Beating Product Constructions for Linear Equations Over Finite Fields

Fred Tyrrell; Paul Hametner

arxiv: 2606.12194 · v2 · pith:GUIGDHYVnew · submitted 2026-06-10 · 🧮 math.CO · math.NT

Beating Product Constructions for Linear Equations Over Finite Fields

Paul Hametner , Fred Tyrrell This is my paper

Pith reviewed 2026-06-27 09:08 UTC · model grok-4.3

classification 🧮 math.CO math.NT

keywords cap setslinear equations over finite fieldsproduct constructionssolution-free setsgenus onedensity improvementtranslation-invariant equations

0 comments

The pith

Any set in a finite field vector space that avoids nontrivial solutions to a genus-one translation-invariant linear equation can be replaced by a strictly denser avoiding set in higher dimension.

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

The paper proves that if a set A inside F_q^n has no nontrivial solutions to such an equation, then there exists a set B inside some F_q^m (m larger than n) that also has no nontrivial solutions and satisfies a strictly larger value of |B| to the power 1/m. A sympathetic reader would care because this rules out the possibility that the highest densities are always obtained simply by taking Cartesian products of one fixed finite set. The argument applies in particular to the cap set problem, showing that no single cap set in F_3^n can be optimal through products alone.

Core claim

For any A ⊆ F_q^n lacking non-trivial solutions to a translation-invariant linear equation of genus one, there is a set B ⊆ F_q^m in some higher dimension which also lacks non-trivial solutions, such that |B|^{1/m} > |A|^{1/n}.

What carries the argument

Existence of a non-product lift from A to a higher-dimensional B that preserves the absence of nontrivial solutions while strictly increasing the normalized density.

If this is right

No fixed cap set in F_3^n yields an asymptotically optimal lower bound on cap set density via direct products.
The maximal density of solution-free sets for these equations is not attained by any finite product construction.
Densities can be strictly increased by moving to higher dimensions while keeping the same forbidden equation.
The same improvement holds for every finite field F_q and every qualifying equation.

Where Pith is reading between the lines

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

Iterating the improvement produces an infinite sequence of sets whose normalized densities are strictly increasing.
The result suggests that optimal constructions, if they exist, must involve genuinely new structure at each scale rather than repetition of a single pattern.
One could test the method on small explicit examples in F_3^3 or F_4^2 to see the size of the density gain.

Load-bearing premise

The equation must be translation-invariant and of genus one, so that no nonempty proper subset of the coefficients sums to zero.

What would settle it

An explicit finite set A in some F_q^n with no nontrivial solutions for which every finite extension B in higher dimension satisfies |B|^{1/m} ≤ |A|^{1/n}.

read the original abstract

We show that for any $A\subseteq \mathbb{F}_q^n$ lacking non-trivial solutions to a translation-invariant linear equation of genus one, meaning that no nonempty proper subset of the coefficients sums to $0$, there is a set $B\subseteq \mathbb{F}_q^m$ in some higher dimension which also lacks non-trivial solutions, such that \[|B|^{1/m}>|A|^{1/n}.\] In particular, this implies that no fixed cap set in $\mathbb{F}_3^n$ gives an asymptotically optimal lower bound by direct products alone.

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 proves that for any avoiding set on a translation-invariant genus-one equation, a strictly denser avoiding set exists in higher dimension, so products are never optimal.

read the letter

The main point is that product constructions from a fixed avoiding set are provably suboptimal for these problems. For any A in F_q^n with no nontrivial solutions to a translation-invariant genus-one equation, the authors construct B in some m > n with the same avoidance property but |B|^{1/m} > |A|^{1/n}. The cap set case in F_3^n follows immediately as a corollary.

What works is the clean general statement. It directly shows why taking products of one good example cannot give the best asymptotic lower bound, and the genus-one condition plus translation invariance are used to make the lift preserve avoidance while improving density. The argument is presented as a direct existence result without obvious circularity or fitted parameters.

