slug: mathar-2024-nonbonding-domino-pairs
bibkey: mathar2024nonbonding
doi: null
url: https://arxiv.org/abs/2404.18806v1
triage: theorem
motivation_gids:
  - D5/S3/StatisticalMechanics/HardCore/NonBondingDominoPairs.result

Mathar’s Conjecture 1 on two non-bonding dominoes

Problem

R. J. Mathar (arXiv:2404.18806, 2024) counts placements of dominoes on an r × c board that are non-bonding: every square of one is at L1 distance at least 2 from every square of another. They are hard dimers with nearest-neighbour exclusion. His Conjecture 1 (equation (21)) reads:

D(r,c,2) = 2c²r² − 2(cr² + c²r) + ½(r² + c²) − 22cr + (59/2)(c + r) − 30; r, c ≥ 3.

Issue #11068 fixes the reading. A domino is a set of two board squares at L1 distance 1, and D(r, c, 2) counts the sets of two pairwise non-bonding dominoes. The identity holds in ℚ for all r, c ≥ 3.

Motivation

The paper left the formula as a conjecture from its transfer-matrix tables. D5/S3/StatisticalMechanics/HardCore/NonBondingDominoPairs.result proves it for all r, c ≥ 3.

Gap

Issue #11068 preregisters the proof route and the literature check. arXiv has only version 1 and no later paper, Semantic Scholar lists no citation, and the OEIS has no entry for the rows. not-found-in-searched-scope.

Route

1. Every domino is horizontal or vertical and is determined by its anchor square.
2. Two distinct dominoes bond exactly when the offset of their anchors lies in a finite list:
   • 11 offsets for equal orientations, the equal domino included;
   • 12 offsets for mixed orientations.
3. The ordered anchor pairs with a given offset form a product of two interval overlaps. For r, c ≥ 3 each overlap is linear in r or in c.
4. Twice D(r, c, 2) is the number of ordered non-bonding pairs: N² minus the 46 offset classes, where N = r(c − 1) + (r − 1)c. Summing gives the polynomial.

Falsifier

The proof would fail if a bonded offset were missing from the lists of step 2, or if an overlap of step 3 were not linear for some r, c ≥ 3.

Evidence

Brute force from the literal definition (#11068):

• the formula matches for all 3 ≤ r, c ≤ 10;
• it fails for every tested board with r ≤ 2 or c ≤ 2;
• replacing the constant term −30 by −29 matches none of the 36 boards with 3 ≤ r, c ≤ 8.

The canonical source is D5/S3/StatisticalMechanics/HardCore/NonBondingDominoPairs.lean. Its public declarations are IsDomino, NonBonding, D2, claim, and result.

• Frozen module statement identity: sha256:f57332fcc4ec0c999803dfb6462e0d206bf0394329542d85010d8b9462d2eadd.
• Result declaration statement identity: sha256:9585653f22b4f09cf7b60c60161fab403e11db0867867aec7dc9b02e3747fdae.
• Freeze event: sha256:39e23454011ce8257a8f38c9c198c58c5074bdba412f9192c5c6f51b726ac9cb.

It has no project-level frozen prerequisite. The proof uses only the standard axioms propext, Classical.choice and Quot.sound; no sorry, native_decide, or new axiom.

Triage

Tier 1 external named conjecture (2024 paper), preregistered in issue #11068 before any Lean. theorem; resolution proved. The public theorem has proof_shape: content. The anchor bijection, the bond-offset classification and the overlap counts are new propositions on its live path. Admission basis open-problem-resolution; utility none.

ASSUMED-UNVERIFIED

The bounded literature check does not establish exhaustive worldwide novelty, priority, or the absence of an independent proof.