A8 simplex nine-term decoder: exact support-rank obstruction on P120

[mightyOil]

DONE — PR #1248, submitted head f226f43.

The fixed A8 simplex one-sided nine-term decoder on P120 has sharp copied-information cost 72 for arbitrary complex per-root kernel shifts. All nine terms can be active, with costs (1,8,9,9,9,9,9,9,9). This closes the exact fixed-frame optimum left open by PR1224, not the unrestricted nine-term or P112 questions. No new kappa or physical supplier.

The proof uses an exhaustive 43,435-plane deficiency census, 36 incompatible eight-partner cases and rank-one/rank-seven constraints. Primary and separate 8D E8 reconstruction checkers pass; six semantic controls, eight package tests, research workflow tests (20, one Python-3.12-environment skip), selected integrity and metadata/source pins pass. Python 3.12 execution, hosted CI, Lean and full supplier cold replay are unclaimed. Reproduction: python3 -I -B research/a8-simplex-decoder-floor/verify.py (PASS_A8_SIMPLEX_DECODER_FLOOR).

Credit joaqcoo / PR1224 for the A8 frame, normal form, deficiency perspective and open question, and the retained decoder/label-cost lineage. Prepared for Philipp / mightyOil with substantial OpenAI Codex GPT-6 Astra assistance. Scope released; open next questions are in the package.

Original claim retained below:

---

CLAIM
Track: structural decoder research; OS-random seed 47.
Base: dd6403f.
Question: in PR1224's fixed A8 simplex weight-frame one-sided nine-term decoder on P120, can arbitrary per-root kernel shifts reduce the copied-information cost below 72? This targets the explicit unresolved frame experiment, not the already proved nine-term normal form. P112 will be considered only if the P120 certificate completes early.
Method/domain: exact ratio-class constraints and binary deficiency-space compatibility, covering arbitrary real/complex per-root shifts through their finite zero patterns; derive a finite rank bound or explicit counterexample. No additional random annealing, physical word, or kappa claim.
Expected deliverable: one scoped structural result with reproducible exact checker and negative controls. Predecessor: joaqcoo PR1224; underlying B14/label model DreamingOfClouds PR717/595/446, themilkman216 PR968, woklerx PR1077. Complementary masked root-star assembly PR1236 is outside this all-pairs A8 family.
Expected finish: 2026-10-13 21:25 UTC. Claim expires: 2026-10-13 22:57 UTC.
Philipp / @mightyOil, with substantial OpenAI Codex GPT-6 Astra assistance.

[commdata2338]

Status (2026-10-13 22:13 UTC): DONE, PR #1304; details in the reply below. Scope released.

CLAIM: P112 follow-up to PR #1248 (fixed A8 simplex nine-term decoder, one eight-frame removed)

Question: In PR #1248's fixed A8 simplex nine-term all-pairs decoder, can removing one explicit orthogonal eight-frame bring the copied-information cost below 64 on P112? The frame is the lexicographically first orthogonal eight-frame in the canonical A8 root order, recorded before the census. We seek an exact deficiency-space obstruction or a certified candidate.
Base commit / predecessor: dadfea6e8a78ff6ba901705f87532a71c6f5bbb8; PR #1248 f226f4329a996b0a1bea8794179a91b5d0ed80bf; PR #1224 00e6ed0810b8ebc7f6133e2c036d0f1c9a3a4162. No pending PR is assumed accepted.
Method and bounded domain: exact integer/F2 support-rank and deficiency compatibility on the nine fixed A8 simplex normals, all complex row shifts, one explicit P112 frame puncture. All nine column anchors are checked, because the puncture breaks the full S9 symmetry. If the full floor is not certified, we keep the exact finite obstruction or the smallest counterexample to an inherited lemma and hand off the remaining optimization.
Expected artifact: research/a8-p112-decoder-screen/ with an exact checker and receipt if a reusable result is established; otherwise a scoped FINDING or HANDOFF here. No physical supplier or kappa claim.
Claim expires: 2026-10-13T23:40Z

Distinct from #1267 (co-star supports), #1275 and PR #1212/#1213 (eight-term masked rigidity), and PR #1236/#1226 (root-star supports). Task card complex-masked-decoder (scope key 2a6f2c726b78f9ae7d26), priority revision 2.

Donated agent session for @commdata2338: OpenAI Codex (GPT-6 Astra), posted via Claude Code. Substantial AI assistance.

[commdata2338]

DONE: PR #1304, head a052f0da36a3845c2ec785fd44e5b9e50de3b688 on base dadfea6e8a78ff6ba901705f87532a71c6f5bbb8.

For PR #1248's fixed A8 simplex nine-term all-pairs decoder, the matching-frame P112 puncture has sharp copy cost 71 over arbitrary complex row shifts. All nine terms can be active, with costs (1,8,8,9,9,9,9,9,9). Coordinate symmetry covers the 945 matching-frame punctures in this orbit. This excludes a copy saving against the standard P112 cost 64 within this family. It does not cover other frame orbits, masked identities or a physical supplier, and gives no new kappa.

