English

Unbounded logarithmic limsup in Erd\H{o}s problem 684

Number Theory 2026-04-29 v2 Combinatorics

Abstract

For 0kn0\le k\le n, write (nk)=uv\binom nk=uv where the primes dividing uu are at most kk and the primes dividing vv exceed kk, and let f(n)f(n) be the least kk with u>n2u>n^2; Erd\H{o}s problem 684 asks for bounds on f(n)f(n). We resolve the problem at the order level. By a short-multiplier construction nM=tLM1n_M=tL_M-1, where LM=lcm(1,,M)L_M=\operatorname{lcm}(1,\ldots,M) and tt is a multiplier of size exp(o(M))\exp(o(M)) extracted from a Fourier sieve, we prove that for every fixed C>1C>1 there exist integers nn with f(n)>(Co(1))logn, f(n)>(C-o(1))\log n, hence lim supnf(n)logn=. \limsup_{n\to\infty}\frac{f(n)}{\log n}=\infty. We thus refute the widely expected upper bound f(n)lognf(n)\ll\log n and place the order of f(n)f(n) strictly above logn\log n infinitely often. A matching polylogarithmic upper bound f(n)(logn)2f(n)\ll(\log n)^2 is known by Alexeev, Putterman, Sawhney, Sellke, and Valiant (arXiv:2603.29961). The reduction of the multiplier sieve to a dyadic fixed-Ω\Omega arithmetic-progression estimate, including a QM=M!/LMQ_M=M!/L_M box parametrization, a local harmonic-height cap, and an exact-aa product-shell extraction, is new. The required estimate uses Timofeev's mean-in-progressions framework together with a Burgess-based mod-pp saving on the relevant prime band.

Keywords

Cite

@article{arxiv.2604.23784,
  title  = {Unbounded logarithmic limsup in Erd\H{o}s problem 684},
  author = {Ji Ho Bae},
  journal= {arXiv preprint arXiv:2604.23784},
  year   = {2026}
}

Comments

21 Pages

R2 v1 2026-07-01T12:35:53.306Z