A linear lower bound on the Ulrich complexity of hypersurfaces

Angelo Felice Lopez; Debaditya Raychaudhury; Yuta Takahashi

REVIEW 1 major objections 6 minor 22 references

A linear lower bound on the Ulrich complexity of hypersurfaces

T0 review · 1 major / 6 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper establishes that the Ulrich complexity of any smooth hypersurface of dimension n grows at least linearly in n, with explicit rank bounds that for large dimensions read n−1 (odd n) and n−2 (even n).

desk verdict Genuinely new linear lower bound for Ulrich complexity; the core argument is convincing, but the very-general even-dimensional case rests on an unpublished Hodge lemma from the authors' own preprint. read the letter →

arxiv 2607.03944 v2 pith:THVA4ADP submitted 2026-07-04 math.AG

Angelo Felice Lopez , Debaditya Raychaudhury , Yuta Takahashi This is my paper · ORCID

classification math.AG MSC 14J7014J6014F06

keywords UlrichbundlescomplexityhypersurfacesCherncharacterlowerboundalgebraicHodgeclassesNewtonidentitiespartitioncombinatorics

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proves the first linear lower bound on the Ulrich complexity of smooth hypersurfaces: the minimal possible rank of an Ulrich bundle on an n-dimensional smooth hypersurface grows at least as a constant multiple of n, and in large dimensions is n−1 (n odd) or n−2 (n even). The previously known bound was only on the order of sqrt(n). This matters because Ulrich bundles are a powerful tool in the study of projective varieties, and their minimal ranks—the Ulrich complexity—are expected to be large; linear lower bounds are a concrete step toward exponential conjectures. The proof combines a character computation via Grothendieck-Riemann-Roch with sharp estimates on a trigonometric power series and a combinatorial domination argument over integer partitions.

What carries the argument

The proof computes the Chern character of an Ulrich bundle on a hypersurface. Grothendieck-Riemann-Roch and the linear resolution of an Ulrich sheaf force the character to equal rd/(1+e^{-H}+...+e^{-(d-1)H}) (under a Hodge-structure hypothesis for even very general X). Twisting by the rational multiple −(d−1)H/2 yields a generating function f(z) whose Maclaurin coefficients b_j satisfy sharp two-sided estimates. Newton's identities convert these estimates into a recursion for the intersection numbers J_s = c_{2s}(F)H^{n−2s}. The combinatorial core is Claim 4.2: expanding J_m as a sum over partitions of m, the 'full' term A_m dominates the sum of all other terms in absolute value whenever r <

What would settle it

Exhibit a smooth hypersurface of dimension n≥6 carrying an Ulrich bundle of rank strictly below the claimed bound (e.g., rank ≤ n−3 for n≥44, or rank < F(m) in the small range), or find a very general even-dimensional hypersurface with an algebraic class in H^n not proportional to H^{n/2}. Either would refute the corresponding part of Theorem 1.

Watch

Extended reading notes

Core claim

The paper's central claim is Theorem 1: if X is a smooth hypersurface of dimension n≥6 and degree d≥3, then any Ulrich bundle on X has rank at least n−1 when n is odd and n≥43, at least n−2 when n is even and n≥44, and at least n when n is even, n≥44, and X is a sufficiently general (very general) member of its moduli space. For smaller dimensions, the lower bound is the explicit number F(m) = min{2m, B(m)} with m = ⌊n/2⌋ (or (n−2)/2), where B(m) is a closed-form rational expression in d and m. These are the first linear lower bounds on the Ulrich complexity of hypersurfaces, improving the prior general inequality uc(X) ≥ sqrt(n+2) − 1. In particular, for smooth cubic hypersurfaces the bound

Load-bearing premise

The proof of the stronger 'very general even' case assumes that for a very general even-dimensional smooth hypersurface, every algebraic cohomology class in middle degree is a rational multiple of H^{n/2}; this is imported from an unpublished companion preprint, and if it fails that part of the theorem collapses, though the odd-dimensional and non-very-general even bounds remain valid.

