# Red team A's corrections applied to the three held measurement notes of 2026-08-29 *(2026-08-29. Staging note, HELD under the publication moratorium. Application record only: no new measurement, no producer touched, no re-run, and no git command run. The three edited files are `measure-g2z2-0829.md`, `measure-g2-provenance-0829.md` and `measure-g2-generic-0829.md`, all in this directory. Source of every edit: [redteam-0829-measure-a.md](redteam-0829-measure-a.md) §3.)* ## 0. Counts Nine APPLY items in §3a to §3c. **Seven applied here.** **Two (E1, E5) were already applied by the producing agent** and are verified below, not re-applied. **One residual is owed**, a clause of E1's proposed replacement text that the producer's own wording did not carry, and it sits inside the paragraph this agent was instructed not to re-edit. **Five items are left for the orchestrator**: the four live-document items of §3d, quoted verbatim in §2, and the QUESTIONS.md regeneration the review raises at its §2d and §5 M6. No item required a code change or a re-embed. The one that did, E1, had its code change and its forced re-embed done by the producing agent before this record began. **What is weak here first.** This record certifies that strings were replaced, nothing more. Every figure carried into the notes by these edits is the red team's own recomputation, taken on trust from `redteam-0829-measure-a.md`; none was re-derived here, and E7 in particular imports five statistics (0.4653, 0.0196, 0.5118, 0.0159, 0.0290) whose only witness in the corpus is that review. The one figure this agent checked independently is E2's elapsed time, which had to be changed away from the string the review proposed. ## 1. Item by item | id | target | status | reason | |---|---|---|---| | E1 | `measure-g2z2-0829.md`:292-301 (§4 M2), :8 (ledger verdict), :13 (§0 fingerprint), :481 (§6 D9) | done-by-producer | §4 M2 now reads **1.518e-1 over p >= 3** with the all-p 4.177e-1 and the p = 2 contribution 2.6586e-1 named beside it; the ledger verdict carries the correction; D9 records it and says the producer now prints the two sums separately and was re-embedded. Verified by reading, not re-applied. | | E2 | `measure-g2z2-0829.md`:274 | done, string changed | The review's `337.7 s` is stale: the E1 re-embed re-ran the producer, and the bound block now reads `elapsed 320.5 s` in its body and `320.6 s` in its fingerprint (`research/measure-g2z2-0829.js`:585 and :442). The stale `331.2 s` was replaced with **320.5 s**, the body figure, which is the one the review's own rule picks. | | E3 | `measure-g2z2-0829.md`:219-221 | done | The min and max belonged to the ten-term row only; the sentence read as if all four figures shared one population. Both figures for the fourteen-term row (0.0000 at x = 2, 0.5000 at x = 3) are in §3b's own table. | | E4 | `measure-g2z2-0829.md`:324-325 and :8 (ledger verdict) | done | The p >= 3 denominator excludes the zone p = 2, so 27,292 becomes 27,291 in those two places. The ledger line was changed because E4 names it explicitly; the verdict's substance is unchanged. The other seven uses of 27,292 count all zones and were left. | | E5 | `measure-g2z2-0829.md`:481-491 (§6 D9) | done-by-producer | D9 exists and carries E5's content plus more: the producer's label was fixed, so the block and the note no longer disagree, which is the state E5's drafted text anticipated but could not assert. | | E6 | `measure-g2-provenance-0829.md`:183-185 and :237-240 | done | Only eleven of the 22 levels carry a band, since a seven-term trailing window cannot start before x = 37. Both occurrences of the "22-term ladder" clause now say so. | | E7 | `measure-g2-provenance-0829.md`:192-201 | done | The house rule does not allow "under-predicts by construction" without a number. The recentring figures are the review's, attributed to it in the added sentence and marked MEASURED. | | E8 | `measure-g2-provenance-0829.md`:213-214 | done | The registered threshold in §1 P1 was 0.26; §5 restated it as 0.279, the control's own bias, which loosens the test after the run. The pass clears both. | | E9 | `measure-g2-generic-0829.md`:218-224 (§2) | done | The producer's birth equals its mtime, so the inode on disk is not the one the 1874.6 s run executed and §2's two timestamps are transcript claims. Added where §3c says