Who should I start this week?

Dynasty still has a bench. Pick two to four players. Last year’s matchup and boom/bust sit next to this week’s Engine line once the season is under way — no “start him” stamp.

Week 3 typical miss 4.2 half-PPR · 46% in the published band · 32% boomed over ceiling · 22% busted under floor (a fair band is about 60% in) (350 skill players). Full accuracy table

Week 4 · 2026 · v1.28-weekly-v1.6 · snapshot Sep 30, 8:01 PM

FAQ

Is this a start/sit ranking?
No. It compares this week’s Engine projection for the players you pick. Dynasty still has a bench, so the number is a lineup helper — you decide who starts.
Where do the numbers come from?
NextDyne Engine weekly projections: season rate adjusted for matchup, rest, Vegas, venue, injury, and recent usage. Raw stats stay in the store; we score them with the format you choose.
Why is the board empty in August?
Weekly Engine numbers cover the regular season, and preseason games don’t count toward them. You can still pick names and see last year’s matchup and boom/bust; the weekly projection shows up in Week 1.
What is the 2025 block on each card?
Last regular season, not this week. Opponent-allowed is how that defense scored against pass or rush in 2025. Script is that offense’s 2025 plays and pass rate. Boom is a 20+ half-PPR week; bust is under 5. The Engine number above it is this week, when we have one.
What do PPR, TE Premium, and Superflex change?
They change how the same raw weekly line is scored. Superflex uses the PPR template (it’s a roster format). If you are signed in with a league, you can score with that league’s rules instead.
How accurate is the Engine?
After official boxes land we grade last week’s pre-kick line on Engine accuracy — mean miss and how often the actual landed in the floor–ceiling band. That is not the confidence chip on a roster.
Do I need an account?
No. Logged-out compare uses Half-PPR unless you pick another template. A league login only upgrades the scoring to your rulebook.