← All problems

Strong-connectivity threshold conjecture for bidirected percolation

Let GG be an infinite, locally finite, transitive graph. Replace each undirected edge of GG by two oppositely directed edges, independently declaring each directed edge open with probability pp. Let pc(G)p_c(G) be the usual Bernoulli bond-percolation threshold, and let psc(G)p_{\mathrm{sc}}(G) be the infimum of the values of pp for which there exists almost surely an infinite strongly connected cluster, meaning an infinite set of vertices such that every pair is mutually reachable by open directed paths. The conjecture is that

psc(G)=pc(G)if and only ifG is amenable;p_{\mathrm{sc}}(G)=p_c(G) \quad\text{if and only if}\quad G\text{ is amenable};

equivalently, psc(G)>pc(G)p_{\mathrm{sc}}(G)>p_c(G) for every transitive non-amenable graph GG.

References

Source: arXiv

Progress summary

Refreshed
Open · progress reported

A new unrefereed paper reports the predicted threshold equality for amenable graphs and separation for spread-out non-amenable ones, but leaves the general non-amenable case open.

The conjecture concerns how strong connectivity in bidirected percolation compares with ordinary percolation across transitive graphs. The latest work separates the amenable and non-amenable regimes without settling the full conjecture.

October 2026 development

Yucheng Liu, Leonardo T. Rolla, and Yuliang Shi report that the threshold equality holds for every transitive amenable graph, while strict separation holds for sufficiently spread-out modifications of every transitive non-amenable graph. The preprint is unrefereed and does not settle every transitive non-amenable graph.

Current status (as of October 2026): The amenable case and a spread-out non-amenable subclass are claimed established, while the full transitive non-amenable case remains open.

Sources

Solutions 0

No solutions have been posted yet.