Erdős Problem a Daythe open tracker
ERDŐS #117 · READER SUBMISSION

I think this settles the problem as stated

VERIFIED2026-11-15 · ANONYMOUS
Community database OPENas of 2025-08-31tags group theoryproblem page · erdosproblems.com/117
Our review · 2026-11-16

Verdict: VERIFIED, as read by AI referees. Within the scope below, the claim holds. arXiv:2611.15068 determines h(n) for every n: h(1) = h(2) = 1, h(n) = n for 3 ≤ n ≤ 6, h(7) = h(8) = 10, and h(n) = 2^⌊(n−1)/2⌋ + 1 for n ≥ 9, with the maximizers for n ≥ 9 exactly the groups isoclinic to an extraspecial 2-group of order 2^(1+2⌊(n−1)/2⌋). Reviewed 16 November 2026. Full write-up: /day/117-abelian-covers-h-n-exact-refereed.

What we did. Two separately briefed AI referees, each told to refute, read the paper line by line with tools: one §§1–4, the other §§5–7. Both returned HOLDS WITH FIXES, with no objection that touches the theorem. The fixes: Lemma 2.2(ii) should be stated for lifts, not for points of the quotient (in its §§4–5 uses the quotient is elementary abelian, so read literally the hypothesis never holds there, but the proof works for lifts; the §6.3 use is fine as printed); Lemma 6.3 with r = 3 needs one line for "at least two"; Lemma 4.4 is correct but terse; a contradictory range on p. 10; an ordering issue in §6.6; and an unproved closing remark in §7. The cited results the seats could reach say what is used (Zarrin; Klein–Metsch; Ceria et al.; Abdollahi–Azad–Mohammadi Hassanabadi–Zarrin, Theorem 1.1). Both seats separately computed ω and a exactly on every nonabelian group of order at most 255 (6,518 groups) and found a ≤ T(ω) every time; a third, computation-only seat did the same for every group of order up to 256 (63,104). One referee also covered every group of orders 625 and 729, and tested the paper's lemmas on many thousands of instances; no check failed, and deliberately wrong variants did fail. Separately, we rebuilt Boris Alexeev's Lean formalization of the August √2 result on our machine.

Scope. No human expert has reviewed this that we know of. The identity h(n) = 1 + √ζ(n), and the claim that the same groups attain ζ(n), come from arXiv:2611.00943. We checked that it states what is used, but we did not referee its proofs. The nonsolvable case relies on the classification of finite simple groups, as the paper says.

Tools, as declared by the submitter: no ai

The submission

Let h(n) be the largest number of abelian subgroups needed to cover a group in which every set of more than n elements contains two distinct elements that commute, and let ζ(n) be the largest index of the centre under the same constraint. Erdős asked for the growth of h, Neumann proved that both quantities are finite, and Pyber bounded both ζ and h exponentially. We determine h(n) for every positive integer n and prove that h(n)=1+ζ(n)‾‾‾‾√ whenever the constraint admits a nonabelian group. The covers are built directly from the commutator geometry of a scalar quotient, and the comparison with ζ is made afterwards. Beyond finitely many small cases, both maxima are attained by the same groups, those isoclinic to an extraspecial 2-group of order 21+2⌊(n−1)/2⌋, confirming Pyber's prediction for the central index in this range.

Reviewed by Claude with Patrick White — an AI and a regular dude, not mathematicians. The verdict says what was checked; it is not community acceptance. ← problem page