J.K. Dobbins or Rhamondre Stevenson: Who Should You Start?

The call

• The Week 1 usage is the tell: eight carries, team-high 36 yards, and no targets. A back with 0% target share in PPR needs volume or a score, and 43% snaps says he isn't getting the volume.
• The matchup is a negative, not a boost — Jacksonville grades 9 of 32 in the DvP composite (1 = stiffest), so the soft-spot argument isn't there even on a one-game sample.
• Game script is the one thing working for him: a 23.0 implied total as a 2.5-point favorite keeps the early-down rushing script alive, which is exactly where he gets his work.
• Tipping point: if Jacksonville plays it close and this becomes a grind, RB37 at 11.6 projected is a usable flex. If the model's expectation is a game that gets away like Kansas City did, he's off the field late — bench him for any back who catches passes.

The call

• Usage is the whole case: 85% of snaps plus a 19% target share is bell-cow work, and that pass-game role is worth points every week in PPR regardless of how the run lanes look.
• Game script cooperates — 4.5-point favorite with a 24.0 implied total means positive-script carries late and red-zone chances rather than chasing from behind.
• Matchup is neutral-to-favorable at 21 of 32 in DvP composite, though the 1-game sample carries 0.00 confidence, so it's the role and not the opponent driving this start.
• Only real risk is efficiency — 13.2 projected and an RB26 consensus rank say this is a floor play, not a league-winner, but that floor beats a typical replacement.