ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let a>0a>0 be a real number satisfying the relevant condition, and write α=log⁡10a\alpha=\log_{10}a. Suppose that the continued fraction expansion of α\alpha satisfies

lim⁡k→∞ak2∑i=1kai2=0.\lim_{k\to\infty}\frac{a_k^2}{\sum_{i=1}^k a_i^2}=0.

For a digit d∈{1,2,…,9}d\in\{1,2,\dots,9\}, let Ed(N,{an})E_d(N,\{a^n\}) denote the Benford error, and assume that dd does not satisfy the bounded-error condition. Central limit theorem for Benford errors. There exist sequences {AN}\{A_N\} and {BN}\{B_N\} such that, for all u<vu<v,

lim⁡N→∞1N#{n≤N:u≤Ed(N,{an})−ANBN<v}=12π∫uve−x2/2 dx.\lim_{N\to\infty}\frac1N\#\left\{n\le N:u\le\frac{E_d(N,\{a^n\})-A_N}{B_N}<v\right\}=\frac{1}{\sqrt{2\pi}}\int_u^v e^{-x^2/2}\,dx.

This conjecture applies Beck's heuristic for interval discrepancies to Benford errors. It predicts Gaussian fluctuations whenever the continued-fraction coefficients satisfy the stated Lindeberg-type condition and the error is not bounded; the source does not provide a resolution.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.