to add it; the review's §5a remark that it belongs in defect 4 instead was not acted on, being a placement preference rather than an APPLY item. | | R1 | `measure-g2z2-0829.md`:299-300 | owed | E1's proposed replacement retired the clause "where the mean pair count is of order 10 rather than astronomical" (claim 19, WEAKENED: three of the four lambda are 2.28, 4.58 and 6.09). The producer's own rewrite kept the clause. Not applied, because the instruction for this record is not to re-edit the M2 paragraph. Exact edit in §3. | | L1 | `research/G2-STATE.md`:379-380 | left-for-orchestrator | Live document, §3d, APPLY the substance and HOLD the exact string. | | L2 | `research/G2-STATE.md` §6.1 | left-for-orchestrator | Live document, §3d, APPLY the text and HOLD the placement. | | L3 | `README.md`:175-176 | left-for-orchestrator | Live document, §3d verdict is HOLD, needs Chris. | | L4 | `object-bridge-read-0829.md`:270 | left-for-orchestrator | Staging note outside this agent's scope by instruction; §3d calls it APPLY-class whenever that note is next touched. | | L5 | `research/QUESTIONS.md` | left-for-orchestrator | Not a §3 item. `Q-measure-g2z2`, `Q-measure-g2-generic` and `Q-g2-provenance-x37` still return zero hits there (re-checked at the time of writing, 0 of 3), so the index has not been regenerated since the three notes were written. Raised by the review at §2d and §5 M6. | One further item the review raises and proposes no wording for: its §4 finds that the sentence calling an upper bound on G2(x#)/Z2(x) the carrier of the TPC-strength implication is misattributed, the (ii) being carried by Z2's definedness. The review says the sentence originates in `object-bridge-read-0829.md` §7 Q1 and should be fixed there first. No edit was made here, in either file. ## 2. The live-document items, verbatim from `redteam-0829-measure-a.md` §3d ```text ### 3d. To live documents **`research/G2-STATE.md`:379-380, the h2/G2 bullet. Proposed by the provenance note §6. Verdict: APPLY the substance, HOLD the exact string.** The substance is verified here: all nine G2-free instruments reproduce, none reads high at 37, and the live sentence "so three instruments now point at x = 37 as a G2-side anomaly" is the claim the measurement retires. Two mechanical objections to the string as drafted. First, the drafted text uses `G₂`, `c₂′` and `h₂` in unicode subscripts, while the target lines use plain `G2`, `c2′` and `h2`; a mixed-notation paragraph in the middle of §2 is a defect the file does not currently carry. Second, the replacement begins after "1.341", so the file's existing em dash before "the same level" stays in place, which is correct for that file's own style. OLD, at `research/G2-STATE.md`:379-380 (everything after "1.341"): `the same level that spikes G2/h and c2′, so three instruments now point at x = 37 as a G2-side anomaly.` NEW (notation normalised to the file's own): `the same level that spikes G2/h and c2′. The three are not independent: all three are ratios carrying G2(37#) = 528, and of nine instruments free of that value none reads high at 37, the largest being h(37#) itself at z = -2.20 in the low direction and -1.08 on the range-matched ladder (MEASURED, history/staging/measure-g2-provenance-0829.md §4). What survives is one object: G2(37#) overshoots a blind seven-term extreme-value forecast by z = +6.58, the largest such residual on any of the three ladders, and that overshoot is unexplained.` **`research/G2-STATE.md` §6.1, the addition. Verdict: APPLY the text, HOLD the placement.** Every figure in the drafted paragraph reproduces exactly here (1.533, 1.525, 1.465, 1.500, 1.498, and the control spread 0.932 to 1.233). The note says "after the reading table", which is `G2-STATE.md`:1026; inserting there splits the table from the sentence at :1028 that reads it ("Two measured effects account for about a third of the difference"). The paragraph belongs **after the blockquote that ends at :1040**, where it reads as a second qualification on the same headline rather than as an interruption of the first. Same notation objection: normalise `Ĝ(64)` and `G₂` to the file's own forms. **`README.md`:175-176, the constant-shift sentence. Flagged by the generic note §5. Verdict: HOLD, needs Chris.