🔢 mathematics

Tropical and algebraic elliptic plane curves with fixed j-invariant

This paper provides an alternative proof of Pandharipande's algebraic enumeration of elliptic plane curves with a fixed $j$ -invariant by tropically enumerating well-spaced elliptic curves and applying the genus 1 correspondence theorem of Cela and Koyama.

Original authors: Alessio Cela, Sae Koyama

Published 2026-08-14

📖 6 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 an architect trying to build a bridge. You have a specific blueprint (the shape of the bridge) and a specific number of pillars you must use. But there's a twist: you want to know exactly how many different ways you can build that bridge using only straight wooden planks, versus how many ways you can build it using flexible, wiggly vines. In the world of mathematics, this is the difference between "algebraic" curves (the rigid, precise blueprints) and "tropical" curves (the simplified, blocky versions made of straight lines).

Mathematicians have long been fascinated by a special type of curve called an "elliptic curve." Think of these as loops, like a rubber band stretched out on a piece of paper. A key feature of these loops is something called a " $j$ -invariant." You can think of the $j$ -invariant as the curve's "fingerprint" or its unique ID card; it tells you the specific shape and size of the loop, regardless of how you rotate or stretch it on the page. For decades, mathematicians have tried to count how many of these loops can pass through a specific set of scattered points on a plane. They found a formula for the rigid, algebraic loops, but when they tried to count the simplified, tropical versions, the numbers didn't quite line up with the original formula. It was like trying to count the bricks in a wall by looking at a shadow, only to find the shadow's shape didn't match the wall's blueprint.

This paper, written by Alessio Cela and Sae Koyama, steps in to fix that mismatch. The authors take the existing "shadow" counting method and refine it, using a new set of rules called "well-spaced" conditions. They prove that when you count these tropical loops correctly—paying close attention to how the loops are spaced out and weighted—their count matches the famous formula for the rigid algebraic loops perfectly. They didn't just guess this; they provided a rigorous, step-by-step proof that works for any number of points and any specific loop shape (as long as it's not a few very special, rare shapes). By doing this, they confirmed that the tropical world is a perfect mirror of the algebraic world for these specific curves, finally closing the gap between the two ways of counting.

The Story of the Loop and the Shadow

Let's dive into the adventure. Imagine you are in a vast, flat field (mathematicians call this the "projective plane"). You drop $3d - 1$ stones on the ground. Your goal is to draw a loop (an elliptic curve) that passes through every single one of these stones. But there's a catch: the loop must have a specific "fingerprint," known as the $j$ -invariant.

For a long time, a mathematician named Pandharipande figured out exactly how many of these loops exist using complex algebra. He found a neat formula: if you take the number of stones and do a little math with it, you get the answer. But there was a problem. When other mathematicians tried to solve the same puzzle using "tropical geometry"—a method that turns smooth curves into jagged, blocky shapes made of straight lines—they got a different answer. It was as if the tropical method was counting ghosts that didn't actually exist, or missing real ones.

The trouble was in the "weights." In tropical geometry, not all loops are created equal. Some loops are more "important" than others, and you have to count them multiple times to get the right total. Previous attempts used a set of weights that were a bit like a guess; they worked to make the numbers look nice, but they didn't actually reflect the real, physical loops they were supposed to represent. The authors of this paper realized that these weights were the culprit. They were "ad hoc," meaning they were made up just to fit the puzzle, rather than being derived from the actual rules of the game.

The "Well-Spaced" Solution

To fix this, the authors turned to a newer, stricter set of rules called "well-spaced" curves. Imagine you are arranging a group of friends around a campfire. If they are too crowded on one side and empty on the other, the fire burns unevenly. A "well-spaced" arrangement means everyone is balanced, with no one too close to the fire and no one too far away. In the math world, this "well-spaced" condition ensures that the tropical loops are balanced in a way that perfectly mimics the real algebraic loops.

