Triangle-free graph criterion for discontinuity in lower-tail probabilities

From papers

Let HH be a fixed graph, let p(0,1)p\in(0,1) be constant, and set

m=p(n2).m=p\binom{n}{2}.

Let GmG(n,m)G_m\sim G(n,m), let NH(Gm)N_H(G_m) denote the number of copies of HH in GmG_m, and write

μH=E[NH(Gm)].\mu_H=\mathbb{E}\left[N_H(G_m)\right].

Triangle-free graph discontinuity conjecture. The function

βloglogP(NH(Gm)<(1nβ)μH)1logn\beta\mapsto\frac{\log\log\mathbb{P}\left(N_H(G_m)<(1-n^\beta)\mu_H\right)^{-1}}{\log n}

has a jump discontinuity at β=1\beta=-1 if and only if HH is triangle-free.

This conjecture proposes a general criterion for whether lower-tail moderate deviations exhibit a discontinuity near δ=n1\delta=n^{-1}. The paper describes the problem as wide open beyond the triangle case and relates the bipartite-graph behavior to Siderenko's conjecture.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Primary source

José Alvarado, Gabriel Dias and Simon Griffiths, “Moderate Deviations of Triangle Counts in the Erdős-Rényi Random Graph G(n,m): The Lower Tail”, arXiv:2403.13792 (2025).

Additional references

8 papers in this index state this conjecture (2016–2024). The statement above is taken from the most recent of them; the others are arXiv:2109.02906, arXiv:2108.06359, arXiv:2009.05428, arXiv:2006.05511, arXiv:1801.06887, arXiv:1708.08439, arXiv:1601.05762.

Solutions 0

No solutions have been posted yet.