Pith. sign in

REVIEW 3 major objections 3 minor 10 references

A Structural Condition on Point Sets with Few Distinct Dot Products

T0 review · 3 major / 3 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A point set with fewer than n^{3/4} distinct dot products must contain a dense line of about √n points.

desk verdict New structural claim at exponent 3/4, but the proof of Theorem 6.3 has a load-bearing gap in §6.2: a maximal well-spaced subset need not be small, so the main theorem is unproven. read the letter →

arxiv 2510.14585 v2 pith:LKUJIFNM submitted 2025-10-16 math.CO math.MG

Anshula Gandhi This is my paper · ORCID
classification math.COmath.MG MSC 52C10
keywords distinctdotproductspointconfigurationsdenselinewell-spacedpointscomplexproductarithmeticprogressionapproximationcombinatorialgeometryslowscaling
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 distinct dot products problem asks for the smallest possible number of distinct pairwise dot products among n points in the plane; known constructions give ~n, while the best proven lower bound is much smaller. This paper proves a structural fact about any configuration that is far from that bound: if |D(P_n)| = o(n^{3/4}), then for every b in (0,1) there is a subsequence in which each configuration contains a line through the origin with about √n points whose successive distances from the origin have ratios falling in (b,1). In other words, slow scaling forces a point-rich line that mimics an arithmetic progression. The contrapositive gives a clean sufficient condition: if every line with about √n points is 'well-spaced' — meaning consecutive ratios are bounded away from 1 — then the configuration determines at least n^{3/4} distinct dot products.

What carries the argument

The engine is the complex dot product p ⋆ q = |p||q| e^{i(arg p − arg q)}, whose real part is the ordinary dot product. Partitioning the real axis into 'buckets' between consecutive line-point radii, the paper shows that if a circle carries N points and a line through the origin carries M well-spaced points, the real projections of the complex dot products put ~N distinct values into each of ~M buckets, giving ≳ N M distinct dot products. With N = n^{1/4} and M = n^{1/2}, this forces ≳ n^{3/4} dot products from any configuration whose rich lines are all well-spaced. The density condition is the contrapositive of this bucket-counting mechanism.

What would settle it

Look for a set L of n^{1/2} points on a line with exactly one consecutive pair whose ratio lies in (b,1) and all other consecutive ratios below b; if its maximal well-spaced subset has size ~n^{1/2}, the proof step in §6.2 collapses. Directly, any sequence of n-point sets with |D(P_n)| = o(n^{3/4}) and no b-dense line of Ω(n^{1/2}) points for some fixed b would refute Theorem 6.3.

Watch

Extended reading notes

Core claim

The central claim is Theorem 6.3: a sequence of n-point sets with |D(P_n)| = o(n^{3/4}) must contain, for any b ∈ (0,1), a subsequence in which each set has a b-dense collinear set L of size Ω(n^{1/2}) — that is, roughly |L| consecutive pairs p,q along the line satisfy b < |p|/|q| < 1. The proof works by contrapositive. If every rich line (with Ω(n^{1/2}) points) is well-spaced, then combining a popular line and a popular circle through the complex-dot-product projection argument yields Ω(n^{3/4}) distinct dot products. Therefore a configuration with fewer than n^{3/4} dot products must contain a rich line that is not well-spaced, and the paper asserts that such a line must be b-dense in the

Load-bearing premise

The proof leans on the unproved claim that any rich line which is not well-spaced must have a maximal well-spaced subset of size o(n^{1/2}), leaving Ω(n^{1/2}) points that each create a bad consecutive pair; this does not follow from the line merely failing the well-spaced condition.

Editorial extensions

If this is right

  • Any point configuration with o(n^{3/4}) distinct dot products must contain a line of Ω(n^{1/2}) points whose consecutive ratios are arbitrarily close to 1 along a subsequence.
  • Any configuration in which every Ω(n^{1/2})-point line has consecutive ratios bounded away from 1 must determine Ω(n^{3/4}) distinct dot products.
  • The structural condition is a necessary condition for sub-n^{3/4} scaling and may combine with additive-combinatorial estimates to improve the lower bound.
  • The argument can be iterated: a slowly-scaling configuration contains about n^{1/2} distinct b-dense lines of n^{1/2} points each (a remark in the paper).
  • Each point in such a configuration forms ≫ n^{1/2} dot products with other points, a direct consequence of the dense popular line.

Reading between the lines

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

  • If the density condition is tight, the true minimum may be n^{3/4} rather than n; a construction with o(n^{3/4}) dot products would need to realize dense √n-lines while somehow suppressing dot product growth.
  • The b-dense condition for every b ∈ (0,1) hints that a diagonal argument over b → 1 could extract a line whose consecutive ratios converge to 1 at a quantitative rate, possibly yielding an arithmetic-progression-like substructure.
  • A natural test is the announced circle analogue: if a slowly-scaling set also forces a dense circle, the projection argument may yield ≳ n dot products and close the gap.
  • The proof's load-bearing step — the assertion that a non-well-spaced rich line has a maximal well-spaced subset of size o(n^{1/2}) — deserves to be isolated as a standalone combinatorial lemma; establishing or refuting it would settle whether Theorem 6.3 holds as stated.
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.

Referee Report

3 major / 3 minor

Summary. The paper studies the minimum number of distinct dot products determined by an n-point set in R^2. It claims a structural necessary condition for 'slow' scaling |D(P_n)| = o(n^{3/4}): for every b∈(0,1) there is a subsequence in which each configuration contains a line through the origin with ≳n^{1/2} points and ≳n^{1/2} consecutive pairs whose radius ratio lies in (b,1) (Theorem 6.3). The proof proceeds by establishing a popular line and a popular circle in any slowly-scaling configuration, proving a lower bound n^{3/4} when all rich lines are well-spaced (Theorem 6.1), and then attempting a contrapositive via maximal well-spaced subsets.

Significance. If established, the result would be a nontrivial structural constraint on point sets with few distinct dot products and could inform the conjectured lower bound |D(P_n)| ≳ n. The line–circle combination arguments and the complex-dot-product framework are workable ideas. However, the main theorem is not proved; the proof rests on a false combinatorial assertion in §6.2.

major comments (3)
  1. [§6.2, after Definition 6.4 (Theorem 6.3)] The proof asserts that because a rich line L fails W_b(L), there is a maximal well-spaced subset S⊂L of size o(n^{1/2}), leaving |L\S|≳n^{1/2} points each breaking well-spacedness. This assertion is false. Example: let m=n^{1/2} and let radii satisfy r_i/r_{i+1}<b for all i except one index j with r_j/r_{j+1}=b. Then L is not well-spaced. Removing r_j gives S with |S|=m-1; the new consecutive pair has ratio (r_{j-1}/r_j)(r_j/r_{j+1}) < b^2 < b (or, at an endpoint, the adjacent ratio is unchanged), so W_b(S) holds. Adding r_j back creates the bad pair, so S is maximal. Thus |S|=Θ(n^{1/2}) and |L\S|=1. The proof cannot produce Ω(n^{1/2}) bad consecutive pairs, so Theorem 6.3's conclusion is unsupported. The same flaw invalidates the iteration claim in the remark.
  2. [§6.1, proof of Theorem 6.1] The argument selects a wedge P_θ of angle cos^{-1}(b) with maximal point count and then asserts that P_θ still contains a popular circle with ≫n^{1/4} points and a popular line with ≫n^{1/2} points. This does not follow from the preceding averaging over wedges: the original popular line and circle could lie outside the chosen wedge. The claim can be repaired by applying Lemmas 3.6 and 4.6 to P_θ itself (since |D(P_θ)|≤|D(P_n)|≪n^{3/4} and |P_θ|∼n), but the proof must say so. As written, the step is a non sequitur.
  3. [Lemma 4.4] The proof of the lower bound on the number of supporting circles is invalid: it derives ≫n^{2-2α} circles from the existence of a popular line with ≫n^{2-2α} points. Points on a line through the origin need not determine distinct radii (opposite points at the same distance lie on the same circle), so the number of circles could be as small as half the number of points; more importantly, the presence of many points on one line says nothing about how many other circles exist. A lower bound on #circles must be argued separately. This lemma is not used in the main proof, but it is stated as part of the conclusion.
minor comments (3)
  1. [Definition 6.2 vs. proof of Theorem 6.3] b-dense is defined with ratios in the open interval (b,1), but the proof concludes ¬W_b(p,q), which only gives ratio ≥ b. Equality at b is not covered; the argument should use a perturbed b'<b or adapt the definition.
  2. [Lemmas 3.2 and 4.2] The proofs that |A·A|≥|A| for real sets and |C·C|≳|C/C| for unit complex sets are stated without justification. These are standard and easily proved (multiplication by a fixed nonzero element is injective; cos identifies at most two differences), but the paper should include the one-line argument.
  3. [Remark after Theorem 6.3] The claim that the argument can be iterated n^{1/2} times to obtain n^{1/2} different b-dense lines is not substantiated. The discard-and-repeat procedure needs a careful accounting of how the rich-line conditions are preserved and how the same circle/line combination continues to yield many dot products.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation does not assume its target and the 3/4 threshold arises from the paper's own inequality, not from a fit or self-citation.