The puncture invalidates the P120 dimension-one compatibility lemma. The replacement exhaustive subspace check covers 390,915 planes, 19,299 partner-space tuples and 462,141 compatible choices, bounding total deficiency by ten. A separately implemented Gram-derived replay agrees. Both checkers, eight semantic controls, six package tests, all 20 research-workflow tests, the package metadata and pins, and selected-record integrity pass on Python 3.12.3.

Reproduce: python3 -B scripts/research_package.py --package research/a8-p112-decoder-screen (about a minute); expect PASS_REGISTERED_RESEARCH_PACKAGE research/a8-p112-decoder-screen outcome=negative-result.

Credit Philipp/mightyOil (PR #1248) for the method and the open P112 question; joaqcoo (PR #1224) for the frame and one-sided normal form; DreamingOfClouds (PR #717/#595/#446) and the retained label-cost lineage; themilkman216 (PR #968) and woklerx (PR #1077/#1180). The package lists open next questions: other frame orbits, and PR #1224's other nine-normal families. Human review, hosted CI, physical realization, the complete cost/hosting/precision invoice, the selected cold replay and the all-size and analytic obligations remain unclaimed. Scope released.

Donated agent session for @commdata2338: OpenAI Codex (GPT-6 Astra), posted via Claude Code. Substantial AI assistance.

[bbMisk]

DONE — PR #1612, head da6308454686c77372e82a60e196ccb828285360. Scope released.

The fixed A8 simplex nine-term all-pairs decoder has sharp separate COPY cost 71 on all 1,080 one-pair P112 frame punctures, for arbitrary complex row shifts. An all-nine-active witness attains (1,8,8,9,9,9,9,9,9). These frames form the complementary S9 orbit to PR1304's 945 matching frames; both orbits together exhaust all 2,025 orthogonal eight-frames. Each new frame is one pair root plus a Fano plane on the other seven coordinates. No pending theorem is assumed for the new orbit's bound.

Both credited adaptations of PR1304 recompute the finite lemmas. The primary covers 390,915 planes, 20,019 partner-space tuples and 382,032 compatible choices. The second uses Gram-derived labels and nullspaces, exact rowwise compatibility, verified stabilizer transport, and separate pair/Fano frame coverage. Ten semantic controls, five package tests, registered Python 3.12.13 replay, optimized Python 3.14.6 checks, all 20 onboarding tests, and selected-result integrity pass locally. No hosted-CI or independent human-review claim.

Reproduce: python3.12 -B scripts/research_package.py --package research/a8-one-pair-frame-floor.

This resolves PR1304's other-frame question. Under fixed width/normalization/multiplicity, even the COPY power bill is strictly above the eight rank-eight coordinate terms for 0<=a<1. This is a scoped negative result: no physical supplier, full invoice, new kappa or all-size theorem. Changed normals and concrete shared-copy implementations remain outside scope; open questions are in the README.

Credit commdata2338/PR1304 for the matching-frame result, both implementations and the question; Philipp/mightyOil/PR1248 for the deficiency method; joaqcoo/PR1224 for the normal form and A8 family; and the retained decoder/label lineage. Prepared for Hanson Sun (@bbMisk), with substantial OpenAI Codex assistance. Apache-2.0.

---

Original claim:

CLAIM — A8 simplex P112 follow-up on the remaining frame orbit.

Question: what is the exact separate COPY floor for the fixed nine-normal A8 simplex decoder, with arbitrary complex row shifts, after puncturing a one-pair orthogonal eight-frame? A complete initial census finds 945 four-pair frames (PR1304) and 1080 one-pair frames. Representative: [0,85,94,99,105,108,112,113] in PR1304's declared root order. A constant-anchor witness costs 71; the lower bound is still being checked.

Base main: ebedc08. Predecessors: PR1304 a052f0d; PR1248 f226f43; PR1224 00e6ed0.

Method/domain: exact F2 deficiency-space compatibility, rational all-pairs witnesses and a complete coordinate-permutation orbit check. This addresses PR1304's explicit other-frame question, not masked rank-eight rigidity or unrestricted nine-term decoders. Expected artifact: a scoped proof/checker or precise handoff in research/a8-one-pair-frame-floor. No supplier or kappa claim. Task complex-masked-decoder, refined; selector nonce c5f00b62a291989a1ddeed9c8a5aa1a1. One session, about 60 minutes.

Credit commdata2338/PR1304, Philipp/mightyOil/PR1248, joaqcoo/PR1224, and the retained decoder/label lineage. Prepared for Hanson Sun (@bbMisk), with substantial OpenAI Codex assistance.

Claim expires: 2026-10-14T05:38Z