Moduli spaces of twisted maps to smooth pairs
Robert Crumplin
TL;DR
The paper develops a universal moduli space ${\mathrm{Orb}}_{\Lambda}(\mathcal{A}_r)$ for twisted maps to smooth pairs, encoding tangency data via line bundle–section data on twisted curves. It provides a tropical/combinatorial classification of irreducible components in terms of inducible essential mod $r$ tropical types, and constructs free semi-abelian actions that organize these components into torsors over boundary strata. It proves that degrees of comparison maps between rooting parameters are monomials in the extra parameter with exponent bounded by $\max(0,2g-1)$, and formulates a precise polynomiality framework through $\widehat{\mathbb{Z}}$-tropical types, canonical liftings, and stratified Chow theory. In genus 1 the virtual class matches the usual fundamental class on finite-type open substacks, recovering Tseng–You-type polynomiality phenomena and enabling a path to relate universal to geometric invariants for smooth pairs; the paper concludes with concrete geometric examples, including Maulik’s setup, illustrating the machinery and its implications for higher-genus comparisons and potential split-virtuallity structures.
Abstract
We study moduli spaces of twisted maps to a smooth pair in arbitrary genus, and give geometric explanations for previously known comparisons between orbifold and logarithmic Gromov--Witten invariants. Namely, we study the space of twisted maps to the universal target and classify its irreducible components in terms of combinatorial/tropical information. We also introduce natural morphisms between these moduli spaces for different rooting parameters and compute their degree on various strata. Combining this with additional hypotheses on the discrete data, we show these degrees are monomial of degree between $0$ and $\max(0,2g-1)$ in the rooting parameter. We discuss the virtual theory of the moduli spaces, and relate our polynomiality results to work of Tseng and You on the higher genus orbifold Gromov--Witten invariants of smooth pairs, recovering their results in genus $1$. We discuss what is needed to deduce arbitrary genus comparison results using the previous sections. We conclude with some geometric examples, starting by re-framing the original genus $1$ example of Maulik in this new formalism.