The soft spot is that the full construction details are not visible in the abstract alone, so the precise mechanism for building B and verifying the density gain cannot be checked here. If the lift is only existential or requires large m, that would be worth clarifying, but the stress-test indicates the logic is internally consistent once the details are examined. No load-bearing gaps appear from what is stated.

This is for additive combinatorics researchers focused on cap sets, slice rank, or equation-avoidance problems. A reader trying to improve lower bounds would find the general suboptimality result useful. It deserves a serious referee because the claim is sharp, the hypotheses are explicit, and the implication for existing construction strategies is concrete.

Referee Report

0 major / 3 minor

Summary. The paper proves that for any set A ⊆ F_q^n without nontrivial solutions to a translation-invariant linear equation of genus one (no nonempty proper subset of coefficients sums to zero), there exists m > n and a set B ⊆ F_q^m without nontrivial solutions to the same equation such that |B|^{1/m} > |A|^{1/n}. This is applied in particular to show that no fixed cap set in F_3^n yields an asymptotically optimal lower bound via direct products.

Significance. If the result holds, it demonstrates that product constructions cannot achieve the supremal normalized density for solution-free sets to translation-invariant genus-one equations over finite fields. This is a general structural statement that rules out optimality of iterated products for a broad class of problems, including the cap set problem, and motivates the search for non-product constructions. The argument relies only on the stated hypotheses and does not introduce fitted parameters or external bounds.

minor comments (3)

The definition of 'genus one' is used throughout but would benefit from an explicit restatement in the introduction or §2 alongside the translation-invariance condition, to make the hypotheses self-contained for readers.
The construction of B from A is the central contribution; a short schematic diagram or pseudocode in §3 would clarify the dimension increase and the preservation of the solution-free property.
The final paragraph on cap sets could include a brief comparison to known explicit constructions (e.g., those achieving densities better than the product of the largest known base cap set) to contextualize the improvement.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript and for recommending minor revision. The report correctly identifies the main contribution: that for any solution-free set A to a translation-invariant genus-one equation, a strictly denser solution-free set B exists in some higher dimension. No specific major comments appear in the report.

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained

full rationale

The paper asserts an existence result: given any solution-free A to a translation-invariant genus-one equation, a higher-density solution-free B exists in higher dimension. This is presented as following directly from the equation properties (translation-invariance and genus one) plus the solution-free assumption on A, via an explicit lift or embedding construction. No equations reduce a claimed prediction to a fitted input by construction, no self-citation is invoked as a load-bearing uniqueness theorem, and no ansatz is smuggled in. The argument is a standard mathematical existence proof whose central claim does not collapse to its inputs; the cap-set consequence is a corollary, not a definitional restatement.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract only; the central claim rests on the stated properties of the linear equation and the avoidance condition, but no further free parameters, axioms, or invented entities are visible.

axioms (1)

domain assumption The linear equation is translation-invariant and of genus one (no nonempty proper subset of coefficients sums to zero).
Explicitly required in the abstract for the sets A and B.

pith-pipeline@v0.9.1-grok · 5612 in / 1255 out tokens · 31735 ms · 2026-06-27T09:08:43.636878+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

26 extracted references · 1 linked inside Pith

[1]

Discrete Analysis , year=

New lower bounds for cap sets , author=. Discrete Analysis , year=
[2]

Nature , volume=

Mathematical discoveries from program search with large language models , author=. Nature , volume=. 2024 , publisher=

2024
[3]

arXiv preprint arXiv:2508.07932 , year=

X -evolve: Solution space evolution powered by large language models , author=. arXiv preprint arXiv:2508.07932 , year=

arXiv
[4]

Designs, Codes and Cryptography , year=

Extensions of Generalized Product Caps , author=. Designs, Codes and Cryptography , year=
[5]

On large subsets of

Ellenberg, Jordan S and Gijswijt, Dion , journal=. On large subsets of. 2017 , publisher=

2017
[6]

Designs, Codes and Cryptography , volume=

Maximal three independent subsets of \ 0, 1, 2\ ^n , author=. Designs, Codes and Cryptography , volume=. 1994 , publisher=