Editorial extensions

If this is right

For every smooth hypersurface of dimension n≥44, the Ulrich complexity is at least n−2, regardless of degree, so the minimal rank of an Ulrich bundle grows linearly with dimension.
For very general even-dimensional hypersurfaces, the lower bound is n for n≥44, two higher than the universal even bound n−2.
For smooth cubic hypersurfaces, the bounds are n−1 (odd n), n−2 (even n), and n (even very general); since Ulrich ranks on cubics are divisible by 3, the effective lower bound is the first multiple of 3 above these values.
For smaller dimensions the bound is the explicit function F(m) = min{2m, B(m)}, which for many low degrees d equals 2m (as tabulated), giving exact rank thresholds that depend only on degree and dimension.
The theorem provides the first linear lower bound, improving the previous general bound uc(X) ≥ sqrt(n+2) − 1.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the Hodge-class statement for very general even-dimensional hypersurfaces (that all algebraic classes in H^n are multiples of H^{n/2}) were proved without genericity assumptions, the stronger n-bound would hold for all even-dimensional hypersurfaces; the current restriction argument yields only n−2 for non-very-general X.
The same partition-domination technique could be exported to other settings where the Chern character of a special bundle is governed by a product formula, e.g., Ulrich bundles on complete intersections; the required input is only a two-sided estimate like the one in Lemma 2.3.
The numerical cutoffs (n≥43/44, m≥21 for F(m)=2m) are artifacts of the chosen estimates; tightening the coefficient bounds in Lemma 2.3 could lower these cutoffs and yield exact 2m thresholds for smaller dimensions.
The theorem does not settle the exponential-rank conjectures for hypersurfaces, but shows that any such lower bound must kick in above a linear growth rate; a natural next test would be to search for Ulrich bundles of rank between n and 2^{n/2}.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

First thing you should know: this paper delivers what it says. It proves the first lower bound on Ulrich complexity of smooth hypersurfaces that is linear in the dimension, improving the known sqrt(n+2)-1 from [BES] to roughly n. The main strategy is clean: use the Ulrich resolution to compute the Chern character of the twisted bundle, then show that a rank r < F(m) would force certain intersection numbers to have alternating signs while also vanishing for rank reasons. The analytic ingredient, Lemma 2.3, bounding the coefficients of the even function d/(e^{uz}+...+e^{-uz}), is a nice piece of elementary analysis, and the partition-sum domination in Claim 4.2 is explicit. I checked the key steps: the recurrence (4.6), the expansion (4.8), and the geometric-series bound (4.19) all hold. For odd-dimensional hypersurfaces and for non-very-general even-dimensional ones, the proof is self-contained and convincing.

The soft spot is exactly where the stress-test note puts it. The very-general even case of Theorem 1 (uc(X) >= n for n>=44, and F(n/2) for n<=42) depends on Lemma 4.1's assertion that every algebraic class in H^n(X;Q) on a very general even-dimensional hypersurface is a multiple of H^{n/2}. That fact is imported from [LR2, Lemma 4.1], an unpublished preprint by two of the three authors, and the paper neither states the lemma nor sketches its proof. If that lemma is true, the argument goes through; if it has extra hypotheses or a gap, the very-general even bounds do not follow. The odd-dimensional and non-very-general even bounds do not rely on it. This is worth fixing before publication: either prove the Hodge statement or at least state it precisely with the hypotheses.

The novelty is real: this is not a reparametrization of the [BES] bound. The relation to the authors' related preprint [LR2] is mentioned in passing, but the novelty boundary between this paper and [LR2] should be made explicit. That is a presentation issue, not a mathematical one.

Bottom line: this paper deserves a serious referee. I would send it to someone who can check the Hodge-theoretic input and the combinatorial estimates. If the [LR2] lemma is supplied, it should be published as is.

