Flash-rainbow tournament edge-coloring formula

Conjecture 1.7 · arXiv:2305.13422

arXiv Conjecture high confidence— first stated 2023-06-01

Authors: António Girão, Freddie Illingworth, Lukas Michel, Michael Savery, Alex Scott

Status open high confidence

Conjecture 1.7 states that $t(l,k) = l^{k-1}$, where $t(l,k)$ is the smallest integer such that every edge-colouring of any tournament on more than $t(l,k)$ vertices contains an $l$-flash or a $k$-rainbow; this is a tournament strengthening of the Lefmann–Rödl–Thomas conjecture. The source paper proves the conjecture asymptotically for $l = \omega(k^2)$ and shows transitive tournaments are not always worst-case, but the full conjecture remains open. No follow-up paper resolving the full statement was found in the literature through May 2026.

Reviewer notes. No follow-up found addressing Conjecture 1.7 specifically. The paper was published in Combinatorica 44 (2024), no. 3, 675–690, with a publisher erratum issued in 2024 correcting Corollary 2.5 (unrelated to the conjecture). The single internal reference (arXiv:2508.14332) is a false corpus match.

Auto-reviewed 2026-05-15 with claude-sonnet-4-6 (web search enabled).

Conjecture. For all positive integers $l$ and $k$, $t(l, k) = l^{k-1}$.

Context

Here $t(l,k)$ denotes the smallest integer such that every colouring of the edges of any tournament with more than $t(l,k)$ vertices contains an $l$-flash or a $k$-rainbow. Since $t(l,k) \geq f(l,k) \geq l^{k-1}$, this is a tournament strengthening of the Lefmann–Rödl–Thomas Conjecture 1.2. The paper proves two upper bounds on $t(l,k)$ confirming it asymptotically for $l = \omega(k^2)$, and also shows that some tournaments are strictly worse than transitive tournaments (Theorem 1.10).

Source paper

Flashes and rainbows in tournaments
António Girão, Freddie Illingworth, Lukas Michel, Michael Savery, Alex Scott · 2023-06-01
https://arxiv.org/abs/2305.13422 PDF source