The paper introduces a new way to count these balanced loops. They looked at two extreme scenarios:

The Giant Loop: Imagine the loop is stretched out so far that it looks like a long, thin string.
The Tiny Loop: Imagine the loop is shrunk down to almost nothing.

By studying these two extremes, the authors could see how the loops behave. They discovered that when you apply the "well-spaced" rules and the correct weights (which they calculated based on the geometry of the loop's core), the tropical count suddenly snaps into place. It matches Pandharipande's algebraic formula exactly.

What They Found (and What They Didn't)

The main finding is a resounding "Yes." The paper proves that the number of tropical elliptic curves passing through $3d - 1$ points with a fixed $j$ -invariant is exactly equal to the number of algebraic curves. The formula is:
$E_{d,j} = \binom{d-1}{2} N_d$
(Where $N_d$ is a famous number calculated by Kontsevich for simpler, non-looping curves).

However, the paper is very careful about what it doesn't do. It explicitly rules out the old method used by Kerber and Markwig. While their method gave the right final number, the authors show that the way they got there was flawed. The old method used weights that didn't correspond to real algebraic curves. The new method uses weights that do correspond to real curves. The paper also clarifies that this result is a mathematical proof, not just a simulation or a guess. It holds true for general points and general $j$ -invariants.

There is one small caveat: the paper mentions that for two very specific, rare $j$ -invariants (0 and 1728), the count needs a tiny adjustment (dividing by 3 or 2) because those specific loops have extra symmetries, like a snowflake that looks the same from many angles. But for almost every other case, the formula works perfectly.

Why This Matters

Why should a curious teenager care about counting loops on a plane? Because this paper bridges two different languages of mathematics. It shows that the simplified, blocky world of tropical geometry isn't just a rough sketch; it's a precise, reliable map of the complex, smooth world of algebraic geometry. By fixing the counting rules, the authors have given mathematicians a new, powerful tool. They can now solve hard problems about complex shapes by turning them into simpler, blocky puzzles, confident that the answer they get is the real, true answer. It's like discovering that if you build a model out of LEGO bricks, you can predict exactly how many bricks are in the real castle, as long as you follow the right building instructions.

Technical Summary: Tropical and Algebraic Elliptic Plane Curves with Fixed j-Invariant

Problem Statement
The paper addresses the enumeration of degree $d$ projective plane curves passing through $3d-1$ general points with a fixed $j$ -invariant. While Pandharipande [Pan97] established an algebraic formula for this count ( $E_{d,j}$ ), previous tropical attempts to reproduce this result contained structural discrepancies. Specifically, Kerber and Markwig [KM09] constructed a weighted polyhedral complex that matched Pandharipande's formula, but their multiplicities were chosen ad hoc to ensure independence from the choice of $j$ and interpolation points, rather than reflecting the actual counts of algebraic curves lifting to those tropical types. Consequently, there was no direct structural link between the tropical and algebraic counts, nor a derivation of Pandharipande's formula from first principles within the tropical framework. Additionally, Len and Ranganathan [LR18] provided a correspondence for toric surfaces, but their cone complex differed from the moduli space of well-spaced curves, potentially leading to inaccuracies in specific curve counts.

Methodology
The authors employ the moduli space of well-spaced tropical curves introduced by Ranganathan, Santos-Parker, and Wise [RSPW19b] and formalized in the correspondence theorem of Cela–Koyama [CK26]. This framework equips the cone complex of tropical curves with geometric weights derived from the number of algebraic lifts, rather than ad hoc assignments.

