Skip to content

Independently check #4416's paid-completion tensor bound at kappa 0.00268647: proof steps and arithmetic hold; gamma margin 8e-6; least v 5745 (no new κ) - #4428

Merged
Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
pinterbanget:pr4416-padding-check
Oct 21, 2026
Merged

Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
pinterbanget:pr4416-padding-check

Conversation

@pinterbanget

Copy link
Copy Markdown
Contributor

Summary

This adds research/pr4416-padding-check/. It is an independent check of 1337isnot1337's #4416, which proposes conditional κ = 0.00268647. The construction tensors zxv's p7 controller v = 6240 times, completes small targets to 11^v and splits paid jobs at 10^v.

Claim type: a written proof check of TENSOR.md §1–3 and §5, plus certified exact arithmetic. No new κ. #4416 is not accepted, and the native p7 audit is not re-checked.

Results

  • The proof steps hold.
    • Product moments multiply.
    • Completion to 11^v loses moment by concavity but buys separation; the bound conservatively keeps T_τ^v.
    • The split tails qJ^τ + r^τ ≤ 2RJ^{−C} and charge(D) < 2L/J^C hold.
    • The rank formula's general exit constant 2^{(111/26)v}, its −log₂η bound, θ < 25/26 and H0 ≥ 7v all hold.
  • The arithmetic is certified.
    • The p7 invoice totals match: 7,688 fields, 55,998 calls, rank-one masses 131,723 and 131,115.
    • At C = 0.00268648: α < 0.999999984404, γ < 0.99819171 and δ < 0.64382471.
    • The p7 mixed root lies in (0.0026864890486, 0.0026864890492).
    • The v = 6240 inequality holds by both routes.
    • Conditional κ = 0.00268647 via paid target padding after native composition #4416's H0 = 4,307,773,805,920 and H1 = 16,201,843,136,005,340,321 reproduce exactly.
  • Margins and minor improvements.
    • γ sits only 8·10⁻⁶ below its stated bound 4991/5000, and the tested saving is 9·10⁻⁹ below the root.
    • The least v for (b, c) = (11, 10) under the direct certified route is 5,745.
    • The same induction, with certified −log₂η, the exact exit constant and θ̄ = 145/151, gives H1 = 15,370,312,178,898,752,926, which is 5.4% smaller.

Evidence

python3.12 -B scripts/research_package.py --package research/pr4416-padding-check --check-only
  PASS_PACKAGE_METADATA_AND_PINS (not a mathematical acceptance check)
python3.12 -B scripts/research_package.py --package research/pr4416-padding-check
  PASS_REGISTERED_RESEARCH_PACKAGE research/pr4416-padding-check outcome=completed   (~10 s, stdlib only)

Head: ba0efeaac4e62d1757d7a9be6fd8aaf8f9bd35dc. Base: canonical main 12fa8c5c, priority revision 7. The audited files' SHA256 prefixes are in the README.

Not covered

  • The native p7 word.
  • The analytic, precision, tape and outer premises.
  • FOLLOWUP.md's 209-profile screen.

Task and overlap

Credit

@1337isnot1337 (#4416, #4414, #4412), @zxv (#4359), and Jacob Alan Sussman (#4268).

Prepared by Ryanson Jonathan with substantial Claude Code (Claude Opus 5.5) assistance; no independent human review. Apache-2.0. This PR does not merge anything or change the frontier.

🤖 Generated with Claude Code

…appa 0.00268647: proof steps and arithmetic hold; gamma margin 8e-6; least v 5745 (no new κ)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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