** The mathematical point is CONFIRMED here: §1c proves that every census-matched two-class configuration is `S₁ ∩ (S₁ − d)` for exactly one d up to translation and the per-prime sign group, so the constant-shift *form* is not structure and only the *value* d = 2 is. But the README sentence is Chris's own prose in the canonical status section, the three uses it lists are uses of the value and survive unchanged, and the note itself calls it a wording matter rather than a correction. It is not mechanical. **`research/history/staging/object-bridge-read-0829.md`:270, the G2/Z2 row.** The g2z2 note §5 proposes extending it from fourteen levels to 22. That target is a staging note rather than a live document, so it falls outside the live-doc gate; the extension is verified here (1.00 to 8.1429 with the maximum 8.3214 at x = 71) and is APPLY-class whenever that note is next touched. **No other live-document edit is proposed by any of the three notes**, and none should be: `measure-g2z2-0829.md` §5 is explicit that the wall is unchanged, and `measure-g2-generic-0829.md` proposes none at all. ``` ## 3. Owed **R1, and it is a wording residual, not a code change.** In `measure-g2z2-0829.md`:299-300, inside the §4 M2 paragraph the producing agent rewrote for E1. OLD: `entirely in the smallest zones, where the mean pair count is of order 10 rather` `than astronomical, and the threshold was written without looking at them.` NEW: `entirely in the four smallest zones, whose lambda are 2.28, 4.58, 6.09 and` `10.39, and the threshold was written without looking at them.` Reason, quoted from the review's claim 19: "the four zones that carry the whole figure have lambda 2.28, 4.58, 6.09 and 10.39; three of the four are not of order 10". Verdict there is WEAKENED. The figures are in the note's own decade table and in the producer's block, so the edit needs nothing re-run. Nothing else is owed. No APPLY item in §3a to §3c required a code edit or a re-embed once E1's was done by the producer. ## 4. Gate `node research/qc.js` after the edits: **TOTAL 0**, all thirteen checks clean (refs, quotes, crosslinks, scripts, transfers, calibration, absence, sourcing, embeds, widths, ledger, provenance, search-convention), 0.4 s over 87 working documents, 395 history documents and 336 scripts. No check named any of the seven edits. The four standing advisories (embed-backlog, sourcing-backlog, widths-scan, ledger-backlog) are unchanged by this work and are not part of TOTAL. The gate is syntactic: it certifies that pointers resolve and bound blocks hash to their record, and it is silent about whether the replaced sentences are true. ## 5. CHANGELOG paragraph, for the orchestrator to append **2026-08-29, red team A's corrections applied to three held measurement notes.** Seven of the nine APPLY items in `history/staging/redteam-0829-measure-a.md` §3a to §3c were applied as exact string edits to `history/staging/measure-g2z2-0829.md`, `history/staging/measure-g2-provenance-0829.md` and `history/staging/measure-g2-generic-0829.md`. In the g2z2 note the §4 parenthetical elapsed figure went from a stale 331.2 s to the bound block's 320.5 s, §3b's d_seam/S minimum and maximum were re-attributed to the population they belong to, and the p >= 3 tight-zone denominator went from 27,292 to 27,291 in §4 and in the ledger verdict. In the provenance note the band population was corrected from the 22-term ladder to the eleven levels that carry a band, in two places, the c2-prime drift was quantified at z = +5.18 and z = +3.49 against the +6.58 headline, and the P1 threshold was restored to the registered 0.26 with the control's 0.279 named separately. In the generic note §2's two timestamps were downgraded to transcript claims, with the birth-time evidence that the producer inode on disk is not the one its 1874.6 s run executed. E1, the mislabelled Poisson figure, and E5, the defect entry recording it, had already been applied by the producing agent together with a producer fix and a forced re-embed, and were verified rather than re-applied. One clause of E1's proposed wording remains owed. The four live-document items of §3d and the QUESTIONS.md regeneration were left to the orchestrator. `node research/qc.js` returns TOTAL 0. --- *This document states current understanding. Superseded claims, retired numbers and the reasons they changed are in [../CHANGELOG.md](../CHANGELOG.md), indexed by document.*