Quinshon Judkins

A play on volume on a stinky offense, and that's fine. 18.1 carries a game is the most of any back I have. He'll be inefficient doing it, and at ADP RB22 you aren't paying for efficiency anyway.

CARRY CLAIM, 22 Sep. The board gives him 18.44 carries a game and he has run 12 and 12 -- a 54% overshoot on the one quantity that is actually stable this early. Found because Matt challenged a $400 FAAB bid I had built on the unadjusted number, in the same hour I re-priced Travis Etienne for exactly this defect. I had already flagged the volume gap in the same analysis and quoted the board reservation anyway. THE TARGETS ARE FINE AND THAT IS WHY THIS IS A CARRY CLAIM, not a whole-player cut: projected 2.93 tgt/g against 2 and 5 actual. Only the rushing volume is wrong. SIZED OFF TOUCHDOWN REGRESSION, NOT OFF HIS POINTS. He has produced 6.0 and 7.3 for 6.65/g with ZERO touchdowns, and pricing a back off two scoreless games is the error this file warns about everywhere -- TD variance dominates a two-game window while volume does not. At roughly 0.4 TD/g his own usage is worth about 9.05/g. Written at 9.5/g x 13.8 games = 131.1, which lands his carries at 13.63 -- DELIBERATELY ABOVE the 12 he has run twice, because two games is a short window, Dylan Sampson is on IR, and a rookie lead back's role more often grows than shrinks. This is a cut from 18.44, not a claim that 12 is his ceiling. GAMES BLANK: he carries no designation and 13.8 is a rookie availability prior, not injury damping, so split_override spends this on rate where the argument lives. Target blanked -- a volume claim promises no rank. NOT SIZED ON THE TEAM SUM. Cleveland's board RB carries total 37.66 against a 26.34 rush budget, which looks damning and is the documented artifact of summing every rostered back including men who never dress together. The evidence here is the PLAYER-level fact alone: 18.44 projected against 12 and 12 run. WHAT IT ACTUALLY LANDED, recorded rather than the intent: 9.50/g x 13.8 = 131.1, RB15 -> RB31, carries 18.44 -> 13.63 as designed. But `split_override` scales EVERY rate proportionally, so the TARGETS fell 2.93 -> 2.17 as well -- against 2 and 5 actual. The reason above argues the targets were fine and the mechanism cut them anyway, because the file has no way to state 'carries down, targets flat'. The season total is the claim and it is right; the target line is now about 1.3/g light of his own usage, which is the price of a one-number override and is stated here rather than discovered later.

This is a view, not a correction — the model said something different and was overruled. See how it works.

Pace extrapolates the selected weeks over 17 games. It is what that stretch was worth, not a projection — no availability, matchup or regression adjustment is applied.