A Moebius inversion formula to discard tangled hyperbolic surfaces
Nalini Anantharaman, Laura Monk
TL;DR
This work introduces a Moebius inversion framework to sieve out tangles from random hyperbolic surfaces, enabling systematic conditioning on tangle-free geometries. Building on Friedman's ideas, it defines a big moduli space and a derived-tangle calculus that yields a unique, multiplicative Möbius function μ supported on tangled configurations, with the key identity 1 = sum μ(τ) over sub-surfaces. The authors then prove a polynomial bound for the number of local topological types of short geodesics in tangle-free surfaces, obviating the exponential proliferation seen in tangled cases. Together, these results provide a robust method to study spectral and geometric statistics under tangle-free conditioning and offer a general framework potentially applicable to other geometric pattern constraints in moduli spaces.
Abstract
Recent literature on Weil-Petersson random hyperbolic surfaces has met a consistent obstacle: the necessity to condition the model, prohibiting certain rare geometric patterns (which we call tangles), such as short closed geodesics or embedded surfaces of short boundary length. The main result of this article is a Moebius inversion formula, allowing to integrate the indicator function of the set of tangle-free surfaces in a systematic, tractable way. It is inspired by a key step of Friedman's celebrated proof of Alon's conjecture. We further prove that our tangle-free hypothesis significantly reduces the number of local topological types of short geodesics, replacing the exponential proliferation observed on tangled surfaces by a polynomial growth.
