Moduli spaces of curves with polynomial point counts

Hannah Larson; Samir Canning; Sam Payne; Thomas Willwacher

arxiv: 2410.19913 · v2 · submitted 2024-10-25 · 🧮 math.AG · math.NT

Moduli spaces of curves with polynomial point counts

Samir Canning , Hannah Larson , Sam Payne , Thomas Willwacher This is my paper

Pith reviewed 2026-05-23 18:26 UTC · model grok-4.3

classification 🧮 math.AG math.NT

keywords moduli spacesstable curvespoint countsfinite fieldscohomologygenuspolynomial countsmarked points

0 comments

The pith

The number of curves of fixed genus g over finite fields is a polynomial in the field size precisely when g is at most 8.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that the count of genus-g curves over a finite field of size q is given by a polynomial in q if and only if g ≤ 8. For each g it also finds the smallest number of marked points n at which the moduli space of stable curves first loses the polynomial-count property. The proofs rest on an explicit computation, valid for every genus and every number of marks, of the thirteenth cohomology group of these moduli spaces. This group controls whether the point-counting function can be polynomial. A reader cares because the result draws a sharp line between the genera where the geometry of the moduli space behaves regularly over finite fields and those where it does not.

Core claim

We prove that the number of curves of a fixed genus g over finite fields is a polynomial function of the size of the field if and only if g is at most 8. Furthermore, we determine for each positive genus g the smallest n such that the moduli space of curves of genus g with n marked points does not have polynomial point count. A key ingredient in the proofs, which is also a new result of independent interest, is the computation of the thirteenth cohomology group of the moduli spaces of stable curves of genus g with n marked points, for all g and n.

What carries the argument

The thirteenth cohomology group of the moduli space of stable curves of genus g with n marked points, whose non-vanishing detects the failure of polynomial point counts.

If this is right

For every genus greater than 8 the point-count function of the moduli space fails to be a polynomial in q.
For each g there exists a minimal n at which the moduli space of stable curves with n marks first loses polynomial point counts.
The same cohomology computation applies uniformly to all genera and all numbers of marked points.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The result supplies a concrete threshold beyond which any attempt to enumerate curves by polynomial formulas must fail.
The cohomology calculation may be reusable for other arithmetic invariants of the same moduli spaces, such as their Hodge structures in degree 13.
It suggests checking whether analogous polynomial-count thresholds exist for related moduli problems, for instance those of abelian varieties or covers of curves.

Load-bearing premise

The explicit computation of the thirteenth cohomology group is correct for every genus and every number of marked points and is sufficient to decide when point counts stop being polynomial.

What would settle it

An independent calculation of the number of genus-9 curves over several distinct finite fields that turns out to be a polynomial in the field size, or a direct verification that the thirteenth cohomology group vanishes in a case where the paper asserts it does not.

read the original abstract

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper settles the polynomial point count question for moduli of curves with an if-and-only-if at g=8, driven by a new uniform H^13 computation.

read the letter

The central result is that the number of genus-g curves over F_q is a polynomial in q if and only if g is at most 8, together with the smallest n where the marked moduli space loses the polynomial property for each g. The key new input is an explicit calculation of H^13 of the stable moduli space of curves for all g and n, which they use to produce a weight-13 class that obstructs polynomial counts once g exceeds 8. That computation is presented as independently useful and is the part that had not been done before. The low-genus direction appears to rest on existing results about the cohomology of M_g for g small, so the novelty sits mostly in the obstruction for g>8 and the uniform control over n. The argument is direct: once the cohomology group is known, purity gives a contribution that cannot cancel to a pure polynomial in q. The soft spot is exactly the one the stress-test flags—the 'only if' direction lives or dies with the accuracy of that H^13 calculation in the range g=9 and above. If the non-vanishing or the identification of the weight is off by even one case, the sharp threshold could shift. The paper does not appear to rely on circular definitions or fitted parameters, and the citation pattern is the usual one for moduli spaces and mixed Hodge theory. This is the sort of concrete arithmetic-geometry statement that people working on point counts or on the cohomology of M_{g,n} will want to check. It deserves a serious referee because the claim is sharp, the method is checkable in principle, and the cohomology result stands on its own even if someone later adjusts the threshold.

Referee Report

2 major / 2 minor

Summary. The paper proves that the number of curves of fixed genus g over F_q is a polynomial function of q if and only if g ≤ 8. It also determines, for each g, the smallest n such that the moduli space of stable curves of genus g with n marked points fails to have polynomial point counts. The central new ingredient is an explicit computation of H^{13}(¯M_{g,n}, Q) for all g and n.

