The pocket holds. The ceiling doesn't.
What the consensus says
The dynasty room has sung one note all spring: rushing upside is the whole ballgame at quarterback. PFF's superflex top-200 leans on six of the nine top-three fantasy QB finishes from 2022 through 2024 coming from passers who ran for at least 500 yards. RotoBaller calls the legs the thing that "separates" the elite tier, and CBS says superflex rewards young dual-threats because they print floor every week. Pocket passers get the back of the hand. Here is the wrinkle: my own engine seemed to disagree, loudly. Its outcome audit, which scores every old ranking against what players actually did over the next three seasons, flagged the pocket_passer archetype as the single most under-rated bucket on the board. Fifty-seven ranks of it. That is the engine accusing itself of fading the Bradys too hard. So I went looking for the bug.
The claim, in plain English
If the engine is really burying pocket passers it should pay for, the tell would show up in the year-over-year tape: at the same level of production, the no-legs guys should hold their value at least as well as the runners. Turn that into a predicate. A pocket passer is the engine's own cutoff, 150 rushing yards a year or fewer:
def is_pocket_passer(qb_season):
return qb_season.rushing_yards / qb_season.games <= 9.375 # <= 150 rush yds / yr
Then I control for the only thing that matters: how good they already were. Band every starter by Year-0 fantasy points per game, so a 17-ppg pocket passer is only ever compared to a 17-ppg runner. The band is the whole game. Without it you are just rediscovering that good QBs are good.
How I beat on it
Standard Year-0 to Year-plus-1 design. Every QB season from 2015 through 2024 with at least six games and 150 pass attempts, paired to the same player's next year, scored under a 10-team superflex half-PPR system. That is 271 quarterback year-pairs, 119 of them pocket passers, split by Year-0 ppg. The full two-gate framework lives here.
What the data actually said
First swing, the aggregate: pocket passers averaged 15.84 ppg the next year, the rest of the field 17.51. A 1.67-point gap, and it points the wrong way for my under-rating theory. Because aggregates lie, I banded it.
| Year-0 ppg | Pocket next-yr | Field next-yr | Gap |
|---|---|---|---|
| elite 22+ | 18.33 (n=3) | 20.79 (n=19) | -2.46 |
| 19-22 | 17.45 (n=18) | 19.18 (n=30) | -1.72 |
| 16-19 | 16.87 (n=38) | 17.36 (n=59) | -0.49 |
| 13-16 | 15.07 (n=41) | 15.81 (n=24) | -0.74 |
| under 13 | 13.48 (n=19) | 14.36 (n=20) | -0.88 |
Every band, the runners come out ahead. Not by a mountain in the middle tiers, but the sign never flips. And the durability stat I expected to rescue the pocket guys buries them instead: the runners repeat as a top-12 fantasy QB the next season 52.6 percent of the time. Pocket passers, 33.6. The honest twist is that pocket passers are not fragile. Their year-over-year retention is 1.03, a hair better than the field's 1.02. They keep exactly what they earn. What they earn is just a lower ceiling, and a lower ceiling means fewer of them hold a starting job once the rushing QBs reset the bar every September. Stable. Just stable lower.
What the engine already figured out
This is the rare one where the consensus and my engine were already holding hands, and I walked in late. The engine does not over-credit pocket passers. It gives the dual-threat archetype its rushing floor and the young-QB ceiling bonus, and hands the pocket passer the flattest decline curve on the board precisely because that retention number is real. The legs build the ceiling, the arm protects the floor, and the engine prices both. So what was the 57-rank flag screaming about? Something other than forward points. That benchmark scores QBs against realized value across all four positions in a scarce 1-QB-replacement world, and it says the engine's cross-positional board placement buries durable starters like Goff, Cousins, and Purdy too far beneath the running backs. Not that it should project them to out-score the runners. Two different bugs wearing the same jersey, and the forward-production fix the headline number begged for would have been flat wrong.
What to do about it
Buy the legs. At any production level the rushing QB is the better hold, and the durability edge you think you are buying with a pocket passer is a mirage: steady, not safe-and-ascending. The pocket passer's real job is the cheap, boring bye-week stabilizer, not the centerpiece. As for the engine, I am not touching the QB projection math, because the tape says it is right. What I am filing instead is the question the audit actually surfaced: whether the cross-positional scarcity weighting drops durable pocket starters too far under the running back tier in 1-QB formats. That is a board-placement calibration, not a forward-points one, and it gets its own test. The number flagged a problem. It just wasn't the problem it looked like.