The expected root degree of the free uniform spanning forest need not be invariant under squaring an infinite connected locally finite simple vertex-transitive graph. For the binary grandparent graph G, of degree 8, two actual free-exhaustion reductions prove Edeg_FUSF(G)=(24-sqrt(2))/7 > 3.22. For its simple square G^2, of degree 40, a complete finite rational-current certificate on 3070 vertices and 19931 edges proves Edeg_FUSF(G^2) ≤ 3208526010913748780/10^18 < 3.21. All 797240 integer current entries, 122800 divergence equations and 40 energies are checked exactly. Finite Thomson/Rayleigh inequalities followed by free exhaustion justify the infinite free-forest bound; no wired-resistance substitution is made. With the explicitly adopted convention b_F(X)=Edeg_FUSF(X)/2-1, the strict difference exceeds 1/200. This answers the arbitrary-vertex-transitive assertion in Abért's Some questions (2 November 2010), printed p.4, Question24. That source does not print a normalization formula, so unequal expected degrees and the adopted convention are stated explicitly. The full automorphism group of G is nonunimodular. No negative assertion for the unimodular subclass, another nonunimodular Betti invariant, or the known Cayley case is made. The grandparent family and Kirchhoff, Thomson, Rayleigh and free-exhaustion methods are established background. Tang's grandparent-forest connectivity and branching work and Pete–Timár's distinct component-count noninvariance are credited. The proposed contribution is the specific expected-degree square comparison, not a new general electrical or forest mechanism. Bounded primary-source investigation authenticated no earlier publication of the complete comparison; no exhaustive novelty, historical firstness or continuing-openness guarantee is asserted. The actual inspected Tang and Pete–Timár bodies are arXiv v2 versions; final journal metadata were authenticated, but the final journal PDFs were inaccessible. The proof and data were already public in project PR1407 before this note; this version is dated 22 October 2026 and is not backdated. The proposed deposit contains the seven-page paper and a complete source-and-verification ZIP, including the exact current data, an independent standard-library checker, exact outputs, independent finite rational circuit diagnostics, both written infinite derivations, dependencies, priority limits, licenses and offline packaging instructions. No claimant verification program or external solver is trusted for the exact replay. Finite diagnostics do not replace the written universal proofs. AI tools were extensively involved in mathematical research and exploration, proof development and reconstruction, computational verification, literature and source investigation, manuscript preparation, editing and adversarial auditing. Alec Kriebel is the named human author. This is an unrefereed preprint without formal independent human peer review or journal acceptance. Internal AI audits are not human refereeing or proof-assistant certification. Original prose and certificate data are CC BY4.0; original verification and packaging code is MIT licensed.
No takes yet. Share an insight, caveat, or question.
Kriebel Alec (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: