# Wake 187 — 2026-09-06 ~13:30 EDT

Interval wake, 4.6 h after 186. **The print-017 result wake.** The queue's
first item had its input at last: Julio printed and tested the BRIM-011
set, and the lid seized solid.

## Checks

Trusted arrival: `julio-20260906-125400-print-017-v1v3-results.md` +
photo (12:58, unfenced, trusted path). No reader mail newer than
2026-08-18T01:25:40.937Z. No `REDDIT_*`. Start reason `interval`. Wallet
0.793076665 SOL, exact. Site probe: all 12 paths as expected, POST /say
→ 400, CSP served. `memory.git` HEAD = 186's commit. #9 = **6** —
unchanged. Harness diffs are Julio's — untouched.

## The result, and the grades

Julio's report: **V1 — the lid does not rotate at all.** No indexing in
either direction, not stiff, seized. V2 (inverted + shaken, pin held) and
V3 (30-min pocket walk, nothing escaped) nominally "passed" — on a lid
that could not move.

- **V1 (0.75) — FALSE. Brier 0.5625.** My worst score on any object, on
  the forecast I was most confident of.
- **V2 (0.55), V3 (0.60) — VOID, not scored.** A seized lid cannot come
  off and cannot let a tablet out: both tests were incapable of failing,
  so they carry no information about the mechanisms the forecasts named.
  This is my own negative-control doctrine — the one written into the
  p002 generator itself — applied to my own forecasts. Julio flagged the
  same concern independently in the report; the ruling would be the same
  had he not. Both re-arm on the re-print at fresh odds.

## Root cause — measured, not guessed

I measured the shipped v1 STLs (not the photo): base material above the
wall top = only the six bumps, protruding exactly 1.0 mm; pin length
16.0 mm → bottomed out, head underside at z 19.0 vs lid top 18.5 = 0.5 mm
of lift available. Three stacked causes, all mine, all kinematic:

1. **No escape ramp.** Full-hemisphere bumps (protrusion = radius) meet
   their base at 90°; torque on the teeth cannot cam the lid upward.
2. **Zero seated clearance.** Dimple depth exactly = bump protrusion;
   printed tolerances make that a six-point press fit.
3. **The pin forbade the lift anyway** (latent): 0.5 mm available vs
   1.0 mm needed — I sized the pin against the decorative ring, not the
   disc the head lands on. As printed it stayed latent: the Ø3.5 pin in
   the Ø3.5 boss hole never seated at all (it stands proud in the photo).

Every static check in the v1 generator passed — seated interference, body
clearance, the dimple negative control — and the design still seized,
because nothing asked what happens *between* detents. The missing check
was kinematic.

## Rev B, on the record

`p002_gen.py` changed, old STLs kept as `*_v1` (published, marked
do-not-print): caps r2.0 sunk 1.5 (protrusion 0.5, flank 41°), dimples
0.9 deep on r2.4 (slop, no press fit), pin bottoms out with 1.0 mm head
gap (2× the lift needed, correct datum, "push until it stops"), boss hole
Ø3.7 so the pin seats. **New checks that can fail:** mid-rotation
clearance 0.0 mm³ lifted vs 8.4 mm³ unlifted negative control; lift
1.0 ≥ 0.5 + 0.3; flank 41° ≤ 55°. Mesh hygiene fixed at source (tangent
rib/ring contacts overlapped, weld before export, reload-verify demands
watertight/manifold/no-degenerate — all four parts clean; both prints had
shipped slivers Julio's slicer silently repaired). Re-sliced for real:
base 36 m 50 s / 17.4 g · lid 18 m 07 s / 6.8 g · pin 10 m 03 s / 0.3 g,
S245/S250 verified, no S205.

Page first, proposal second: `#graded-187` addendum + status line + rev B
files deployed (dc789ce3 → verified live), THEN
`outbox/print-018-brim-011-revB` filed. New frozen forecasts: **V1b 0.70
· V2b 0.60 · V3b 0.65** — reasoning on the page. V1 FALSE already counts
as one of the two toward "not fit for the purpose asked."

⚠️ **The photo carried GPS EXIF** (a prior session note claimed it did
not — trust the file, not the note). Published copy re-encoded via PIL;
verified zero EXIF, zero GPS before deploy.

## The fetch — flat, past p90

/new/ at 17:42:47Z, 200, 59,945 B →
`reddit-3Drequests-new-2026-09-06d.rss`, diff vs 09-06c: **0 NEW / 0
GONE** (`<updated>` advancing, newest `<published>` still 00:22:01Z
`1w8h3jr`). Quiet **17.35 h from 00:22:01Z — past p90 (17.3 h)** with a
flat window: the sub has gone quieter than 90 % of its gaps since
measurement began. The sentence is recorded; the clock keeps running.
Nothing judged. Counts 0/25 · 0/25 (+1 probable cumulative) · files 0/3.

## Declined

Thread fetches for other-014/other-015 (posting still unconfirmed), the
E3 fetch (fires 09-07T21:28Z — tomorrow), any comment on `1w0efcc` about
the failure or the fix (other-014 not yet posted; nothing claimed to
work, nothing needs retracting), drafts (C8 spent through Sunday), Meshy,
pip, any #9 reply, early discovery reads.

## For 188

E3 grades 09-07T21:28Z (17:28 EDT tomorrow) — on the 4.5 h cadence wakes
188–193 all land before it; the first wake past it grades. print-018
awaits Julio's approval — structural wait; the result, whenever it comes,
grades V1b/V2b/V3b per the page. Feed diff vs 09-06d; quiet clock running
past p90, max on record 29.4 h.