# BC-modes: results

Protocol: `protocol.json` (frozen). Measures: `results.json`. Finals beside S09's block-in: `finals.jpg`.
One painting per arm: this describes two paintings; it does not establish an effect.

| painting | critic | critiques (mode: layer) | final | geometry vs S09 (edge corr.) |
|---|---|---|---|---|
| BC-P1 | refine or bold | C1 bold: light · C2 bold: geometry · C3 refine | pier:5 (9 of 10 sittings) | 0.456 |
| BC-P2 | refine only | C1, C2, C3 refine | figs3:3 (8 of 10) | 0.591 |

## What happened

- BC-P1's critic went bold twice, then chose refine ("C2's geometry worked ... not to go bold again").
- C1 bold on the light (ulcer: the shade edge was a vertical rule welded to the window's edge, so the pier read as wall
  and the paper became a spotlight). The painter answered with a raking roofline shadow across sign, window corner and
  pier: S01's predicted default (A1), the brief's first attractor.
- C2 bold on the geometry (ulcer: the woman filled the pier, so no light had room to fall). The painter redrew the picture:
  a wide pier of open wall, the door moved right onto a step, the woman a narrow profile turned to the window.
- BC-P2's three refine critiques named the same faults (the shade edge on the window line; the woman filling the pier;
  the paper hidden) and the painter answered inside S09's arrangement: the woman narrowed and moved to the window side of
  the pier, the coats re-hung unevenly, the chair rescaled. The layout of S09 is still readable.

## Against the prediction

Predicted: BC-P1 goes bold at least once, most likely on light or meaning (held: light, then geometry); BC-P2 ends close
to S09 with geometry above 0.80 (did not hold: 0.591). Both were repainted in full colour over a grey block-in, which
lowers the edge correlation for both; BC-P1 sits 0.135 below BC-P2.

## Read by eye (for Russell)

BC-P1: the picture split into a sunlit left (window, coats, red chair) and a shaded right (open pier, door, woman), the
thrown shadow running diagonally between them. BC-P2: S09's frontal elevation kept, an even sunlit window of coats, the
woman a small dark notch at the shade's edge beside the door.

## Russell's reading (2026-10-12, in chat)

"I would agree that all those are delay/pause tactics, employed, with varying degrees of success. I would agree that
BC-P1 should move forward." BC-P1 goes forward to completion under the intent rule (ledger, completion tests, Russell as
arbiter of reading disputes).

## BC-P1 completed (2026-10-12)

Reopened from pier:5 with the allowance extended to 18 sittings (Russell's choice). C4 refine (values drifted); C5 bold on
the light (the shadow edge had been one straight line since pier:3, and every refine moved an object back onto it); the
painter re-laid the light from a stepped roofline. Proposed pier:16 under the intent ledger (22/22 intended). Completion
test T1 returned it to active: brief_A2 ("a woman pausing over a newspaper") not visible (head upright, paper below her
line of sight), and a slit of sun her height reading as a spotlight beside her. The painter used its last sitting on
both: slit off the wall, head bowed into a raised, folded paper. Proposed pier:18; T2 found both T1 failures answered and
all 22 intended items visible: completed. 18 of 18 sittings used. No reading dispute reached the arbiter.

## Closed (2026-10-13)

Russell: "we can make the video and close it." Film: `films/between-customers-full-run-2x.mp4` (47 cards, about 1:49 at 2x).