Answers are grounded in this paper and cite their sources.
Consider, for n ≥ 3, dice with n faces whose face values lie in {1,…,n}, and let Gn be the directed graph on these dice in which X → Y if X beats Y in exactly ⌊n2/2⌋ + 1 of the n2 pairs of faces, the smallest possible advantage above one half. Easlick showed that n = 6 is the smallest n for which the longest directed cycle of Gn has length exactly n, and conjectured that it is also the largest; this was soon proved in a Lean formalisation, using cycles whose length is linear in n. We show that for every n ≥ 7 the graph Gn contains directed cycles of every length from 3 to (15n − 67)/2, and that the length c(n) of its longest cycles grows exponentially: c(n) ≥ 2(n+4)/5 for every n ≥ 7, and log2 c(n) ≥ (1/4 − o(1)) n, while trivially c(n) < 4n. The proofs combine two embeddings, Gn ↪ Gn+1 for even n and Ga × Gb ↪ Ga+b for odd a and b, with a detour lemma, a few computer-verified seed cycles and a Gray-code construction. We also show that the non-trivial strong component of G7 is not Hamiltonian, and record exact counts of short cycles for n ≤ 9. The seed cycles (certificates), the verification programs and the LaTeX source are in supplementary.zip (see anc/README.txt there).
No takes yet. Share an insight, caveat, or question.
Vadim Genkin (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: