Skip to content

Prove the sharp all-rank binary dual-polar cut ceiling (no κ) - #7551

Open
petalas wants to merge 1 commit into
CrocSwap:mainfrom
petalas:research/w4-1113d-stopping-obstruction
Open

petalas wants to merge 1 commit into
CrocSwap:mainfrom
petalas:research/w4-1113d-stopping-obstruction

Conversation

@petalas

@petalas petalas commented Nov 13, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Scoped written proof with exact finite corroboration, produced by Nick Petalas. The exact Boolean-cut ratio on the binary symplectic dual-polar graph is

max_S p_m(S)/p_1(S) = 2 - 2^(1-m), for every integer m >= 1.

Here p_i is cut disagreement on a uniform distance-i pair. Point cuts attain the bound. The first nonconstant adjacency eigenspace uniquely maximizes the spectral quotient. This resolves the missing all-rank maximization in jacobalansussman's PR #4628, Theorem 6 of author-evidence/induction-across-block-sizes.md. Its finite evidence through rank 12, point-cut witness and coordinate-family interpretation retain predecessor credit. No predecessor scientific executable was replayed.

The new proof factors the distance-eigenvalue polynomial and compares every sector by an elementary uniform inequality. It closes the ceiling for constant-dimension coordinate certificates and their nonnegative sums. This is not a physical call-count theorem or a new kappa. Only six files under research/binary-dual-polar-cut-ceiling/ are added.

Evidence

  • Before: the cited induction note leaves the all-rank comparison open after finite ranks through 12. After: PROOF.md proves it for every positive rank, independently of finite checks.
  • The new checker compares 1,224 factored polynomials with 40,424 recurrence equations through rank 48 and checks all 8,256 nonconstant spectral quotients through rank 128.
  • A separate literal binary-span oracle constructs the 3, 15 and 135 Lagrangians at ranks one through three, checks every pair's intersection numbers, enumerates all 6 and 32,766 nontrivial rank-one/two cuts, and checks 81 sharp point cuts. Fifteen arithmetic/domain controls pass. These are corroboration of the new proof, not a replay of predecessor certificates.
  • Local CPython 3.12.15, standard library only: warning-strict normal and optimized direct replays produce identical generated hash-bound receipts. Their measured command durations are 0.293 and 0.294 seconds; these are deadline checks, not speedup claims. Both registered checks pass, with a complete registered invocation taking 0.892 seconds.
  • Committed-byte/mode/source-history preflight and official six-file export integrity pass at 9aec48e7507d4fb6e52170f40505b72e6e9e41f8. Both modes also pass in a read-only Git-free export with empty PATH. All seven exported files, including the export receipt, retain their bytes and modes; direct and exported replay receipts agree. In-memory syntax/function-length and staged whitespace checks pass.

From the repository root, the exact reproduction commands are:

/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -W error -B research/binary-dual-polar-cut-ceiling/verify.py
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -O -W error -B research/binary-dual-polar-cut-ceiling/verify.py
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -B scripts/research_package.py --package research/binary-dual-polar-cut-ceiling
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -B scripts/check_research_commit.py --package research/binary-dual-polar-cut-ceiling --check-source-history --check-modes

Any CPython 3.12 executable may replace that local path. Expected scientific marker: PASS_BINARY_DUAL_POLAR_CUT_CEILING. The verifier generates its receipts and writes no files. Each registered command has a 120-second limit.

Scope and credit

The complex-rank-shape-decoder card explicitly changes to this certificate limitation. Base is canonical main e54523517d0f03e58352e647c403a5af3e525a57, fetched again before committing. The catalogue's #5599 assessment is historical; selected #7000 and all five framework assumptions remain unchanged. The source induction note, selected record and original card are pinned as historical provenance, without runtime scientific dependencies. Advance CLAIM.

Credit jacobalansussman for the source question, finite evidence, point-cut count and coordinate-certificate application, with eumemic and the kept-subspace/native lineage credited there. Dual-polar spectra and spectral Dirichlet-form comparison are classical; Lee and Tanaka, SIGMA 14 (2018), 009 give standard background. This package derives the binary formulas directly. No global literature-priority claim is made.

This result is restricted to the stated binary graph and certificate model. It does not classify all equality cuts, extend to other fields, exclude arbitrary noncoordinate certificates, or supply a physical word. Native realization, source/target access, ownership/restoration, stock, compiler, precision/tape, analytic recovery and the complete supplier invoice remain unproved here. Selected cold replay, Lean, Docker, independent human and formal proof-kernel review are unrun or unclaimed.

Open next questions

Which Boolean cuts attain equality at rank at least three? Their indicators must lie in the constant and first adjacency eigenspaces. Start with the 135-vertex rank-three graph; a zero-one classification or complete census would identify every sharp coordinate certificate. Neither is supplied here.

Related open work

No open PR studying this exact dual-polar cut ratio was found in the refreshed 846-PR corpus or the live precommit search. Direct predecessor #4628 and adjacent temporal-cut #6706 are merged. This proof supplies no competing physical program or temporal rank-cut value. The refresh also checked the latest claims and 118 recent account PR bodies. Sparse local-note and unpublished-work coverage remain incomplete.

The completion refresh read this account's seven other open PRs created since the CLAIM: #7547, #7548, #7549, #7550, #7552, #7553 and #7554. They study principal-demand stabilization, rounded commutator error, phase-B packing, cactus query acquisition, integer-gauge optimization, an ordinary four-cycle and modular phase-cache memory. None studies this dual-polar cut ratio or computes a competing value. No extra mathematical computation audited those results.

CI notes

CI did not certify this package. At unchanged head 9aec48e7507d4fb6e52170f40505b72e6e9e41f8, run 43522454629, plan job 132219860649, failed in Onboarding helper tests. I read the completed checks once and inspected the actual job log. It reports Ran 760 tests and FAILED (failures=1, skipped=24). The sole failing test is:

test_research_current_status.CurrentCiNavigationTests.test_current_ci_section_matches_selected_record
AssertionError: '`make frontier7000-verify`' not found

The mismatch is inherited from canonical main e54523517d0f03e58352e647c403a5af3e525a57. I read its Git objects directly: certificates/selected-result.json names make frontier7000-verify, while docs/ci-verification.md still has Current selected-family replay: #6030 and make frontier6030-verify. The selected record, CI document and failing test have no diff between that base and this head. Only this research package changes.

The dependent replay job 132221977413 was skipped because plan failed. Hosted CI executed no package replay. All scientific and export evidence above remains local.

I independently inspected the failed-step logs for rohanarun's run 42905094314, dillona's run 42903313790, and petalas's run 43522269247. Each has the identical sole failing test, missing-command assertion, 760-test total and 24 skips. These are the other runs verified here; no broader failure census is claimed.

No fix, retrigger commit, push, shared-file edit or workflow rerun was made. Maintainer repair of main is separate from this contribution. All registered checks already passed locally; independent human and formal proof-kernel review remain unclaimed.

Merge Danger

Door: two-way. This is an isolated research package.

Blast Radius: package. Six new files and two bounded opt-in replay commands; no shared implementation, frozen predecessor or selected-frontier file changes.

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