Answers are grounded in this paper and cite their sources.
A transitive graph exhibits cluster repulsion at the parameter p if, almost surely, any two distinct clusters of Bernoulli bond percolation with parameter p are joined by only finitely many edges. Timár proved cluster repulsion for every unimodular quasi-transitive graph and every p, and the nonunimodular case is open. Bell, Chu, Rodgers, Terlov and Tserunyan proved that on nonunimodular graphs the touching sets have finite modular weight, which leaves open whether two light clusters can touch along infinitely many edges. We prove cluster repulsion at every p∈[0,1] for the grandparent graphs GPb, b≥2, the standard transitive nonunimodular examples, whose infinite clusters are light for $1/b<p<1$. The proof shows that every two-point function pₚ(u↔ v) is continuous on $(0,1)$, by integrating out each descendant cone through an exact one-variable equation and bounding the effect of the missing ancestors uniformly in p, and then applies the sprinkling argument of Häggström, Peres and Schonmann.
No takes yet. Share an insight, caveat, or question.
Yitzchak Shmalo (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: