Random walks and quadratic number fields
Bence Borda
TL;DR
The work uncovers a novel bridge between random walks on the circle and the arithmetic of real quadratic fields. By studying S_n α mod 1 with α a quadratic irrational, Borda shows that convergence to uniformity in the quadratic Wasserstein metric W_{T,2} is governed by deep invariants of the associated real quadratic field, including fundamental units and special zeta values of Z-modules. The paper develops a two-pronged approach: a general badly approximable regime giving precise log-N rate constants via Fourier-analytic sums, and a quadratic-irrational regime where these constants are expressed explicitly through Dedekind zeta values and Zagier-type formulas for all quadratic irrationals. This synthesis connects probabilistic limit theorems for circle rotations to classical theories of binary quadratic forms, Pell equations, and zeta-valued constants, and provides concrete computational methods for c1(α) and c2(α).
Abstract
We establish a novel type of connection between random walks and analytic number theory. Working with a random walk on the circle group $\mathbb{R}/\mathbb{Z}$ in which each step is a random integer multiple of a given quadratic irrational $α$, we show that the rate of convergence to uniformity in the quadratic Wasserstein metric (also known as the periodic $L^2$ discrepancy) is governed by deep arithmetic invariants of the ring of algebraic integers of the real quadratic field $\mathbb{Q}(α)$, such as fundamental units and special values of zeta functions.
