Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated

Problem 841

../

claims/: The 2 claim pages of Problem 841, one per claimant's result; the problem's standing derives from them.


Statement. Let tnt_n be minimal such that {n+1,…,n+tn}\{n+1,\ldots,n+t_n\} contains a subset whose product with nn is a square number (and let tn=0t_n=0 if nn is itself square). Estimate tnt_n.

Formulation. Erdős originally asked, as the site's commentary records, whether the integers nn with tn≥n1−o(1)t_n\ge n^{1-o(1)} have density zero. The answer is yes. By the bound of Granville and Selfridge on their claim page, tn=P(n)t_n=P(n) when P(n)>2n+1P(n)>\sqrt{2n}+1 and tn≤3n/2+1t_n\le3\sqrt{n/2}+1 otherwise, where P(n)P(n) is the largest prime factor of nn. So for fixed 0<δ<1/20<\delta<1/2 and large nn, tn≥n1−δt_n\ge n^{1-\delta} forces P(n)=tn≥n1−δP(n)=t_n\ge n^{1-\delta}, and the integers with P(n)≥n1−δP(n)\ge n^{1-\delta} have density log⁡(1/(1−δ))\log(1/(1-\delta)), which tends to 00 with δ\delta (a remark of this page); Theorem 1.1 of Bui, Pratt and Zaharescu gives the same conclusion. The site's Statement asks instead for an estimate of tnt_n.

"Estimate tnt_n" is open-ended. The site labels the problem SOLVED after listing Selfridge's exact value tn=P(n)t_n=P(n) when P(n)>2n+1P(n)>\sqrt{2n}+1, with tn≪n1/2t_n\ll n^{1/2} otherwise, and three results of Bui, Pratt and Zaharescu. First, the proportion of n≤xn\le x with tn≤nct_n\le n^c tends to that with P(n)≤ncP(n)\le n^c, which is ρ(1/c)\rho(1/c). Second, at least x1−o(1)x^{1-o(1)} integers n≤xn\le x have tn≤exp⁡(O(log⁡nlog⁡log⁡n))t_n\le\exp(O(\sqrt{\log n\log\log n})). Third, tn≫(log⁡log⁡n)6/5(log⁡log⁡log⁡n)−1/5t_n\gg(\log\log n)^{6/5}(\log\log\log n)^{-1/5} for sufficiently large non-square nn. The page's standing reads the question as answered by these results; they do not determine the order of tnt_n for an individual nn.

Status. SOLVED, in the site's label (page last edited 14 October 2025), which describes the Statement above.

Source. erdosproblems.com/841, accessed 2026-09-04. Cite as: T. F. Bloom, Erdős Problem #841, https://www.erdosproblems.com/841.

References.

  • [BPZ24] Bui, Hung M. and Pratt, Kyle and Zaharescu, Alexandru, A problem of Erdős-Graham-Granville-Selfridge on integral points on hyperelliptic curves. Math. Proc. Cambridge Philos. Soc. (2024), 309-323.
  • [ErSe92] Erdős, Paul and Selfridge, J. L., Problems and Solutions: Solutions: 6655. Amer. Math. Monthly (1992), 791-794. Granville and Selfridge cite the item as P. T. Bateman, P. Erdős and J. L. Selfridge, Getting a square deal, Amer. Math. Monthly 99 (1992), 791-794.
  • [Gu04] Guy, Richard K., Unsolved problems in number theory. 3rd ed., Problem Books in Mathematics, Springer, New York (2004), xviii+437 pp. Section B30 "A small set whose product is square", pp. 128--129, states the problem with tnt_n in the page's notation, records that the Thue--Siegel theorem gives tn→∞t_n\to\infty faster than a power of ln⁡n\ln n, a sentence on which the section prints comments by Granville and Silverman, and reports Selfridge's bound tn≤max⁡(P(n),3n)t_n\le\max(P(n),3\sqrt n) with P(n)P(n) the largest prime factor of nn. Library home: guy_2004_unsolved_problems_number_theory.

Formalization. Statement in formal-conjectures. A Lean development of the Bui–Pratt–Zaharescu results by OpenAI Codex, held in Boris Alexeev's lean-proofs repository, is recorded as the claimants' formalization link on their claim page; this corpus has not built it.

Progress

Not yet compiled.

Known Results

Not yet compiled.

Linked library material

These entries are derived from explicit links on library pages. They are navigation only and do not by themselves record mathematical progress.