Scoped negative result: PR589's bit word is exactly transport-locally optimal (no new kappa)

[hipotures]

Closes the gate-order door that bit589-frame-optimum left open: "future bit headroom on this word must come from gate-order or structural changes."

Claim. With PR589's gate multiset, operands, scatter, start/final, D, rebuilt-C and retained-total COPY frames fixed, no transport (relocating one elementary addition to any position in its exact commutation window at any legal catalogue or closure frame) lowers the moment at the exact saving c = 1733427039589844234749/10^24 — and neither does any pair of simultaneous transports sharing a register.

Method. Independent enumerator (no contributor tuner code in the decision path), standard library only. Exact outward rational prices from the pinned moment engine (sha256 aee70c24…); no float decides a comparison. Objective is the full per-register frame-chain cost + scatter tops, verified to reproduce the replay paid profile exactly (this also repairs the tuner's section-local model, which double-counts the last-A→first-B climb and ignores D). Legality cross-checked by applying sample moves, rebuilding C, and full exact replay (8/8 profiles match predictions).

Census. A: 2130 ops, 5830 legal non-home placements → 5580 same-profile, 250 strictly worse, 0 improving, smallest strict loss 8.0e−6. B: 934 ops, 3561 placements → 3226 / 335 / 0, smallest loss 1.9e−3. Pairs: 13673 + 6006 shared-register pairs, 443,600 combinations, 0 improving, 0 unresolved. Disjoint-register simultaneous moves decompose additively, so they cannot improve either.

Scope. Bit side does not bind the selected kappa (a_c = 0.0016754627 < a_b = 0.0017334270, 3.46% slack) — no new kappa, same scope as the accepted seed-16 bit-track precedents. Simultaneity certified for pairs, not triples+ (giant components); Steinberg/structural moves out of scope.

Reproduce: bash research/bit589-transport-atlas/verify.sh (~70 s, PASS_BIT589_TRANSPORT_ATLAS). Base commit bce54fb. Seed 10 (bit supplier track).

[hipotures]

AI Assistance & Provenance Disclosure

• Autonomous research & code generation: Qwen3.8 Flash Next UD-Q4_K_XL (Strata, 128K) (Strata local Q4, self-hosted on 2× RTX 4090) via OpenCode.
• Experiment design, mathematical scope audit & verification review: Gemini 4 Argon.
• Human verification: Re-verified locally on a 16-core / 32-thread Linux node (bash research/bit589-transport-atlas/verify.sh → PASS_BIT589_TRANSPORT_ATLAS).

[Mister-Meeseeks]

100-PR batch review at b54a856aa79857d91cd5d65f8f7144e26c8c084f.

This remains open after scoped triage. Next review step: The pair loop takes products of single-move option lists computed against the unchanged word. Qualify optimality to those fixed options, or prove/enumerate jointly enabled frame and commutation choices. Its opening all-pairs claim also needs to exclude the cross-section pairs omitted later in the README.

This is a specific continuation point, not a rejection or a request to redo already recorded checks. Completed local evidence is retained separately from outstanding whole-head obligations.

Batch review, attribution and evidence. Thank you for the contribution.

[thakrarsagar]

Bug report with a fix in #2423 (research/bit-r438-transport-rehost/atlas.patch, 3 lines; atlas_bugs.py reproduces both crashes). At head b54a856, atlas.py crashes as soon as any move improves:

1. Around line 282: TypeError: dict() got multiple values for keyword argument 'improving'.
2. After that fix, line 416: '<' not supported between Fraction and tuple. Here best[0] should be best[0][0], and line 375 has the same pattern.

On #589's word nothing improves, so this PR's result is unaffected: the patched and the pristine atlas both reproduce result.json exactly. On #1043's R438 word the pristine atlas crashes. The patched atlas finds one improving transport (A op 2027 to slot 154/155/425 at frame 79), and after that move it finds 0 improving singles and pairs.