---

Claim. The conjectured asymptotic of Problem 1131,

$$\min I=2-(1+o(1))\frac{1}{n},$$

is false. Liam Price posted the claim in the site's discussion thread on 26 April 2026, writing that GPT-5.5 Pro claims a disproof of the asymptotic the problem displays, with a link to a write-up on an online editor. The read link exposes only the editor's shell and no document to a reader without an account, so the corpus records the target of the claim from the thread post: the displayed asymptotic is the only asymptotic the problem states, and a disproof of it says that $n(2-\min I)$ does not tend to $1$. Price is the claimant, with the system named as the post names it.

Submission note. Posted to the site's forum by Liam Price on 26 April 2026:

GPT-5.5 Pro claims a disproof of the claimed asymptotic here.

Covers. The second question of Problem 1131, whether $\min I=2-(1+o(1))/n$: the answer claimed is no. The first question, the minimal value of $I$, is not determined. The bounds $2-O((\log n)^2/n)\le\min I\le 2-2/(2n-1)$ of Erdős, Szabados, Varma and Vértesi, which the site's commentary records, are consistent with either answer; Brutman and Toledano's numerical study (Comput. Math. Appl. 34 (1997), no. 12, 37--47) had pointed to a limit of $n(2-\min I)$ near $1.09$, as the problem page records.

Depends on. No page of this wiki.

Acceptance. None recorded. Nat Sothanaphan replied in the thread on 27 April 2026 that a standard check found no issues, and noted that the Brutman--Toledano paper had earlier given numerical evidence against the asymptotic; Sothanaphan is not the site's curator, so the reply is recorded here and not listed as reviewed evidence. The site labels the problem OPEN (page last edited 2026-01-23) and no journal publication, referee report or Lean proof of the result is recorded, so no evidence is listed and the claim is pending.