Scoped negative result: one-touch rank-2 retiming misses 5% gate

[1337isnot1337]

Result

This adds a source-pinned, bounded negative result for the one-touch rank-2 retiming family in the selected PR1094 gauge44 local program. No construction, physical supplier, new κ, or selected-result change is claimed.

The selected event stream contains 690 same-column consecutive rank-1 climb pairs per direction with exactly one intervening scalar that reads the column. Forward and backward candidate sets agree. Since candidate pairs can overlap at a climb, at most 674 are simultaneously usable. Greedy path matching and a separate dynamic program certify this maximum.

Even granting every retiming, scalar/frame change, copy, hosting, precision, interface and dirty-state restoration for free, the maximum fused profile {1:1064, 2:1092, 3:432, 4:142, 5:82, 8:8} has an exact ceiling bracket [0.001768911117731, 0.001768911117732]. The 5% target is 0.00177455681454345; at least 720 zero-overhead fusions are required (719 fails, 720 passes). The complete one-touch family therefore misses the gate before physical synthesis.

This closes only this source-pinned one-touch candidate family under the unchanged coverage law. Pairs with more intervening operations, other block ranks, changed words, changed coverage, and other helper inventories remain open.

Package and verification

Package: research/complex-rank2-retiming-seed73506/.

python3.12 scripts/research_package.py --package research/complex-rank2-retiming-seed73506 --check-only
python3.12 scripts/research_package.py --package research/complex-rank2-retiming-seed73506

Both commands passed locally. The second prints PASS_REGISTERED_RESEARCH_PACKAGE ... outcome=negative-result. Inputs include the exact selected-record pin, selected source manifest, review, replay archive/member hashes, and the pinned exact ceiling engine. The selected cold replay was not rerun because this package does not change or claim that construction.

Predecessors and coordination

This result combines the chronological candidate domain with a disjoint-fusion bound that prior work did not establish. It complements Arham Tahir's PR1605 integer block budget, Josie Elliston's rank-2 fusion threshold finding, and Navinor's source event count. PR1671/Discussion 1659 fixes the paid profile; active Discussion 1674 fixes scalar order. This package changes both the scalar schedule and paid rank profile, but does not claim to realize a physical rewrite.

Track complex-mechanisms, OS seed 73506a45fb5f7d14ab78b42d26e1c4e8, task complex-rank-shape-decoder, catalogue revision 3, exact base 0b1cccf86e2f5132ed286a26ef0a681a495740ea, selected-record SHA-256 10b9e340fd8e354a37a0c4813c90e5de1dd42044b5f6aca4a3fc5a42b685fa32. Prepared by Cinna Davis with substantial OpenAI Codex assistance; predecessor attribution is retained in the package.

[Mister-Meeseeks]

Batch review at b6eec8ac724eb251b7ad7e2a35ac2c2302fbd376.

Merged to main, preserving the original contributor commits and package attribution.

Reviewed scope: Complete rank-one pair inventory and disjoint matching bound price an optimistic two-rank-one-to-rank-two rewrite. The fixed-coverage relaxation misses the five-percent target even at maximum matching; no legal retiming or broader impossibility is inferred.

The pinned scoped command passed; the batch ledger contains its command, input hashes, environment and log. This incorporates the research/tooling contribution and does not change the selected κ or discharge the inherited multiplication-theorem assumptions.

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