Table of Contents
Fetching ...

Divisor Structure of p-1 in Mersenne Prime Exponents

Jesus Dominguez

Abstract

According to the Wagstaff heuristic, the probability that a Mersenne number $M_p = 2^p-1$ is prime mainly depends on the size of the exponent $p$. We investigate whether the secondary arithmetic structure in $p-1$ is linked to noticeable variation among Mersenne prime exponents, motivated by the cyclotomic decomposition of $2^{p-1}-1$. We introduce the normalized divisor-structure parameter $S(p)=\log τ(p-1)/\log\log p$, which is a scale-invariant way to measure the divisor structure of $p-1$. Using the currently known Mersenne prime exponents (excluding small cases), we compare $S(p)$ with nearby prime controls through percentile analysis, stratified conditional likelihood estimation and stratified permutation testing. Across all methods, Mersenne prime exponents tend to exhibit elevated values of $S(p)$ within local exponent windows compared to nearby primes of comparable size. The observed effect is moderate and remains stable across different non-parametric tests. It is also compatible with the classical asymptotic behavior. An analytic mechanism has not yet been established, and a theoretical explanation of the observed phenomenon remains open.

Divisor Structure of p-1 in Mersenne Prime Exponents

Abstract

According to the Wagstaff heuristic, the probability that a Mersenne number is prime mainly depends on the size of the exponent . We investigate whether the secondary arithmetic structure in is linked to noticeable variation among Mersenne prime exponents, motivated by the cyclotomic decomposition of . We introduce the normalized divisor-structure parameter , which is a scale-invariant way to measure the divisor structure of . Using the currently known Mersenne prime exponents (excluding small cases), we compare with nearby prime controls through percentile analysis, stratified conditional likelihood estimation and stratified permutation testing. Across all methods, Mersenne prime exponents tend to exhibit elevated values of within local exponent windows compared to nearby primes of comparable size. The observed effect is moderate and remains stable across different non-parametric tests. It is also compatible with the classical asymptotic behavior. An analytic mechanism has not yet been established, and a theoretical explanation of the observed phenomenon remains open.
Paper Structure (12 sections, 65 equations, 2 tables)