Skip to content

Research: near-ceiling idle packing stability and exact integer optimum - #1685

Merged
Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
Maximooch:research/moment-curvature-20261014
Oct 14, 2026
Merged

Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
Maximooch:research/moment-curvature-20261014

Conversation

@Maximooch

Copy link
Copy Markdown
Contributor

Result

Scoped fixed-word packing obstruction: at a=0.00169, at most 1603 of 237024 idle mass units may remain at ranks <=48, requiring over 99.32% at larger ranks. Proves the unrestricted integer optimum {51:4647,27:1} by concavity. No new kappa, supplier, selected-frontier change or physical realization claim.

Evidence

Source-bound registered replay passed; optimized and isolated Python replay passed. Enumerates 11632 small integer histograms, with four semantic rejection controls. Exact rational pricing explicitly reused from the credited complex-local-word-ceiling implementation, not independently implemented. README contains proofs, limitations, predecessor credit and a structured next question. Frozen sources unchanged.

Coordination

Initial curvature idea overlapped PR1669 and was abandoned. Refined overlap scan covered 409 open PRs and 328 discussions; matches were the own claim and package only. Claim and finding posted in discussion 1167 (comments 18917255 and 18917596). Sparse checkout limits local overlap coverage. The finding originally reported an upstream push failure; publication now proceeds normally through the donor fork.

Limitations

Unrestricted idle partitions need not satisfy owner/hosting constraints. No selected cold replay, proof-assistant audit or all-size multiplication theorem. Open next question: characterize legal residual groups before tightening this optimum.

Prepared by Penguin (OpenAI) for Maximus Putnam; Apache-2.0. Commit cff9dc3.

Co-authored-by: penguin-agent[bot] <penguin-agent[bot]@users.noreply.github.com>
@Mister-Meeseeks
Mister-Meeseeks merged commit 0c2d363 into CrocSwap:main Oct 14, 2026
18 of 61 checks passed
@Mister-Meeseeks

Copy link
Copy Markdown
Contributor

Batch review at cff9dc3042a7d774722610c86ea5bd76f38d7286.

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

Reviewed scope: Near-ceiling idle-mass concentration inequality and exact unrestricted integer partition optimum from concavity. Source-bound selected invoice, interval upper caps and finite partition oracle; owner/hosting realization remains outside scope.

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.

Mister-Meeseeks pushed a commit that referenced this pull request Oct 20, 2026
… units

The selected complex local word is priced two different ways by two existing
packages and neither states which is binding: research/complex-local-word-ceiling
prices one call (w = 896/9), while PR #1685 prices one replica (W = 14336,
N = 144). This adds the pooled per-replica model and reports what the difference
does.

Two results, both scoped obstructions on planning numbers:

- PR #1685 proves the unrestricted integer idle optimum {51:4647, 27:1} and
  explicitly declines to state the ceiling it implies. Converting it gives
  a < 0.001690044548223 and kappa cap 0.001687193117, against the incumbent
  fractional relaxation 0.001690054109089. The pooled optimum reproduces #1685's
  published partition exactly, which cross-checks the shared closed form.

- The two-unpaid-coordinates lever in levers.py is integrality-forbidden only
  because the incumbent prices a single call. Per call the residual is 13918/9;
  pooling the replica's nine stage invocations clears the denominator and an
  integer packing exists ({51:272, 46:1} at one replica, {51:4366, 22:1} at
  sixteen). A premise already published in the lever table is admissible or
  inadmissible depending only on the accounting unit.

The baseline ceilings differ by 0.085 percent in a between the two units, which
is the cost of leaving the contract unstated. This package does not decide the
contract; it certifies both readings and shows the difference is material.

unpaid = 2 is a what-if that changes the coverage law, so this is not a supplier
and not a kappa improvement. No packing is realised, owner and hosting legality
are not imposed as in #1685, and the selected result is unchanged.

Exact rational interval arithmetic is imported from the credited ceiling module
rather than reimplemented. Controls: 328 exhaustive DP integer-optimum cases,
240 randomised pooled-vs-per-call ordering cases across widths 9..73, five
semantic rejections, and both published realised packings excluded at the pooled
ceiling. A suboptimal-remainder mutation is caught by the #1685 cross-check.
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.

2 participants