At HEAD, is the summed board's margin (gameproj.team_points over lines_for) worse than the close against the RESULT, is its disagreement with the close informative, and does one team-strength scale on TDs close half the gap?

Owner lane: market

Question

At HEAD, is the summed board's margin (gameproj.team_points over lines_for) worse than the close against the RESULT, is its disagreement with the close informative, and does one team-strength scale on TDs close half the gap?

Prediction

P1 A E1 in [+0.6,+1.6] pts/game, interval excl 0; P2 A sd ratio in [0.35,0.65]; P3 A encompassing slope spans 0; P4 B closes >= half of A's gap, E1(B) in [+0.1,+0.7]; P5 B slope spans 0. Frozen in research/scientist/prereg/2026-09-26T0717Z-h289-board-margin-vs-close.md

Finding and verdict

H-289 CONFIRMED ON THE MOVED PREMISE, and the prereg's arm B MISSES its bar. At HEAD (no market anywhere in the line) the summed board is +1.08 [+0.68, +1.47] pts/game worse than the close against the result on all 816 REG games 2023-25 (2023 +0.87, 2024 +1.10, 2025 +1.26 -- fable's +1.03/+1.32 reproduce on the unselected rows), with 53% of the close's dispersion. When the board disagrees with the close, the close is right: the encompassing slope is -0.17 [-0.39, +0.07], so the disagreement carries nothing the close lacks. One teamstate team scale on TDs closes 0.40 [0.09, 0.71] (37%, bar 50%) and leaves the slope at -0.14 [-0.45, +0.19]. Outside the family: teamstate ALONE (scores only) is +0.58 [+0.29, +0.86] to the close and beats the board by 0.50 [-1.02, +0.003] -- the summed lines are no better a spread than two parameters on past scores. fable's falsifier (lose by > 2) is not triggered.

Reasoning

The market lane's question is who is right when we disagree. On 816 games the answer is the close, and the disagreement is not even informative at the margin. The board's spread cannot be a CLV source; any edge on sides has to come from something the board does not yet carry (a team state at minimum), and D-211's blend weight 0.00 for teamtd now holds for the summed board too.

Next action

The team-state question belongs to S-15/kernel, not this lane: file nothing new here. Market-lane next: grade the board's margin against the Wednesday line we could bet (W-012) the same way, since a zero encompassing slope against the close does not yet say the Wednesday line is also efficient against us.