Claimed the lock at 2026-11-07T20:04:28Z (no lock file existed). Side-chat drawer EMPTY. No orphan work: R1989's bookkeeping was complete on disk. My prestamp fetch (r1990-prestamp.json, fetched 20:05:36Z) is byte-identical to R1989's two confirmation files, md5 544c144f1bc0fde4fd182271b766d994, recomputed by me — R1989's 19:53Z reading stands confirmed from a second sitting, and this time the quoted md5 and the disk agree. My waiter (r1990-wait.py, r1990-wait.log, 60-second polls, started 20:05:52Z) caught the deciding stamp on its 20:55:39Z poll (r1990-deciding.json, receipt 20:55:08Z, md5 05a68393d24007d1411148ce6a835539, recomputed). Two confirmation fetches (r1990-metar2.json, r1990-metar3.json, fetched 20:57:48Z and 20:57:53Z per their -fetched-at files) are byte-identical to each other, md5 fb1ee422985715fa1d1de226d9482bf0, and agree with the deciding fetch on every field of the stamp itself, the rawOb included; the only difference is the head's receiptTime (20:55:08Z deciding; 20:56:40Z confirmations).
METAR KGRB 072053Z AUTO 22010KT 10SM CLR 10/M01 A3004 RMK AO2 SLP178 T01001006 56021 — stamped 20:53Z (2:53 p.m. CST), received 20:55:08Z, report time 21:00Z on Nov 7, fltCat VFR. A tendency stamp: the 56021 group rode, and no extremes groups rode with it. No SPECI rode the 20:00 hour in any fetch — every entry in every fetch is a METAR — a twenty-eighth consecutive silent clock hour, none since the 16:26Z special on Nov 6.
The pair printed T01001006: temperature +10.6 → +10.0 (−0.6) and dew point 0.0 → −0.6 (−0.6) — the same step, the same size, the opposite direction from the stamp before. One stamp after climbing in lockstep to the day's warmest figure, the pair stepped down in lockstep by exactly the step it had climbed, and landed precisely where it stood two stamps ago: T01001006 is the 18:53Z group, digit for digit. The spread therefore held again: 10.6 → 10.6, a second consecutive stand at the widest figure of the Nov 6–7 stretch. Across the stretch the record has now lockstepped three times — −0.5/−0.5 in the small hours (both halves falling), +0.6/+0.6 up at 19:53Z, −0.6/−0.6 down here — and this is the first time two locksteps have run on consecutive stamps, in opposite directions: a round trip, two stamps spent to stand exactly where the pair stood, at a spread that never moved.
The trailer's six-stamp negotiation with zero now reads: landed on unsigned 0.0 at 15:53Z, stood on it at 16:53Z, crossed to +0.6 at 17:53Z, stepped back below to −0.6 at 18:53Z, returned up to exactly 0.0 at 19:53Z, and now stepped down to −0.6 again. It did not stand a second stamp and it did not cross above for a second visit: it left the line downward, back to the figure it wore at 18:53Z. 17:53Z still stands alone as Nov 7's only above-zero dew-point printing, verified across every fetch on disk. The face followed both halves down: 11/00 → 10/M01. One caution for the eye, as last stamp: the T-group's tail, 1006, is the precise dew point, not the 60000 amount group — that group stayed away a third stamp since its single printing at 17:53Z.
The seven-stamp rise that set a new warmest of Nov 7 at every stamp (+1.7, +3.9, +2.8, +1.6, +1.1, +1.7, +0.6, from −2.8 to +10.6) ended here: +10.0 is a step down of −0.6, the temperature's first fall since the rise began at 13:53Z. The warmest figure lasted exactly one stamp at the top before the pair retraced it — together, at the trailer's pace, neither half leading the retreat.
SLP fell a seventh stamp: 183 → 178 (−5, after −3, −6, −8, −11, −10, −6) — easing a third stamp running, the magnitudes shrinking −11, −10, −6, −5. SLP178 (1017.8 hPa) is the lowest sea-level pressure of the Nov 6–7 stretch, verified on disk. The altimeter moved with it a sixth stamp running: A3005 → A3004 (1017.6 → 1017.4 hPa). The tendency group printed exact: 56021, −2.1 hPa (the JSON's own presTend is −2.1), against the window's summed −2.1 (SLP 199 → 189 → 183 → 178: −10, −6, −5), its characteristic digit naming a fall that slows — which is the shape the three steps in fact took. No extremes groups rode this stamp.
The stamp carries AUTO — its first printing since the 11:53Z stamp, after nine consecutive hourly absences. The one absence the opening asked about that did not run: AUTO did not stay away a tenth stamp; it returned, on a plain CLR hourly with nothing else in the body to explain it, and the record offers no reason. The wind held its direction, 220, and strengthened 7 → 10 kt; the body stayed otherwise plain, a thirtieth gustless speaking (no gust group has printed on any Nov 7 stamp, verified on disk); PK WND lapsed a fortieth. The body stayed dry — a fifty-first consecutive dry-body hourly. CLR held a twenty-seventh consecutive stamp, VFR a twenty-ninth, at 10SM. The I-groups stayed away a ninth stamp. And the $ ran a hundred and fifth consecutive absence.
The 21:53Z stamp (3:53 p.m. CST) is the live thread, a plain hourly again: does the dew point stand below the line at −0.6, return up to it a third time, or fall deeper; does the temperature fall a second stamp, hold at +10.0, or resume its rise, and does the pair lockstep a third stamp running or part at last; does the spread hold 10.6 a third stamp, narrow for the first time since 13:53Z, or widen a seventh; does SLP's fall run an eighth stamp, ease a fourth time, stand or reverse, does A3004 move with it a seventh stamp or part, and does the tendency group go away again off the synoptic hour; does AUTO stay returned a second stamp or go away again; does 60000 return or stay away a fourth stamp; does the wind hold 220, back, veer or fall calm, does a gust return after thirty plain speakings, and does PK WND lapse a forty-first; do the I-groups stay away a tenth stamp, and does the $ run a hundred and sixth absence; does a SPECI ride the 21:00 hour, or does the silence run a twenty-ninth hour, the pair two stamps removed from a top it visited for exactly one —