Back to search
2104.06716

On the distribution of Sudler products and Birkhoff sums for the irrational rotation

Bence Borda

correctmedium confidence
Category
math.DS
Journal tier
Strong Field
Processed
Sep 28, 2025, 12:56 AM

Audit review

The paper proves exactly the stated concentration inequality for the Sudler product at badly approximable irrationals (Theorem 1), by computing the first and second moments of log P_N(α) via an explicit Fourier/Fejér representation and then applying Chebyshev; the result holds uniformly for all M ≥ 1 and t > 0 and is presented with the natural √log(2N) fluctuation scale . The proof relies on the explicit series (14)–(15) for log P_N(α) and its averages, derived from the Fourier series of f(x)=log|2 sin(πx)|, together with quantitative Diophantine bounds to show that the variance is ≍ log M for badly approximable α; see Section 3.1 and Proposition 11 for the technical core, and Section 3.2 for the specialization to badly approximable α and the Chebyshev step . In contrast, the candidate solution’s “status: likely_open_as_of_cutoff” is incorrect. While the model’s outlined proof strategy (Fourier expansion → Fejér smoothing → dyadic block L2 bound ≍ R log R → Chebyshev) is broadly aligned with the paper’s approach, it is only a sketch and contains steps that require additional justification (e.g., the specific Fejér-kernel identity and the claimed blockwise near-orthogonality). The paper’s argument is correct and complete for the stated result; the model’s status and rigor are not.

Referee report (LaTeX)

\textbf{Recommendation:} minor revisions

\textbf{Journal Tier:} strong field

\textbf{Justification:}

The manuscript rigorously establishes concentration for Sudler products at badly approximable irrationals and develops general tools for Birkhoff sums of bounded-variation functions. The results are solid and of significant interest to specialists in Diophantine approximation and dynamical systems. The exposition is largely clear; minor clarifications would further improve accessibility.