1994
[7]

Bulletin of the American Mathematical Society , volume=

New applications of the polynomial method: the cap set conjecture and beyond , author=. Bulletin of the American Mathematical Society , volume=
[8]

arXiv preprint arXiv:2103.06481 , year=

Improved explicit upper bounds for the Cap Set Problem , author=. arXiv preprint arXiv:2103.06481 , year=

arXiv
[9]

arXiv preprint arXiv:2007.03528 , year=

Breaking the logarithmic barrier in Roth's theorem on arithmetic progressions , author=. arXiv preprint arXiv:2007.03528 , year=

arXiv 2007
[10]

arXiv preprint arXiv:1907.01449 , year=

Formalizing the solution to the cap set problem , author=. arXiv preprint arXiv:1907.01449 , year=

Pith/arXiv arXiv 1907
[11]

Annals of Mathematics , number =

Ernie Croot and Vsevolod Lev and P. Annals of Mathematics , number =
[12]

A symmetric formulation of the Croot-Lev-Pach-Ellenberg-Gijswijt capset bound , author=

A symmetric formulation of the Croot-Lev-Pach-Ellenberg-Gijswijt capset bound , url=. A symmetric formulation of the Croot-Lev-Pach-Ellenberg-Gijswijt capset bound , author=
[13]

New Bounds on cap sets , volume =

Bateman, Michael and Katz, Nets , year =. New Bounds on cap sets , volume =
[14]

Sul massimo ordine delle calotte in S4,3 , volume =

Pellegrino, Giuseppe , year =. Sul massimo ordine delle calotte in S4,3 , volume =
[15]

and Ferret, S

Edel, Y. and Ferret, S. and Landjev, I. and Storme, L. , title =. J. Comb. Theory Ser. A , pages =. 2002 , issue_date =

2002
[16]

Maximal caps in AG (6, 3) , volume =

Potechin, Aaron , year =. Maximal caps in AG (6, 3) , volume =
[17]

On Subsets of Finite Abelian Groups with No 3-Term Arithmetic Progressions , author=. J. Comb. Theory, Ser. A , year=
[18]

Armin Biere and Katalin Fazekas and Mathias Fleury and Maximillian Heisinger , title =. Proc. of
[19]

Yves Edel's webpage , urldate =

Yves Edel , url =. Yves Edel's webpage , urldate =
[20]

arXiv preprint arXiv:2406.12290 , year=

Improving behrend's construction: Sets without arithmetic progressions in integers and over finite fields , author=. arXiv preprint arXiv:2406.12290 , year=

arXiv
[21]

2023 , note =

Elsholtz, Christian and Lipnik, Gabriel and Siebenhofer, Melanie , title =. 2023 , note =

2023
[22]

Acta arithmetica , volume=

Solving a linear equation in a set of integers I , author=. Acta arithmetica , volume=
[23]

Discrete Mathematics , volume=

Solving linear equations in a vector space over a finite field , author=. Discrete Mathematics , volume=. 2021 , publisher=

2021
[24]

A tribute to Paul Erdos (A

Sum-free subsets , author=. A tribute to Paul Erdos (A. Baker, B. Bollob
[25]

arXiv preprint arXiv:2409.07216 , year=

Problems from BCC30 , author=. arXiv preprint arXiv:2409.07216 , year=

arXiv
[26]

2025 , type =

Pollak, Florian , title =. 2025 , type =

2025

[1] [1]

Discrete Analysis , year=

New lower bounds for cap sets , author=. Discrete Analysis , year=

[2] [2]

Nature , volume=

Mathematical discoveries from program search with large language models , author=. Nature , volume=. 2024 , publisher=

2024

[3] [3]

arXiv preprint arXiv:2508.07932 , year=

X -evolve: Solution space evolution powered by large language models , author=. arXiv preprint arXiv:2508.07932 , year=

arXiv

[4] [4]

Designs, Codes and Cryptography , year=

Extensions of Generalized Product Caps , author=. Designs, Codes and Cryptography , year=

[5] [5]

On large subsets of

Ellenberg, Jordan S and Gijswijt, Dion , journal=. On large subsets of. 2017 , publisher=

2017

[6] [6]

Designs, Codes and Cryptography , volume=

Maximal three independent subsets of \ 0, 1, 2\ ^n , author=. Designs, Codes and Cryptography , volume=. 1994 , publisher=

1994

[7] [7]

Bulletin of the American Mathematical Society , volume=

New applications of the polynomial method: the cap set conjecture and beyond , author=. Bulletin of the American Mathematical Society , volume=

[8] [8]

arXiv preprint arXiv:2103.06481 , year=

Improved explicit upper bounds for the Cap Set Problem , author=. arXiv preprint arXiv:2103.06481 , year=

arXiv

[9] [9]

arXiv preprint arXiv:2007.03528 , year=

Breaking the logarithmic barrier in Roth's theorem on arithmetic progressions , author=. arXiv preprint arXiv:2007.03528 , year=

arXiv 2007

[10] [10]

arXiv preprint arXiv:1907.01449 , year=

Formalizing the solution to the cap set problem , author=. arXiv preprint arXiv:1907.01449 , year=

Pith/arXiv arXiv 1907

[11] [11]

Annals of Mathematics , number =

Ernie Croot and Vsevolod Lev and P. Annals of Mathematics , number =

[12] [12]

A symmetric formulation of the Croot-Lev-Pach-Ellenberg-Gijswijt capset bound , author=

A symmetric formulation of the Croot-Lev-Pach-Ellenberg-Gijswijt capset bound , url=. A symmetric formulation of the Croot-Lev-Pach-Ellenberg-Gijswijt capset bound , author=

[13] [13]

New Bounds on cap sets , volume =

Bateman, Michael and Katz, Nets , year =. New Bounds on cap sets , volume =

[14] [14]

Sul massimo ordine delle calotte in S4,3 , volume =

Pellegrino, Giuseppe , year =. Sul massimo ordine delle calotte in S4,3 , volume =

[15] [15]

and Ferret, S

Edel, Y. and Ferret, S. and Landjev, I. and Storme, L. , title =. J. Comb. Theory Ser. A , pages =. 2002 , issue_date =

2002

[16] [16]

Maximal caps in AG (6, 3) , volume =

Potechin, Aaron , year =. Maximal caps in AG (6, 3) , volume =

[17] [17]

On Subsets of Finite Abelian Groups with No 3-Term Arithmetic Progressions , author=. J. Comb. Theory, Ser. A , year=

[18] [18]

Armin Biere and Katalin Fazekas and Mathias Fleury and Maximillian Heisinger , title =. Proc. of

[19] [19]

Yves Edel's webpage , urldate =

Yves Edel , url =. Yves Edel's webpage , urldate =

[20] [20]

arXiv preprint arXiv:2406.12290 , year=

Improving behrend's construction: Sets without arithmetic progressions in integers and over finite fields , author=. arXiv preprint arXiv:2406.12290 , year=

arXiv

[21] [21]

2023 , note =

Elsholtz, Christian and Lipnik, Gabriel and Siebenhofer, Melanie , title =. 2023 , note =

2023

[22] [22]

Acta arithmetica , volume=

Solving a linear equation in a set of integers I , author=. Acta arithmetica , volume=

[23] [23]

Discrete Mathematics , volume=

Solving linear equations in a vector space over a finite field , author=. Discrete Mathematics , volume=. 2021 , publisher=

2021

[24] [24]

A tribute to Paul Erdos (A

Sum-free subsets , author=. A tribute to Paul Erdos (A. Baker, B. Bollob

[25] [25]

arXiv preprint arXiv:2409.07216 , year=

Problems from BCC30 , author=. arXiv preprint arXiv:2409.07216 , year=

arXiv

[26] [26]

2025 , type =

Pollak, Florian , title =. 2025 , type =

2025