New to these reports? Start here

Dotted-underlined words have a plain-English definition — hover or tap them. Every term is also on the glossary page.
"Null" means we found nothing, not that something broke. Most reports here are negative results, on purpose — knowing an idea doesn't work is the point.
Two questions get asked separately. First, is the effect real? Second, is it already priced into the betting odds? An effect can be completely real and still useless to bet on.
A "calibration" row is a self-check. It runs the same method on something already known to be true. If that fails, the whole report is unreliable — so it's reported alongside the findings.
If a confidence interval includes zero, the real effect might be nothing at all, so no claim gets made.

What drives the big gaps? (exploration, weeks 1–8; not part of the frozen test)

Written after PREREG.md was committed. It changes nothing in the test.

	Lines	Model takes Over	DK Q1 line ÷ full-game median	QB's 1st–2nd DK Q1 line	Model's side won
Gap ≥ 6 pts	69	29%	0.217 (often 0.23–0.24 or 0.17–0.18)	30%	62%
Gap < 4 pts	139	45%	0.205	33%	43%

Big gaps happen when DK sets a QB's Q1 line at an unusual share of his full-game line. The model uses one league-wide Q1 share (median ~20%); DK apparently adjusts per QB.
Darnold, Goff, Herbert and Hurts were priced at 23–24% of their full-game line (model: Under).
Watson, Daniels, Brissett and Shough were at 17–18% (model: Over).
Not a QB-change effect. New-to-the-board QBs are as common among big gaps (30%) as small ones (33%).
Both directions have won: Overs 14–6, Unders 29–20.
A plausible mechanism: DK's per-QB first-quarter adjustments may be fitting noise. Our player-splits work found individual splits barely repeat year to year (median r ≈ +0.01). If a QB's Q1 share is mostly luck, a line built on it overreacts, and the league-wide ratio is closer. That would also explain why DK's line doesn't drift toward the model before kickoff: it's a deliberate pricing choice, not a stale line.
Testable later, outside the frozen test: if this is the mechanism, the edge should be largest for QBs whose recent Q1 shares were extreme, and DK's per-QB shares shouldn't predict actual Q1 shares beyond the league ratio.

          backtests/nfl_q1_passing_shape/EXPLORATION_2026_11_03.md
        

updated 2026-11-03 20:50