Polynomial 3-graph Ramsey numbers via blowup characterization

Conjecture 1.1 · arXiv:2411.13812

arXiv Conjecture high confidence— first stated 2025-10-29

Status open high confidence

Conjecture 1.1 from arXiv:2411.13812 proposes a full biconditional characterization of 3-graphs H for which r(H,K_n^(3)) grows polynomially in n. The source paper establishes the conjecture for the special cases when H is tightly connected or has at most two tight components, but the full general statement remains open. A March 2026 follow-up (arXiv:2603.16069) by co-author Xiaoyu He constructs 3-graphs with quasipolynomial growth rates n^{Theta(log n)}, contributing to the broader program of classifying growth rates but not resolving the conjecture in full generality.

Cited literature (1)

  • Xiaoyu He · arXiv preprint · arXiv:2603.16069

    Constructs a 3-graph H_2 on six vertices with r(H_2, K_n^(3)) = n^{Theta(log n)}, establishing a quasipolynomial (superpolynomial but subexponential) growth regime for specific 3-graphs, directly relevant to the classification program of Conjecture 1.1 but not resolving the full biconditional.

Reviewer notes. The conjecture is very recent (October 2025); no paper fully resolving it was found in the indexed literature. The source paper itself proves the conjecture for H tightly connected or with at most two tight components. A related preprint arXiv:2507.09434 (Ascoli, He, Yu; July 2025) proves the Erdos-Hajnal conjecture for k=3 via iterated-blowup extremizers and establishes polynomial-to-exponential transition results, but predates arXiv:2411.13812 and cannot be a follow-up to Conjecture 1.1.

Auto-reviewed 2026-05-14 with claude-sonnet-4-6 (web search enabled).

Conjecture. For a $3$-graph $H$, there exists a constant $c$ depending only on $H$ such that $r(H,K_{n}^{(3)})\leq n^{c}$ for all $n$ if and only if $H$ is a subgraph of an iterated blowup of an edge.

Context

The paper proposes a full classification of those $3$-graphs $H$ for which the off-diagonal Ramsey number $r(H,K_{n}^{(3)})$ grows polynomially with $n$. The 'if' direction (polynomial upper bound when $H$ is iterated tripartite) was already known to Erdős and Hajnal; the 'only if' direction is new. The problem was first raised explicitly by the first author at an AIM workshop in 2015.

Notes. First author (Conlon) posed the problem at a 2015 AIM workshop (ref [5]); the formal conjecture as stated here appears to be the paper's own formulation.

Source paper

When are off-diagonal hypergraph Ramsey numbers polynomial?
David Conlon, Jacob Fox, Benjamin Gunby, Xiaoyu He, Dhruv Mubayi, Andrew Suk, Jacques Verstraëte, Hung-Hsun Hans Yu · 2025-10-29
https://arxiv.org/abs/2411.13812