Skip to content

Prove a 5t/2 homogeneous mixed-Q8 lower bound - #3578

Merged
Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
petalas:research/w4-b19-q8-saturation-20261018
Oct 20, 2026
Merged

Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
petalas:research/w4-b19-q8-saturation-20261018

Conversation

@petalas

@petalas petalas commented Oct 18, 2026

Copy link
Copy Markdown
Contributor

Summary

homogeneous inverse-generator targets, both positive i/j labels, arbitrary restored dirt
- lower bound n >= 2t; four identical targets could still use eight
+ 2ni+nj >= 4t and 2ni+3nj >= 6t, hence n >= 5t/2
+ four identical targets require ten or twelve; ten forces counts 6/4

The author-written proof charges both boundary ranks using the homogeneous zero-product block and both orientations of PR3570's quadratic/cubic identities. It closes the homogeneous eight-call handoff. The ten-call decision and mixed 2/2 boundary remain open. No physical supplier, new kappa or selected-frontier change is claimed.

Exact used PR3570/3536/3558 proof bytes, origins, attribution and source closure are retained. Canonical investigation.

Evidence

  • Before: PR3570's count relaxation admitted four homogeneous targets at eight balanced-label calls.
    After: the new inequalities reject eight and force six target-positive and four opposite calls at ten. Twelve is attained by direct words.
  • At actual commit 4395a9cf3edf9ba36918c48477f36dcde110f45a, CPython 3.12 registered science and optimized fault controls pass, as does scripts/check_research_commit.py --package research/q8-homogeneous-mixed-floor --check-source-history for all 15 files. Scoped Ruff E/F/B/UP passes on the exact committed scientific bytes.
  • Scientific replay checks 390 generic words on 9,360 independent dirty columns, 1,593 triangular rank fixtures, 186 zero-product pairs and 36 direct endpoints with 378 omissions. Exact normal/optimized scientific and optimized control commands also pass in an owned Git-free committed-byte export.
  • Reproduce with python3.12 -Werror -B scripts/research_package.py --package research/q8-homogeneous-mixed-floor. PROOF.md, RESULT.json and VALIDATION.md separate the universal proof, finite evidence and unrun obligations. Independent human/formal proof-kernel acceptance, the selected cold replay, Lean, Docker and hosted completion are unclaimed.

Merge Danger

Door: two-way

Blast Radius: isolated

One new opt-in research package. No shared engine, selected record, frozen predecessor or supplier changes.

@Mister-Meeseeks
Mister-Meeseeks merged commit 65d9fb4 into CrocSwap:main Oct 20, 2026
2 checks passed
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