De'Von Achane

I moved him down from RB3 and it's a volume call: 13.6 carries a game is the lightest workload in this tier and 4.8 targets doesn't quite cover the difference.

RESTORED 22 Aug at Matt's direction, having been lifted this morning. Lifting it was defensible on its own terms -- the row said 'Moved to RB5 from RB3' and the board reached RB5 natively, so its stated target held without it -- but that reasoning missed what it was doing to the top of the position. WHAT IT BUYS. Native he is RB4 at 243.4 (17.26/g x 14.1g), which sits 3 ranks ahead of ADP RB7 and 1 ahead of ECR RB5, and puts FOUR backs in tier 1 where every outside board runs two. 211.5 puts him RB5 and leaves tier 1 as Gibbs, Bijan and McCaffrey -- three, the MIN_TIER_SIZE floor, and the closest the cut can get to the consensus shape. THE COST, STATED PLAINLY BECAUSE IT IS REAL. 211.5 over 14.1 games implies 15.00/g, which is BELOW ALL THREE of his seasons (16.29, 15.35, 18.08) and 2.05 below his own 2023-25 at the model's RB decay of 0.50, which comes to 17.05. The model's native 17.26 is only 0.21 above that weighted history, so the model is well calibrated on him and this override is not correcting a model error. WHAT IT IS INSTEAD: a claim that his 2025 was a peak -- 18.08/g on a career-high 14.88 carries a game -- rather than a new level, and that Miami will not hand it back. Both markets agree with the destination even if not the rate: ECR has him RB5 exactly and ADP RB7. There is no way to reach RB5 at a rate his record supports; RB5 requires under 15.11/g, so the tension is structural rather than a choice about how to word it. Recorded so the next reader knows this is a view about regression, held against a calibrated projection, and not a bug being patched.

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.