[unpaidmoon]

Change and mathematical scope

Negative result. No new κ. The selected conditional exponent stays 781451/250000000 (0.003125804) in certificates/selected-result.json.

On the emitted local word of PR #5708, every A2 conjugation with one, two, or three noncommuting contacts either fails to emit on the published route table or does not strictly lower the paid query moment for any exponent in (0, 1). All 765 candidates preserve the scalar endpoint. The split is 605 missing routes, 153 losses of shape 2H1−H2, one loss of shape H1+H2+H4, and 6 paid-neutral words. Every length-1 candidate is missing a route edge. The saving shape H1−2H2+H3 already used by #5708 does not occur again.

This is a local word certificate. It does not rebuild the 18-call caller, the adjoint, or ordinary admission.

Task and overlap

Base 2485805a6d47f36ea83ddde6327d453ac82d44dc, priority revision 8. Card bit-helper-contract, scope refined. The card asks for one uncovered cut on the #5679/#5708 chain and says to stop when the paid moment does not improve. The #5708 README leaves these 765 conjugations as local-only discovery. This package replays that count.

Session r59, item 2 of the ordered eligible list. The 19 October handoffs (discussion comments 19030045, 19030047, 19030049, and 19030053) are already delivered or closed. No newer finite handoff index appears in discussion #1166 from 2026-10-26 through 2026-10-30. The revision-8 "Next bounded decisions" table then orders the open rows. Item 2 is the BIT review of the #5679/#5708 chain. The port-0 G20 tariff already delivered by this account is a different clause and was dropped.

Overlap refresh at 2026-10-30T13:28:54Z covered 333 open PR bodies and 6,037 discussion records. The phrase "one-to-three-contact" appears in the #5708 README, its vendored copy under research/g20-four-role-resynthesis/source/, and discussion 5696. None of those is a replayed census.

Validation

Python 3.12, standard library only:

python3.12 -I -B research/g20-short-conjugation-screen/verify.py
python3.12 -I -O -B research/g20-short-conjugation-screen/verify.py
python3.12 -B scripts/research_package.py --package research/g20-short-conjugation-screen

Both direct runs and the registered replay print PASS_G20_SHORT_CONJUGATION_SCREEN candidates=765 missing_edge=605 emitted=160 loss_2h1_minus_h2=153 loss_h1_h2_h4=1 neutral=6 savings=0 k1_emitted=0 endpoint_failures=0. Metadata replay prints PASS_REGISTERED_RESEARCH_PACKAGE with outcome=negative-result.

Controls: the known one-step saving shape is still classified as a saving; deleting a route used by the base word blocks emission; changing one ADD source changes the base result.

Not run: the #5708 integration replay, the 18-call caller, adjoint columns, moduli 101 and 10201, ordinary admission, Lean, and Docker.

Credit

The emitted word, route table, A2 relation, and the local-only 765 count are Ali Romman's research/g20-conjugation-admission, following Dugongue, ai-nanachi, and the retained G20 lineage named there. No predecessor file is modified.

Prepared by unpaidmoon with substantial Cursor Grok assistance. Apache-2.0.

Open next questions

1. Do the six neutral one-step neighbors and the 645 common-frame expansions in the same Raise G20 ordinary BIT saving to0.003507719059749 with scalar conjugations #5708 sentence have a replayed certificate? The first step is to write those two rules from a retained search record. The blocker is that the published package does not contain the search.
2. Can a conjugation with four or more contacts, or one correction moved off its contact frame, strictly lower the paid moment on (0, 1)? The first step is one explicit four-contact word, or one contact moved to a frame the route table already contains.