On the Number of Prime Factors of Consecutive Integers

Cheuk Fung (Joshua) Lau · 2026

Read On the Number of Prime Factors of Consecutive Integers by Cheuk Fung (Joshua) Lau (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-15042.txt.

Introduction

Lau proves there are infinitely many integers n whose shifted values n plus k all carry few prime factors. For every shift k from 2 upward, the prime-factor count stays within a logarithmic bound in k. This sharpens a 2025 result of Tao and Teräväinen, which allowed a bound growing linearly in k. The proof refines their probabilistic argument with a sharper sieve and moment estimates, and the paper adds a conjecture under which the bound is essentially sharp.

Catalog record

CREATORCheuk Fung (Joshua) Lau
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE84035
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.15042

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On the Number of Prime Factors of Consecutive Integers Cheuk Fung (Joshua) Lau Abstract . We prove that there are infinitely many n n such that ω ⁡ ( n + k ) ≪ log ⁡ k \omega(n+k)\ll\log k for all integers k ≥ 2 k\geq 2 . This improves on a result of 7 , who has O ⁡ ( k ) O(k) in place of O ⁡ ( log ⁡ k ) O(\log k) . As corollaries, we make progress on a number of questions posed by Erdős. The proof is based on a quantitative refinement of the Tao-Teräväinen probabilistic argument, combining a more efficient sieve procedure with stronger exponential concentration-of-measure estimates. Moreover, we formulate a conjecture on integers with many prime factors based on Cramér-type random models. Assuming this conjecture, the main bound is essentially sharp. 2020 Mathematics Subject Classification Primary 11N56; Secondary 11N36 1. Introduction In this paper we are interested in when there are strings of consecutive integers each having ‘few’ prime factors. More specifically, we study variants of the following question. Question . Given a function f:\mathbb{N}\to\mathbb{N} , are there infinitely many integers n such that \omega(n+k)\leq f(k) for all integers k\geq 1 ? Here \omega(n)=\sum_{p\mid n}1 is the number of distinct prime factors of n , and one can ask similar questions for the closely related functions \Omega(n):=\sum_{p^{j}\mid n}1 (the number of prime factors counting multiplicity) or \tau(n):=\sum_{d\mid n}1 (the number of divisors of n ). One can also ask for \omega(n-k)\leq f(k) for all 1\leq k<n , which has essentially the same difficulty. Erdős asked several questions of this type, including (in increasing order of difficulty): Conjecture 1 ( 2 , 7 ) . There are infinitely many n such that \omega(n+k)\leq\Omega(n+k)\ll k for all integers k\geq 1 . Conjecture 2 ( 3 ) . There are infinitely many n such that \omega(n-k)\leq k for all integers 1\leq k<n . Conjecture 3 ( 2 ) . There are infinitely many n such that \tau(n+k)\ll k for all integers k\geq 1 . Conjecture 4 ( 3 ) . For \varepsilon>0 , there are infinitely many n such that \omega(n-k)\leq(1+\varepsilon)\frac{\log k}{\log\log k} for all integers 1\ll_{\varepsilon}k<n . Moreover, there are infinitely many n such that \Omega(n-k)\leq(1+\varepsilon)\frac{\log k}{\log 2} for all integers 1\ll_{\varepsilon}k<n . These are listed as Erdős Problem #248, #413, #826 and #679 respectively on the website erdosproblems maintained by Thomas Bloom cataloguing problems posed by Erdős. Conjecture 1 was recently solved by 7 . In this paper, we prove the following result, which is an improvement on Theorem 1.1 of 7 . Theorem 1.1 . There exists a positive constant C such that, for infinitely many positive integers n , one has \omega(n+k)\leq\Omega(n+k)\leq C\log k for every integer k\geq 2 . Using random models, we conjecture this is best possible up to a constant. Conjecture 5 . Theorem 1.1 is best possible up to a constant, that is, for any \varepsilon>0 and n sufficiently large in terms of \varepsilon , there exists integers k_{1},k_{2}\geq 2 such that \omega(n+k_{1})>(1-\varepsilon)\log k_{1} and \Omega(n+k_{2})>(1-\varepsilon)\log k_{2}/\log 2 . As a corollary of Theorem 1.1 , we have a weaker version of Conjecture 3 (Erdős Problem #826). Corollary 1.2 . There is an absolute constant C such that there are infinitely many n satisfying \tau(n+k)\ll k^{C} for all integers k\geq 1 . With the same proof, a version of Theorem 1.1 for n-k also holds. Theorem 1.3 . There exists a positive constant C such that, for infinitely many positive integers n , one has \omega(n-k)\leq\Omega(n-k)\leq C\log k for every integer 1<k<n . We also speculate that the analogous version of Conjecture 5 holds. Conjecture 6 . Theorem 1.3 is best possible up to a constant, that is, for any \varepsilon>0 and n sufficiently large in terms of \varepsilon , there exists integers 1<k_{1},k_{2}<n such that \omega(n-k_{1})>(1-\varepsilon)\log k_{1} and \Omega(n-k_{2})>(1-\varepsilon)\log k_{2}/\log 2 . Theorem 1.3 misses the first part of Conjecture 4 (Erdős Problem #679) by a \log\log k factor, and misses the last part by a constant. Moreover, Conjecture 6 disproves the first claim in Conjecture 4 (Erdős Problem #679). Using Theorem 1.3 , we immediately obtain a corollary, which is a weaker version of Conjecture 2 (Erdős Problem #413). Corollary 1.4 . There are infinitely many n such that \omega(n-k)\leq k for all integers 1\ll k<n . 2. Outline In this section, we outline the proof of Theorem 1.1 . Following the broad strategy of 7 , we construct a random variable \mathbf{n} taking values in [x,2x] and show that with positive probability, \omega(\mathbf{n}+k)\ll\log k for all k\geq 2 . By the union bound, it suffices to prove that (2.1) \sum_{k=2}^{\infty}\mathbb{P}\left(\omega(\mathbf{n}+k)\geq C\log k\right)<1 for a sufficiently large constant C . As noted in 7 , the terms where k is very large (e.g., k\geq x^{1/100} ) are handled by trivial bounds on \omega(n+k) . The main challenge lies in the range k<x^{1/100} . To prove ( 2.1 ), we construct \mathbf{n} by weighting the uniform distribution on [x,2x] with a product of independent Selberg-type sieves. Specifically, we choose n\in[x,2x] with probability proportional to w(n)\coloneqq\mathds{1}_{p\mid n\,\forall p\leq 0.15\log x}\prod_{k=1}^{K}w_{R_{k}}(n+k)^{2}, where w_{R_{k}}(n+k) are sieve weights of Goldston–Pintz–Yıldırım type and K=(\log x)^{1/1000} . Morally, these weights act as a proxy for the indicator function that n+k has no prime factors less than the sieve level R_{k} . A key technical requirement for this construction is that the product of the sieve levels must satisfy \prod_{k=1}^{K}R_{k}<x^{\theta} for some \theta<1 . However, to ensure that n+k has few prime factors, we want each R_{k} to be as large as possible. To balance these requirements, we choose a polynomial decay for the sieve levels: R_{k}=x^{c/k^{50}} for a small constant c>0 . This choice of polynomial decay is a quantitative refinement of 7 and is essential for the sharp bounds required in Theorem 1.1 . Under this weighting, an integer n+k behaves like a random integer conditioned to have no prime factors smaller than R_{k} . If n+k has no prime factors below R_{k} , it trivially has at most \log(2x)/\log R_{k}\ll k^{50} prime factors, but this is far too large for our purposes. Instead, we exploit the fact that a typical integer with no prime factors below R_{k} has roughly \log\left(\frac{\log x}{\log R_{k}}\right)\approx\log(k^{50}/c)\asymp\log k prime factors. To ensure that all n+k simultaneously exhibit this typical behavior, we prove a strong concentration of measure result. Specifically, we show that the number of prime factors \omega(n+k) satisfies a central limit theorem-type bound under the sieve weights, with mean and variance proportional to \log(\log x/\log R_{k}) . From this concentration, we deduce that the probability of the tail event is small: (2.2) \mathbb{P}\left(\omega(\mathbf{n}+k)\geq C\log k\right)\ll e^{-c^{\prime}C\log k}=\frac{1}{k^{c^{\prime}C}}. To establish ( 2.2 ), we derived an upper bound for the exponential moment. By contrast, 7 obtained a comparable bound for \mathbb{P}(\omega(\mathbf{n}+k)\geq Ck) using second-moment estimates. This perspective plays an important role in enabling our quantitative refinement. By choosing C sufficiently large, we ensure this probability is less than 1/(100k^{2}) , allowing the sum in ( 2.1 ) to converge and remain less than 1. 3. Acknowledgements I am very grateful to my supervisor, James Maynard, for suggesting this problem and for many helpful comments and

Read the complete text