The methodology proceeds as follows:

Algebraic Setup: The paper distinguishes between the Gromov-Witten virtual count ( $E^{vir}_{d,j}$ ) and the genuine enumerative count ( $E_{d,j}$ ). Using the splitting axiom of Gromov-Witten theory, the authors prove that $E^{vir}_{d,j} = d^2 N_d$ , whereas the enumerative count is the target of the tropical analysis.
Tropical Correspondence: The authors utilize the correspondence theorem from [CK26], which relates the algebraic count to a weighted sum over maximal cones in the moduli space of well-spaced tropical curves $W_{\Gamma_d}(\mathbb{R}^2)$ . The weight of a cone $\sigma$ is given by $m(\sigma) \cdot \det_{ev \times j}(\sigma)$ , where $m(\sigma)$ is the geometric multiplicity (automorphism-weighted number of lifts) and $\det_{ev \times j}$ is the lattice index of the evaluation map.
Limit Analysis: To compute the sum, the authors analyze the behavior of the enumeration in two limiting cases for the $j$ $j$ -invariant:
- Large $j$ : The cycle of the tropical curve contracts to a single bounded edge. The authors classify these curves by their deficiency (0 or 2) and relate them to rational tropical curves by removing the contracted edge. This case is used to establish the main theorem for general $j$ -invariants.
- Small $j$ : The cycle remains non-contracted but small. The authors classify curves by deficiency (0, 1, or 2) and analyze the combinatorial structures of the loops (e.g., 3-valent vs. 4-valent vertices). This analysis yields a distinct, interesting formula (Equation 2 in the paper) involving the sum of terms $(2\text{Area}(T)^2 - 1)/2$ over the Newton subdivision, which is shown to be equivalent to the result derived from the large $j$ case.
Combinatorial Reduction: In both limits, the authors demonstrate that the weighted sum over elliptic tropical curves reduces to a sum over rational tropical curves (genus 0). They utilize Pick's Theorem and properties of the Newton subdivision dual to the rational curves to show that the total weight equals $\binom{d-1}{2}$ times the Mikhalkin count $N_d$ .

Key Contributions and Results

Correction of Literature: The paper corrects an oversight in [KM09] and [LR18] by utilizing the geometrically grounded multiplicities of [CK26]. It demonstrates that while the cone complexes in [KM09] and [CK26] differ (particularly regarding the presence of certain cones and their assigned weights), the final enumerative result remains consistent with Pandharipande's formula.
Theorem 1.1 (Main Result): For general points in $\mathbb{P}^2$ and a general $j$ -invariant, the tropical enumeration $E^{trop}_{d,j}$ equals the algebraic enumerative count $E_{d,j}$ :
$E_{d,j} = E^{trop}_{d,j} = \binom{d-1}{2} N_d$
where $N_d$ is the number of degree $d$ rational curves through $3d-1$ points. The proof relies primarily on the large $j$ limit analysis.
Explicit Formula: The paper provides an explicit combinatorial formula for the tropical count:
$E^{trop}_{d,j} = \sum_{C} \left( \sum_{T} \frac{2\text{Area}(T)^2 - 1}{2} \right) \text{mult}(C)$
where $C$ ranges over rational tropical curves, $T$ over triangles in the dual Newton subdivision, and $\text{mult}(C)$ is Mikhalkin's multiplicity.
Distinction of Counts: The paper explicitly proves (Theorem 1.2) that the Gromov-Witten count differs from the enumerative count, establishing $E^{vir}_{d,j} = d^2 N_d$ .

Significance
The paper claims significance primarily as an alternative proof of Pandharipande's algebraic enumeration formula using tropical geometry. By leveraging the recent correspondence theorem for well-spaced curves [CK26], the authors provide a structural justification for the formula that was previously missing. The work validates that the "ad hoc" multiplicities used in earlier tropical works [KM09] coincidentally yielded the correct count, but the underlying geometric weights from the moduli stack of well-spaced curves are necessary for a rigorous correspondence. The paper does not propose new applications or future implications beyond establishing this correspondence and correcting the theoretical framework for elliptic curve enumeration in toric varieties.

The Story of the Loop and the Shadow

The "Well-Spaced" Solution

What They Found (and What They Didn't)

Why This Matters

Technical Summary: Tropical and Algebraic Elliptic Plane Curves with Fixed j-Invariant

More like this