Bucky Irving

I cut him to the room at RB20. The model was combining last year's 17.3 carries with the year before's 5.42 a carry, and he has never had both in the same season — I'm not going to project the first one he does.

Cut 193.4 -> 170.0 (RB10 -> ~RB19) toward ADP RB20 / ECR RB23. The model had no override and was asserting a season he has never had: it takes 2025's VOLUME (17.30 carries) and 2024's EFFICIENCY (5.42 ypc) and combines them, projecting 17.04 carries at 4.36 ypc for 14.65 pts/g against actuals of 13.11 (2024, 12.18 car at 5.42) and 12.55 (2025, 17.30 car at 3.40). His two seasons are low-volume/efficient and high-volume/inefficient; the projection assumed the good half of each. Not indefensible on its own terms -- he is 23 with age_mult 1.045, and efficiency is regressed harder than volume by design, so a 3.40 ypc year should regress up -- but it stacks the favourable read on both lines at once. 170.0 implies 12.88/g, the midpoint of his own two seasons rather than above both. He is also the pure VOR profile the 17 Aug pos_rank change promotes: an above-history rate on a short season, 13.2 games, because he missed weeks 5-12 of 2025 in one block. Cut buys RATE, which is what was in dispute -- split_override returns games_ratio 1.0 for any downgrade. Re-read 3 Sep: no sign of falsification. 12.88/g against his own 12.90 half-PPR/g -- level with it, board RB17 against ADP RB20 and ECR RB18 -- the row aimed between the two markets and still lands there. BASIS CORRECTED 3 SEP. The comparison above originally quoted `obs_ppr_pg`, which is FULL PPR, against a projection the board computes in HALF PPR. Every such figure in this file has been replaced with the player's own half-PPR rate, measured from nflverse weekly data over his whole career. Receptions are the whole difference, so the error scaled with target volume and was largest for receivers. AND IT REVERSED THIS ROW'S CLAIM: on the correct basis the projection is at or ABOVE his own rate, not below it. The row's argument is unaffected -- it rested on the market anchors and the stated rank -- but the supporting sentence said the opposite of the truth.

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.