Wiki
Wiki

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

Updated

Problem 888

../

claims/: The 2 claim pages of Problem 888, one per claimant's result; the problem's standing derives from them.


Statement. What is the size of the largest A⊆{1,…,n}A\subseteq \{1,\ldots,n\} such that if a≤b≤c≤d∈Aa\leq b\leq c\leq d\in A are such that abcdabcd is a square then ad=bcad=bc?

Formulation. The wording asks for the size F(n)F(n) of the largest such set. The site reads it as asking for the order of magnitude and marks the problem solved on F(n)≍nlog⁡log⁡n/log⁡nF(n)\asymp n\log\log n/\log n, and the formal-conjectures statement asserts the same Θ\Theta bound; this page takes that reading.

Status. SOLVED (LEAN): the site labels the problem SOLVED (LEAN). The order is nlog⁡log⁡n/log⁡nn\log\log n/\log n: Erdős credits Sárközy with the bound o(n)o(n) (a proof was posted in the thread on 22 January 2026), the primes give ≫n/log⁡n\gg n/\log n, the primes together with the squarefree semiprimes give (1+o(1)) nlog⁡log⁡n/log⁡n(1+o(1))\,n\log\log n/\log n (posted in the thread in August 2025), and the matching upper bound ≪nlog⁡log⁡n/log⁡n\ll n\log\log n/\log n is the result of a manuscript of 25 April 2026 credited to GPT-5.5 Pro prompted by Chojecki, which the site's curator, Thomas Bloom, accepted (claim page). A research note released on 16 September 2026, its proof credited to GPT-6 Astra, sharpens the order to the asymptotic F(n)∼nlog⁡log⁡n/log⁡nF(n)\sim n\log\log n/\log n with a Lean development of its own; it is the site's one proof-claim entry and has no reply or review (claim page). A thread post of 18 January 2026 claimed the exact value π(n)+#{pq≤n:p<q primes}\pi(n)+\#\{pq\le n: p<q\text{ primes}\}, plus one for n≤14n\le14, with a working document as its write-up; replies in the same week showed that products of seven distinct primes can be added for large nn and that its Lemma 2 fails, and the post, not a dated manuscript, gets no claim page. The "(Lean)" suffix refers to a Lean proof registered by the formal-conjectures catalog, which was not built here. The site's commentary was last edited 28 May 2026; its thread and proof-claim form stand as described here as of 2026-10-07.

Source. erdosproblems.com/888, accessed 2026-09-04. Cite as: T. F. Bloom, Erdős Problem #888, https://www.erdosproblems.com/888.

Formalization. Statement in formal-conjectures.

Progress

Not yet compiled.

Known Results

Not yet compiled.