBS average, so Raiola’s 29.5 against UCLA is credited +38.6 to 68.1. Starts against FCS teams are not adjusted and are left out of the adjusted views. Garbage time is not removed.