Skip to content

Prove the sharp global exact-sum Gaussian radial optimum (no kappa) - #7696

Open
aliromman wants to merge 1 commit into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-036
Open

aliromman wants to merge 1 commit into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-036

Conversation

@aliromman

@aliromman aliromman commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

Result and scope

The sharp global forward radial slack is 1/sqrt(2) for injective exact-sum Gaussian-pair rounding with Euclidean error at most one. A new Hall family proves the lower bound. An explicit bijection attains the supremum and has the same inverse outward and absolute radial suprema. Each individual input has a strictly smaller radial change. The lower bound assumes no phase covariance, period, square identity or surjectivity.

Read the proof, exact receipt, direct map and verifier and scope record. This resolves the global question after the finite lower bound sqrt(13)-3. The construction pairs adjacent vertices within each half-grid square ring. It gives an explicit negative provider, exact sums, common-translation covariance and wire-swap covariance. Phase covariance and square-to-swap fail and have retained counterexamples.

No new kappa, native supplier, complete multiplier invoice or selected-assumption discharge is claimed. All five selected assumptions remain. A physical caller must accept the actual inverse and missing symmetries; range, scratch, tape costs, tensor behavior and native tariff remain open. This completes the advance claim.

Evidence and reproduction

Normal and optimized CPython 3.12.16 replay passed on macOS arm64. A separate directed walk checks 129 complete rings and 66564 half-grid points against the direct map. Direct four-coordinate calculations check both providers, exact Gaussian error, both inverse identities and radial inequalities on all 14641 inputs in [-5,5]^4, plus 80 signed inputs through 4097 bits. The two explicit sharpness sequences are checked at 69 indices.

Thirty-two complete Hall squares check the corner and exterior-edge exclusions. Separate rational-threshold neighbor censuses give deficits 8 versus 6 at slack 1/2, 18 versus 16 at 3/5, and 2178 versus 2176 at 7/10. A larger symbolic square checks slack 707/1000. Invalid domains, zero radial slack and the excluded phase/square identities are controls. All decisions use integer arithmetic. The written proof establishes the all-input result; finite checks corroborate it.

Metadata, committed bytes, source history, modes, eight-file export and archive integrity pass. A read-only optimized replay without Git or a usable PATH passes and preserves all nine payload/manifest hashes and modes.

With CPython 3.12 and its standard library, from the repository root:

python3.12 -I -W error -B research/gaussian-exact-sum-global-optimum/verify.py
python3.12 -I -W error -B -O research/gaussian-exact-sum-global-optimum/verify.py
python3.12 -B scripts/research_package.py --package research/gaussian-exact-sum-global-optimum
python3.12 -B scripts/check_research_commit.py --package research/gaussian-exact-sum-global-optimum --check-source-history --check-modes

Expected markers:

PASS_GAUSSIAN_EXACT_SUM_GLOBAL_OPTIMUM
PASS_REGISTERED_RESEARCH_PACKAGE
PASS_RESEARCH_COMMIT_BYTES

Hosted CI did not certify this package. The plan job failed in the shared test_current_ci_section_matches_selected_record: the base guide names frontier6030 while the selected record requires make frontier7000-verify. The log reports 760 tests, one failure and 24 skipped; package replay was skipped. This package changes no shared CI file. Local replay remains the scientific evidence. Full predecessor scientific replays, selected cold replay, Lean and Docker were not run. Independent human and formal-kernel review are unclaimed.

Sources and credit

Prepared by @aliromman with substantial OpenAI Codex assistance. Ali Romman's PR #7673, pinned proof, supplied the eighteen-to-sixteen finite obstruction and global question. Nick Petalas (@petalas) supplied the first finite witness and forward-only exact-sum question in PR #7652, pinned source. Both finite witnesses are prior work and are the first positive members of the new family.

Jerry Chen (@CubeJerry) supplied the half-grid transform and annular geometry in PR #7570, pinned proof. That stronger square-to-swap class has the different sharp constant sqrt(19)-sqrt(13). Its paired-ring construction and connector edges are distinct from this individual-ring matching.

Ali Romman's PR #6509, pinned proof, supplied the ideal pair and exact-sum contract. Its physical port evidence is not transferred here. Credit details preserve the source roles. No source code was copied; new work is Apache-2.0.

Base: canonical main e54523517d0f03e58352e647c403a5af3e525a57, priority revision 8. Card uniform-transfer-audit, scope changed to this local precision contract. Fresh handoffs, open PRs, Discussions and source-specific intake show no competing global exact-sum claim. Closed overlap uses the complete 6175-record snapshot and an updated-order delta with no later changes. Sparse local and unpublished work remain coverage limits.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant