🔢 mathematics

Genus one correspondence between tropical and algebraic curves

This paper establishes a complete genus-1 correspondence between algebraic and tropical curves in toric varieties by proving that their enumerative counts agree through an algebro-geometric proof utilizing logarithmic deformation theory, thereby generalizing the celebrated genus-0 Nishinou–Siebert theorem.

Original authors: Alessio Cela, Sae Koyama

Published 2026-07-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Alessio Cela, Sae Koyama

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to count how many unique, looping paths a hiker can take through a vast, multi-dimensional mountain range. In the world of mathematics, this "mountain range" is a geometric shape called a toric variety, and the "hiker" is an elliptic curve (a fancy name for a doughnut-shaped loop).

For a long time, mathematicians had two ways to count these paths:

The Algebraic Way: Doing the hard, heavy lifting with complex equations and calculus. This is like trying to count every single step the hiker takes by analyzing the soil, the wind, and the hiker's heartbeat. It's accurate but incredibly difficult.
The Tropical Way: Simplifying the mountain range into a skeleton made of straight lines and vertices (like a wireframe model). This is called "tropical geometry." It's much easier to draw and count on paper, but for a long time, no one was sure if counting these wireframes actually gave the correct answer for the real, complex mountains.

The Big Problem
In the 1990s and 2000s, mathematicians figured out that for simple, non-looping paths (genus 0), the wireframe count matched the real count perfectly. But as soon as you added a loop (genus 1, like a doughnut), the wireframes started behaving strangely. Some wireframes looked valid but couldn't possibly correspond to a real mountain path. Others were valid but needed special "weights" to count correctly.

For over two decades, the question remained: Can we trust the wireframe count for looping paths in any dimension?

The Solution: The "Well-Spaced" Rule
Alessio Cela and Sae Koyama, the authors of this paper, say: "Yes, but only if the wireframes follow a specific rule called 'well-spacedness'."

Think of a wireframe loop like a rubber band stretched around a set of pegs.

If the rubber band is too loose or the pegs are in a weird arrangement, the band might snap or slide off. In math terms, these are "bad" wireframes that don't correspond to real curves.
The authors discovered that a wireframe is "real" (algebraic) if and only if it is "well-spaced." This means the distances between the parts of the loop are balanced in a very specific way, ensuring the rubber band can actually exist in the real world.

The Magic of Multiplicities
Even when a wireframe is "well-spaced," it doesn't always count as just "one" path. Sometimes, a single wireframe shape can represent many different real paths.

The authors created a new scoring system (multiplicities).

Imagine you are counting the wireframes.
If a wireframe is simple, it gets a score of 1.
If a wireframe has a complex knot or a specific geometric feature (like a vertex where too many lines meet), the authors provide a formula to calculate exactly how many real paths it represents.
They call this the loop multiplicity and the saturation index. It's like realizing that one sketch of a house actually represents 50 different real houses because of how the windows and doors are arranged.

The "Lifting" Trick
The paper solves a puzzle that mathematician David Speyer started in 2005. He knew the wireframes could exist, but he couldn't prove how many real paths they lifted to.

The authors used a technique called logarithmic deformation theory. You can think of this as a magical bridge.

They built a bridge that connects the simple, flat wireframe world to the complex, curved algebraic world.
They proved that if you take a "well-spaced" wireframe and try to "lift" it up to the real world, it will always land on a real curve.
Furthermore, they counted exactly how many times it lands.

The Result
The paper proves a Correspondence Theorem:

The number of real, doughnut-shaped paths in a complex mountain range is exactly equal to the number of "well-spaced" wireframe paths, provided you use their new scoring system to weight them.

Why This Matters (According to the Paper)

It's a Complete Solution: Before this, the answer was only known for 2D surfaces (like a flat sheet of paper). This paper solves it for any number of dimensions.
It's "Genuinely Enumerative": Unlike some previous methods that gave "virtual" counts (mathematical ghosts that help with calculations but aren't real counts), this method counts actual, existing curves.
It Fixes Old Mistakes: The authors show that previous attempts to count these loops (by Kerber and Markwig) used the wrong weights for certain complex shapes. Their new method corrects these errors and aligns with the known correct answers for simple cases.

In a Nutshell
The authors took a confusing, high-dimensional geometry problem and showed that you can solve it by drawing simple line graphs, as long as you check that the lines are "well-spaced" and apply a specific, clever formula to count the complex ones. They turned a mountain of difficult calculus into a manageable puzzle of counting and balancing wireframes.

Technical Summary: Genus One Correspondence Between Tropical and Algebraic Curves

Problem Statement
The paper addresses the fundamental problem of determining which tropical curves arise from algebraic curves and establishing a precise enumerative correspondence between them. While the genus 0 case is well-understood—where the balancing condition is the sole requirement for realizability, leading to the celebrated Nishinou–Siebert correspondence theorem—the situation in positive genus is significantly more subtle. In genus 1, the existence of an algebraic lift for a tropical curve is not guaranteed by combinatorial data alone; it requires the curve to satisfy a "well-spaced" condition. Previous work by Speyer provided sufficient conditions for realizability but noted that his framework was qualitative rather than quantitative, making it unsuitable for deriving exact combinatorial counts. Furthermore, while virtual Gromov–Witten invariants have been expressed as weighted counts of tropical curves (e.g., by Abramovich–Chen–Gross–Siebert), these weights are defined in terms of virtual invariants rather than explicit combinatorial data. This paper aims to resolve the enumerative problem for genus 1 curves in arbitrary-dimensional toric varieties, providing a complete, purely tropical formula for the count of algebraic elliptic curves.

Methodology
The authors employ a purely algebro-geometric approach, diverging from the degeneration techniques used in prior genus 0 correspondence proofs. The methodology relies on the following key pillars:

Logarithmic Deformation Theory: The proof utilizes the moduli space $W_\Gamma(X)$ of well-spaced stable logarithmic maps from genus 1 curves to a smooth projective toric variety $X$ . This space, introduced by Ranganathan–Santos-Parker–Wise, parametrizes maps with prescribed contact orders along the toric boundary.
Minkowski Weights and Intersection Theory: Instead of degeneration, the authors apply the theory of Minkowski weights (following Fulton–Sturmfels) to compute intersection numbers on toric varieties. They establish a correspondence by analyzing the pullback of cohomology classes under the evaluation map $ev \times st: W_\Gamma(X) \to X^n \times M_{1,1}$ .
Lifting Criteria and Fine Logarithmic Maps: A central technical component is the development of a criterion (Theorem 5.4) for lifting a tropical curve to a logarithmic map. This involves analyzing the existence of meromorphic sections on the components of the domain curve that satisfy specific gluing conditions at the nodes. The authors distinguish between "fine" logarithmic maps (over fine, not necessarily saturated log schemes) and "fs" (fine and saturated) maps, computing the number of fs enhancements for a given tropical type.
Combinatorial Enumeration: The authors explicitly enumerate the number of algebraic lifts for each tropical curve type. This involves calculating the saturation index of the associated monoids, the lattice index of the tropicalization map, and the automorphism groups of the underlying maps.

Key Contributions and Results

The Correspondence Theorem (Theorem A): The main result establishes that the genuinely enumerative count of algebraic genus 1 curves in a toric variety $X$ passing through general subvarieties coincides exactly with the count of corresponding well-spaced tropical curves, weighted by explicit combinatorial multiplicities. This generalizes the Nishinou–Siebert theorem to genus 1 in arbitrary dimensions.
Explicit Multiplicities (Theorem D): The paper provides a complete formula for the weight $n_{Aut}(\sigma')$ $n_{A u t} (σ^{'})$ assigned to each maximal cone $\sigma'$ $σ^{'}$ of the moduli space of tropical well-spaced maps. This weight depends on:
- The automorphism group of the tropical curve.
- The saturation index of the cone (related to the number of fs logarithmic enhancements).
- The loop multiplicity (an index related to the cycle of the tropical curve).
- The number of interior lattice points of a specific polytope $P(\sigma')$ constructed from the direction vectors of the edges adjacent to the cycle.
- In low dimensions (curves in toric surfaces), these weights simplify to expressions involving the area of dual polygons or determinants of direction vectors.
Realizability Criterion (Theorem 5.4): The authors provide an effective criterion to determine if a tropical curve admits a logarithmic lift. This resolves the realizability problem by showing that a lift exists if and only if specific meromorphic functions exist on the components of the domain curve satisfying prescribed divisor and gluing conditions.
Integrality of Invariants (Proposition 1.2): The paper proves that for $n \ge 1$ free markings, the generic point of every irreducible component of the moduli space $W_\Gamma(X)$ has a trivial automorphism group, ensuring that the enumerative counts are integers.

Significance and Claims
The authors claim that this work provides the first complete, purely tropical enumeration of genus 1 curves in toric varieties of arbitrary dimension. Unlike previous virtual approaches where weights are defined via Gromov–Witten invariants, the weights derived here are purely combinatorial and computable. The paper asserts that this framework allows for the direct computation of enumerative invariants without recourse to virtual cycles.

The authors highlight that their results correct and extend previous ad-hoc multiplicity assignments (such as those by Kerber and Markwig) by deriving weights directly from the geometry of algebraic curves. They note that while their weights differ from previous proposals in certain cases, the final enumerative formulas for plane elliptic curves with fixed $j$ -invariant agree with known algebraic counts (e.g., those of Pandharipande).

The paper concludes by noting that the methodology opens the door to computing enumerative counts for elliptic curves in $\mathbb{P}^3$ and potentially extending the correspondence to higher genera, though the latter requires constructing higher-genus analogues of the moduli space $W_\Gamma(X)$ . The work is presented as a definitive resolution to the genus 1 enumerative problem in the toric setting, bridging the gap between tropical combinatorics and algebraic geometry through rigorous logarithmic deformation theory.

More like this