A five-vertex counterexample to ferromagnetic Potts censoring
Authors/Creators
Description
Anonymous AI-assisted unrefereed candidate: an exact five-vertex counterexample shows that deleting one predetermined interior heat-bath update can reduce finite-horizon total-variation distance for the zero-field uniformly interacting ferromagnetic three-state Potts model started monochromatically. At activity 30, the exact full-minus-censored gap is 7905357280856578194954129502105 / 766036711510586802141859485820665762204, strictly positive.
The package includes a readable proof, labelled graph, all 243-state final distributions, a 27-entry marginal certificate, three producer-side exact implementations, adversarial controls, exact intermediate-horizon comparisons and an earlier all-sites-updated witness. Five differentiated producer-coordinated internal editorial reports and their frozen submission are retained. These do not establish unaffiliated validation or journal peer review.
The result is not a mixing-time, random-scan, practical speed-up, minimality or all-colour-count theorem. Historical priority and external specialist review remain unestablished. The inaccessible AIM endpoint is distinguished from its dataset transcription and Holroyd's primary-source statement. Original prose and data are CC0; original code is MIT. Third-party papers and supplied-review prose are not redistributed or relicensed. Publisher: Evidence Press.
Files
potts-censoring-counterexample-0.1.0-candidate.pdf
Additional details
Related works
- Is identical to
- https://github.com/ipitchford/potts-censoring-counterexample/releases/tag/v0.1.0-candidate (URL)
- Is supplement to
- https://evidencepress.org/releases/potts-censoring-counterexample/ (URL)