English

Statistical convex-cocompactness for mapping class groups of non-orientable surfaces

Geometric Topology 2024-04-18 v1 Dynamical Systems

Abstract

We show that a finite volume deformation retract Tεt(Ng)/MCG(Ng)\mathcal{T}_{\varepsilon_t}^{-}(\mathcal{N}_g)/\mathrm{MCG}(\mathcal{N}_g) of the moduli space M(Ng)\mathcal{M}(\mathcal{N}_g) of non-orientable surfaces Ng\mathcal{N}_g behaves like the convex core of M(Ng)\mathcal{M}(\mathcal{N}_g), despite not even being quasi-convex. We then show that geodesics in the convex core leave compact regions with exponentially low probabilities, showing that the action of MCG(Ng)\mathrm{MCG}(\mathcal{N}_g) on Tεt(Ng)\mathcal{T}_{\varepsilon_t}^{-}(\mathcal{N}_g) is statistically convex-cocompact. Combined with results of Coulon and Yang, this shows that the growth rate of orbit points under the mapping class group action is purely exponential, pseudo-Anosov elements in mapping class groups of non-orientable surfaces are exponentially generic, and the action of mapping class group on the limit set in the horofunction boundary is ergodic with respect to the Patterson-Sullivan measure. A key step of our proof relies on complexity length, developed by Dowdall and Masur, which is an alternative notion of distance on Teichm\"uller space that accounts for geodesics that spend a considerable fraction of their time in the thin part.

Keywords

Cite

@article{arxiv.2404.11293,
  title  = {Statistical convex-cocompactness for mapping class groups of non-orientable surfaces},
  author = {Sayantan Khan},
  journal= {arXiv preprint arXiv:2404.11293},
  year   = {2024}
}

Comments

60 pages, 5 figures. Comments welcome!

R2 v1 2026-06-28T15:57:07.955Z