Off-diagonal tripartite Ramsey 2^O(n log n) bound
Informal conjecture on tripartite off-diagonal Ramsey numbers · arXiv:2404.02021
Status open high confidence
No follow-up paper resolving the conjecture that $r(H,K_{n,n,n}^{(3)})\leq 2^{O_H(n\log n)}$ for all $H$ was found. A closely related successor paper (arXiv:2411.13812, same author group, November 2024) addresses the polynomial-vs-exponential threshold for $r(H,K_n^{(3)})$ — the non-tripartite complete 3-graph — but does not treat the tripartite $K_{n,n,n}^{(3)}$ version. The known upper bound of $2^{O_H(n^{2-\epsilon})}$ and the $2^{\Theta(n\log n)}$ threshold in constructions for $K_n^{(3)}$ remain the closest context.
Reviewer notes. The conjecture concerns $r(H,K_{n,n,n}^{(3)})$, the off-diagonal Ramsey number against the complete tripartite 3-uniform hypergraph. The companion paper arXiv:2411.13812 (Conlon et al., November 2024) characterises when $r(H,K_n^{(3)})$ is polynomial, but $K_n^{(3)}$ and $K_{n,n,n}^{(3)}$ are distinct objects and that paper does not address the tripartite setting. A 2026 paper (arXiv:2603.16069, He–Nie–Post–Verstraete) studies quasipolynomial growth for a different Ramsey function. No paper resolving the $2^{O_H(n\log n)}$ upper bound for the tripartite case was found after an exhaustive 5-call search.
Context
The authors note that $r(H,K_{n,n,n}^{(3)})\leq 2^{O_H(n^2)}$ for all $H$, and using the methods of [5] one can show a stronger bound $2^{O_H(n^{2-\epsilon})}$ for some $\epsilon>0$ depending on $H$. They suspect the true bound is $2^{O_H(n\log n)}$, which would help explain why the threshold $2^{\Theta(n\log n)}$ is so prevalent and hard to surpass in constructions for $K_n^{(3)}$.
Notes. Introduced with 'We suspect that it may even be the case' in the discussion following Conjecture 4.6.
Source paper
On off-diagonal hypergraph Ramsey numbers
David Conlon, Jacob Fox, Benjamin Gunby, Xiaoyu He, Dhruv Mubayi, Andrew Suk, Jacques Verstraete · 2024-04-02
https://arxiv.org/abs/2404.02021