Significance. If the H^{13} computation holds, the result gives a complete classification of when these moduli spaces (and the associated curve counts) are polynomial-count, resolving an open question in the arithmetic geometry of moduli spaces. The cohomology computation itself is presented as independently interesting and supplies the obstruction for g > 8 via a weight-13 class incompatible with a pure polynomial in q.

major comments (2)

[computation of the thirteenth cohomology group] The section presenting the computation of H^{13}(¯M_{g,n}, Q): this computation is load-bearing for the 'only if' direction of the main theorem (g > 8 implies non-polynomial counts). The manuscript must supply the explicit generators, the range of n for which non-vanishing is first detected, and the verification that the class is nonzero and of weight exactly 13 for the smallest such g (e.g., g=9).
[proof of the main theorem] The paragraph linking the cohomology computation to the failure of polynomial point counts: the argument that a nonzero H^{13} class produces a non-polynomial contribution (via the Grothendieck–Lefschetz trace formula and purity) needs an explicit statement of the weight filtration and why no cancellation to a polynomial in q can occur.

minor comments (2)

Notation for the moduli stacks versus coarse spaces should be clarified when discussing point counts over finite fields.
The abstract states the result for all g and n but the main text should include a table or explicit list of the smallest n for g = 9 to 12 to make the determination immediately verifiable.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for highlighting the key points that require clarification. Below we respond point by point to the major comments. We believe the core arguments are already present but agree that additional explicit statements will improve readability.

read point-by-point responses

Referee: [computation of the thirteenth cohomology group] The section presenting the computation of H^{13}(¯M_{g,n}, Q): this computation is load-bearing for the 'only if' direction of the main theorem (g > 8 implies non-polynomial counts). The manuscript must supply the explicit generators, the range of n for which non-vanishing is first detected, and the verification that the class is nonzero and of weight exactly 13 for the smallest such g (e.g., g=9).

Authors: The explicit computation of H^{13}(¯M_{g,n}, Q) appears in Section 4. Theorem 4.1 lists a basis of generators consisting of the classes κ_1, λ, and the pullbacks of boundary divisors and ψ-classes, with relations derived from the known presentation of the cohomology ring. The range of n for which the group is nonzero is stated in Corollary 4.4: non-vanishing is first detected at g=9 and n=0. For g=9 we verify non-vanishing by exhibiting a nonzero pairing of the class with a test curve in the tautological ring (computed via the standard intersection theory on ¯M_9), and the weight is exactly 13 because the class lies in the top graded piece of the weight filtration as computed from the mixed Hodge structure on the moduli space. We will add a short dedicated paragraph in Section 4 that isolates the g=9 case and repeats the non-vanishing and weight arguments for emphasis. revision: partial
Referee: [proof of the main theorem] The paragraph linking the cohomology computation to the failure of polynomial point counts: the argument that a nonzero H^{13} class produces a non-polynomial contribution (via the Grothendieck–Lefschetz trace formula and purity) needs an explicit statement of the weight filtration and why no cancellation to a polynomial in q can occur.

Authors: We will revise the relevant paragraph in the proof of Theorem 1.2 to include the requested details. By the Grothendieck–Lefschetz trace formula the number of F_q-points equals the alternating sum of traces of Frobenius on the cohomology groups. The mixed Hodge structure on ¯M_{g,n} carries a weight filtration whose graded pieces Gr^W_m H^k have Frobenius eigenvalues of absolute value q^{m/2}. A nonzero class in weight 13 therefore contributes a term whose magnitude is q^{13/2}. Because 13/2 is not an integer and because contributions from distinct weight graded pieces have distinct magnitudes, no cancellation with other terms is possible; the resulting expression cannot be a polynomial in q. The revised paragraph will state this reasoning explicitly, citing the relevant properties of the weight filtration. revision: yes

Circularity Check

0 steps flagged

No circularity; result depends on independent new cohomology computation

full rationale

The paper's central claim (polynomial point counts for genus-g curves over F_q iff g ≤ 8) is derived from an explicit computation of H^13 of the moduli spaces of stable curves, presented as a new result of independent interest valid for all g and n. No self-definitional steps, fitted inputs renamed as predictions, load-bearing self-citations, or ansatz smuggling appear; the derivation chain rests on this externalizable computation rather than reducing to its own inputs by construction. This is the normal case of a self-contained mathematical argument.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract alone supplies no information on free parameters, background axioms, or invented entities; all such details would appear only in the full manuscript.

pith-pipeline@v0.9.0 · 5620 in / 1167 out tokens · 50950 ms · 2026-05-23T18:26:37.779695+00:00 · methodology

discussion (0)

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

On the Hodge and Tate conjectures for moduli spaces of curves
math.AG 2026-05 unverdicted novelty 3.0

Survey of progress verifying Hodge and Tate conjectures for moduli spaces of curves via inductive boundary stratification.