Referee Report

1 major / 6 minor

Summary. The paper proves a lower bound on Ulrich complexity of smooth hypersurfaces X⊂P^{n+1} of dimension n≥6 and degree d≥3. For odd n the bound is n−1 for n≥43 and min{2m,B(m)} with m=(n−1)/2 for small n; for even n it is n−2 (respectively n for very general X) with analogous small-n values. The proof works with the Chern character of an Ulrich bundle, twists by −uH to turn the character into a series with coefficients b_j, and then uses a partition expansion to show that if the rank r is below the threshold F(m), the Chern class c_{2m} of the twisted bundle is nonzero, contradicting r<2m. The even non-very-general bound is obtained by restricting to a hyperplane section.

Significance. If the result holds, it is the first linear lower bound on Ulrich complexity of hypersurfaces, substantially improving the previous sqrt(n+2)−1 bound of [BES]. The core of Section 4 is an explicit, parameter-free contradiction argument: the coefficient estimates in Lemma 2.3, the partition expansion (4.8), and the domination estimate (4.19) are all checkable and appear sound. The odd-dimensional and non-very-general even bounds are essentially self-contained. The stronger very-general even bound, however, depends on a Hodge-theoretic input imported from an unpublished preprint of two of the authors, so the paper as it stands is only conditional for that portion of the main theorem.

major comments (1)

[§4, Lemma 4.1; Theorem 1 (even very-general case)] The very-general even-dimensional bounds uc(X)≥n (n≥44) and uc(X)≥F(n/2) (n≤42) require the assertion that for a very general even-dimensional hypersurface every algebraic class in H^n(X;Q) is a multiple of H^{n/2}. This is imported from [LR2, Lemma 4.1], an unpublished preprint by two of the three authors, and the lemma is neither stated nor proved in the present paper. If that lemma fails, or carries extra hypotheses on n or d, the very-general even part of Theorem 1 does not follow. Since this is load-bearing for part of the main claim, please include a complete statement and proof of the needed Hodge-theoretic fact, or reformulate Theorem 1 so that the very-general even bounds are explicitly conditional on it.

minor comments (6)

[Theorem 1 statement] The last displayed formula ends with "F(n/2). ." — a stray double period, and the entire theorem lacks a closing period.
[§4, proof of Theorem 1] The sentence "Set k=2sin (3.1)" should read "Set k=2s in (3.1)".
[Lemma 2.2 proof] In the displayed definition of G_N(z), the product notation "NY i≥1" should be \prod_{i=1}^N.
[§4, Lemma 4.1] The statement that i_*: H^{2k}(P^{n+1};Q)→H^{2k}(X;Q) is an isomorphism for all k≠n/2 is not immediate for k>n/2. It follows from hard Lefschetz, but a one-sentence justification (or a precise reference) would be useful.
[Introduction / Theorem 1] The term "very general" is used but never defined. Please specify the precise countable union of proper loci in the moduli of degree-d hypersurfaces that is excluded.
[References] Ref. [LR2] is cited for a fact that is load-bearing for part of the theorem. If the preprint is not yet published, its statement should be reproduced in the paper or an appendix should give the proof; at minimum, update the reference if it has appeared.

Circularity Check

1 steps flagged · score 4.0 of 10

Very-general even-dimensional bound rests on unpublished [LR2] Hodge lemma from two of the authors; odd and non-very-general even bounds are self-contained.

self citation load bearing [Section 4, proof of Lemma 4.1 (and hence the very-general even case of Theorem 1)]
"while for even n and k= n/2, any algebraic class in H^n(X;Q) is of type aH^{n/2} (see for example [LR2, Lemma 4.1])"

The only justification for the Hodge-class input needed to identify ch(E) with i^*\beta in the very-general even case is a lemma from [LR2], an unpublished preprint by two of the three current authors. If that lemma is absent, incomplete, or false, Lemma 4.1 and the very-general even bounds uc(X) >= n / F(n/2) do not follow. The paper supplies no statement, proof, or independent verification, so the branch is load-bearing self-citation rather than a derivation from established common knowledge.

full rationale

No fitted-parameter-as-prediction or definitional circularity is present: F(m) and B(m) are thresholds chosen so that the combinatorial inequality in Claim 4.2 holds; the proof of the claim is a direct coefficient estimate, not an assumption of the theorem. For odd n, and for even n via hyperplane sections, the proof uses only the standard Ulrich resolution and the Chern-character calculation, which is self-contained. The only load-bearing reliance on the authors' own prior work is [LR2, Lemma 4.1] in Lemma 4.1 for very general even hypersurfaces; this does not reduce the main theorem to its conclusion, but it is an unpublished self-citation on which a substantial branch of Theorem 1 depends. Hence moderate score.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters or invented entities. The paper's own contribution is a fixed inequality; the threshold functions B(m) and F(m) are defined, not fitted. The main unproved input is the [LR2] Hodge class lemma for the very-general even case; all other inputs are standard.

assumptions (5)

domain assumption Work over the complex numbers; all varieties are smooth projective.
Preamble: 'We work over the complex numbers.' Needed for Hodge theory and intersection theory.
domain assumption An Ulrich bundle on a smooth hypersurface of degree d has a linear resolution 0→O_{P^{n+1}}(-1)^{⊕rd}→O_{P^{n+1}}^{⊕rd}→i_*E→0.
Used in Lemma 4.1; cited to [B, Prop. 2.1(ii)].
domain assumption For a very general even-dimensional hypersurface, every algebraic class in H^n(X;Q) is a multiple of H^{n/2}.
Used in Lemma 4.1 for the even very-general case; cited to [LR2, Lemma 4.1], an unpublished same-author preprint.
standard math Euler product sinh(z)=z∏(1+z^2/(π^2ℓ^2)) and convergence of infinite products (Ahlfors).
Lemma 2.3 coefficient bounds.
standard math Grothendieck–Riemann–Roch for closed embeddings and the splitting principle/Newton identities for Chern classes.
Lemma 3.4 and equation (4.1).

how reviews work

0 comments

Cite this review

Pith. "Pith review of A linear lower bound on the Ulrich complexity of hypersurfaces." pith.science (2026). https://pith.science/paper/THVA4ADP

@misc{pith2026260703944,
  author       = {Pith},
  title        = {Pith review of: A linear lower bound on the Ulrich complexity of hypersurfaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/THVA4ADP}},
  note         = {Machine review of arXiv:2607.03944}
}

read the original abstract

We give a lower bound on the Ulrich complexity of hypersurfaces in terms of their dimension.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 1 linked inside Pith

[1]

L. V. Ahlfors. Complex analysis. An introduction to the theory of analytic functions of one complex variable. Internat. Ser. Pure Appl. Math. McGraw-Hill Book Co., New York, 1978. xi+331 pp

1978
[2]

Beauville

A. Beauville. An introduction to Ulrich bundles. Eur. J. Math. 4 (2018), no. 1, 26-36

2018
[3]

R. O. Buchweitz, G. M. Greuel, F. O. Schreyer. Cohen-Macaulay modules on hypersurface singularities. II. Invent. Math. 88 (1987), no. 1, 165-182

1987
[4]

Bl\"aser, D

M. Bl\"aser, D. Eisenbud, F. O. Schreyer. Ulrich complexity. Differential Geom. Appl. 55 (2017), 128-145

2017
[5]

Ciliberto, F

C. Ciliberto, F. Flamini, A. L. Knutsen
[6]

Casanellas, R

M. Casanellas, R. Hartshorne
[7]

Costa, R

L. Costa, R. M. Mir\'o-Roig, J. Pons-Llopis. Ulrich bundles. De Gruyter Studies in Mathematics, 77, De Gruyter 2021

2021
[8]

D. Erman. Matrix factorizations of generic polynomials. Preprint arXiv:2112.08864

arXiv

Show all 22 references

[9]

Eisenbud, F.-O

D. Eisenbud, F.-O. Schreyer. Resultants and Chow forms via exterior syzygies. J. Amer. Math. Soc. 16 (2003), no. 3, 537-579

2003
[10]

W. Fulton. Intersection theory. Ergeb. Math. Grenzgeb. (3), 2. Springer-Verlag, Berlin, 1998, xiv+470 pp

1998
[11]

Faenzi, Y

D. Faenzi, Y. Kim. Ulrich bundles on cubic fourfolds. Comment. Math. Helv. 97 (2022), no. 4, 691-728

2022
[12]

Herzog, B

J. Herzog, B. Ulrich, J. Backelin. Linear maximal Cohen-Macaulay modules over strict complete intersections. J. Pure Appl. Algebra 71 (1991), no. 2-3, 187-202

1991
[13]

Kim, F.-O

Y. Kim, F.-O. Schreyer. An explicit matrix factorization of cubic hypersurfaces of small dimension. J. Pure Appl. Algebra 224 (2020), no. 8, 106346, 13 pp

2020
[14]

Lazarsfeld

R. Lazarsfeld. Positivity in algebraic geometry. II. Positivity for vector bundles, and multiplier ideals. Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics, 49. Springer-Verlag, Berlin, 2004

2004
[15]

A. F. Lopez, R. Mu\ n oz
[16]

A. F. Lopez, R. Mu\ n oz, J. C. Sierra
[17]

A. F. Lopez, D. Raychaudhury. Ulrich subvarieties and the non-existence of low rank Ulrich bundles on complete intersections. Trans. Amer. Math. Soc. 379 (2026), no.\ 8, 5809-5846

2026
[18]

A. F. Lopez, D. Raychaudhury. A lower bound on the Ulrich complexity of hypersurfaces. Preprint 2025, arXiv:2503.13396

2025
[19]

A. F. Lopez, J. C. Sierra
[20]

G. V. Ravindra, A. Tripathi. Rank 3 ACM bundles on general hypersurfaces in ^5 . Adv. Math. 355 (2019), 106780, 33 pp

2019
[21]

G. V. Ravindra, A. Tripathi. On the base case of a conjecture on ACM bundles over hypersurfaces. Geom. Dedicata 216 (2022), no. 5, Paper No. 49, 10 pp

2022
[22]

R. P. Stanley. Enumerative combinatorics. Volume 1.. Cambridge Stud. Adv. Math., 49. Cambridge University Press, Cambridge, 2012. xiv+626 pp

2012

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.

[1] [1]

L. V. Ahlfors. Complex analysis. An introduction to the theory of analytic functions of one complex variable. Internat. Ser. Pure Appl. Math. McGraw-Hill Book Co., New York, 1978. xi+331 pp

1978

[2] [2]

Beauville

A. Beauville. An introduction to Ulrich bundles. Eur. J. Math. 4 (2018), no. 1, 26-36

2018

[3] [3]

R. O. Buchweitz, G. M. Greuel, F. O. Schreyer. Cohen-Macaulay modules on hypersurface singularities. II. Invent. Math. 88 (1987), no. 1, 165-182

1987

[4] [4]

Bl\"aser, D

M. Bl\"aser, D. Eisenbud, F. O. Schreyer. Ulrich complexity. Differential Geom. Appl. 55 (2017), 128-145

2017

[5] [5]

Ciliberto, F

C. Ciliberto, F. Flamini, A. L. Knutsen

[6] [6]

Casanellas, R

M. Casanellas, R. Hartshorne

[7] [7]

Costa, R

L. Costa, R. M. Mir\'o-Roig, J. Pons-Llopis. Ulrich bundles. De Gruyter Studies in Mathematics, 77, De Gruyter 2021

2021

[8] [8]

D. Erman. Matrix factorizations of generic polynomials. Preprint arXiv:2112.08864

arXiv