[Paper Review] SL(2,R)-invariant probability measures on the moduli spaces of translation surfaces are regular
This paper proves that any $SL(2,\mathbb{R})$-invariant probability measure on the moduli space of normalized translation surfaces of genus $g$ is regular, by showing that the set of surfaces with two non-parallel saddle-connections of length $\leq \rho$ has measure $o(\rho^2)$ as $\rho \to 0$. The result resolves a key technical assumption in the Eskin-Kontsevich-Zorich formula for Lyapunov exponents.
In the moduli space $H_g$ of normalized translation surfaces of genus $g$, consider, for a small parameter $ρ>0$, those translation surfaces which have two non-parallel saddle-connections of length $\leq ρ$. We prove that this subset of $H_g$ has measure $o(ρ^2)$ w.r.t. any probability measure on $H_g$ which is invariant under the natural action of $SL(2,R)$. This implies that any such probability measure is regular, a property which is important in relation with the recent fundamental work of Eskin-Kontsevich-Zorich on the Lyapunov exponents of the KZ-cocycle.
Motivation & Objective
- To establish the regularity of $SL(2,\mathbb{R})$-invariant probability measures on the moduli space $\mathcal{H}_g$ of normalized translation surfaces of genus $g \geq 1$.
- To resolve a technical gap in the Eskin-Kontsevich-Zorich formula for sums of Lyapunov exponents, which requires the measure to be regular.
- To show that the set of translation surfaces with two non-parallel saddle-connections of length $\leq \rho$ has measure $o(\rho^2)$ under any such invariant measure.
- To provide a geometric and dynamical justification for the integration by parts argument used in the Eskin-Kontsevich-Zorich formula.
Proposed method
- Use the $SL(2,\mathbb{R})$ action to analyze the dynamics of saddle-connections via the decomposition $g_t R_\theta n_u$.
- Apply conditional measure techniques to study the behavior of measures along orbits of the diagonal subgroup $g_t = \mathrm{diag}(e^t, e^{-t})$.
- Estimate the norm of vectors under $SL(2,\mathbb{R})$ transformations using Euclidean norms and trigonometric identities.
- Construct a reference measure $m_0$ on a fundamental domain $X_0^*$ and relate it to the full invariant measure $m$ via scaling and integration.
- Use angular separation and minimal angle estimates between saddle-connections to control the contribution of singular configurations.
- Apply integration by parts and curvature-based arguments to relate geometric quantities to Siegel-Veech constants, relying on the regularity condition.
Experimental results
Research questions
- RQ1Does every $SL(2,\mathbb{R})$-invariant probability measure on $\mathcal{H}_g$ satisfy the regularity condition required for the Eskin-Kontsevich-Zorich formula?
- RQ2What is the asymptotic measure of the set of translation surfaces with two non-parallel saddle-connections of length $\leq \rho$?
- RQ3How does the dynamics of the $SL(2,\mathbb{R})$ action control the geometry of short saddle-connections?
- RQ4Can the integration by parts argument in the Eskin-Kontsevich-Zorich formula be justified without additional assumptions?
- RQ5What is the role of angular separation between short saddle-connections in the regularity of invariant measures?
Key findings
- The set of translation surfaces with two non-parallel saddle-connections of length $\leq \rho$ has measure $o(\rho^2)$ with respect to any $SL(2,\mathbb{R})$-invariant probability measure on $\mathcal{H}_g$.
- This implies that any such invariant measure is regular, as defined in the Eskin-Kontsevich-Zorich work.
- The regularity condition is satisfied because the measure of surfaces with short, non-parallel saddle-connections decays faster than $\rho^2$.
- The proof relies on controlling the norm of vectors under $g_t R_\theta$ actions and using angular separation to exclude problematic configurations.
- The result confirms that the integration by parts argument in the Eskin-Kontsevich-Zorich formula is valid under $SL(2,\mathbb{R})$-invariance alone.
- The key estimate is $m(Y(T,\frac{\pi}{2},X_0^*)) < \frac{1}{4}\eta \rho^2$ and similar bounds, which together imply the $o(\rho^2)$ decay.
Better researchstarts right now
From reading papers to final review, dramatically reduce your research time.
No credit card · Free plan available
This review was created by AI and reviewed by human editors.