full rationale

I walked the claimed derivation chain from the assumptions |D(P_n)| << n^{3/4} to Theorem 6.3. The argument first establishes a popular line and popular circle via averaging and imported external lemmas ([7, Lemma 3.1] is cited for the supporting-lines upper bound, and the ratio/product-set facts are standard external input, not self-citations). Theorem 6.1 then proves a sufficient condition: if every rich line is well-spaced, the configuration determines at least n^{3/4} dot products. Theorem 6.3 is the contrapositive: if |D(P_n)| = o(n^{3/4}), some rich line is not well-spaced, and the paper attempts to convert that failure into a b-dense rich line. The threshold 3/4 is not fitted: it emerges from the paper's own inequality n^{3-3α} < n^α in the footnote, so that is an independent derivation step. The proof's genuine weakness is in §6.2, where the paper asserts without derivation that a non-well-spaced rich line has a maximal well-spaced subset S of size o(n^{1/2}) with Ω(n^{1/2}) points in the complement, each producing a bad consecutive pair. This is a logical gap and a serious correctness concern; a single bad pair can coexist with a maximal well-spaced set of size |L|-1 = Θ(n^{1/2}). However, that is not circularity: the conclusion (existence of a b-dense rich line) is not assumed among the hypotheses, is not defined in terms of the conclusion, and no parameter is fitted to the target quantity. The paper also contains no load-bearing self-citation chain or imported uniqueness theorem that forces its choice. Under the standards of this review, a gap in the proof should be recorded as a correctness risk, not as circularity. The honest finding is therefore no significant circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 1 invented entities

The central claim rests on standard averaging and product-set lower bounds, plus one external lemma from [7]. No constants are fitted to data; the 3/4 exponent is derived from balancing popular line and circle sizes.

assumptions (5)
  • domain assumption External lemma [7, Lemma 3.1]: there exists p with |{p·q}| ≳ n^{1/2}|L_n|^{1/2}.
    Used in Lemma 3.5 to upper-bound the number of supporting lines; taken as a black box.
  • standard math Product sets of positive reals satisfy |A·A| ≥ |A|.
    Used in Lemmas 3.2 and 4.2; follows by fixing one factor.
  • standard math Rotation and uniform scaling preserve the number of distinct dot products.
    Used throughout for normalization; orthogonal transformations preserve dot products and scaling preserves distinctness.
  • standard math cos is strictly decreasing on [0,π] and injective on [0,π].
    Used to count distinct real projections of complex dot products in §5.
  • standard math Averaging: if M lines/circles contain n points, some line/circle has ≥ n/M points.
    Used in popular line and popular circle lemmas.
invented entities (1)
  • Complex dot product p⋆q
    purpose: Counting device: bookkeeps both magnitude and angle difference so that real projections can be counted bucket-wise in §5.
    Pure proof artifact (Def 5.2); carries no observable prediction outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Structural Condition on Point Sets with Few Distinct Dot Products." pith.science (2026). https://pith.science/paper/LKUJIFNM

@misc{pith2026251014585,
  author       = {Pith},
  title        = {Pith review of: A Structural Condition on Point Sets with Few Distinct Dot Products},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LKUJIFNM}},
  note         = {Machine review of arXiv:2510.14585}
}
abstract

The distinct dot products problem, a variant of the Erd\H{o}s distinct distances problem, asks "Given a set $P_n$ of $n$ points in $\mathbb{R}^2$, what is the minimum number $|D(P_n)|$ of distinct dot products they determine?" The best proven lower bound is $|D(P_n)| = \Omega(n^{2/3+7/1425})$, due to work by Hanson$\unicode{x2013}$Roche-Newton$\unicode{x2013}$Senger, and a recent improvement by Kokkinos. However, the best known construction determines $\Theta(n)$ dot products. We provide a structural condition that a point configuration $P_n$ would have to satisfy in order to have 'few' dot products, by which we mean that $|D(P_n)| < n^{\frac{3}{4}(1-\epsilon)}$ for some $\epsilon > 0$.

Figures

Figures reproduced from arXiv: 2510.14585 by the authors.

Figure 1
Figure 1. Any n points in geometric progression form ∼ n dot products. 3 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Any n points on a line form ≳ n dot products. Therefore, throughout the rest of this paper, while we refer to any finite set of points on a line as a line configuration, we refer to any finite set of points in geometric progression on a line as an optimal line configuration. 3.2. Bounding the number of supporting lines. We have just seen that if all points in a configuration lie along one line through the origin, we… view at source ↗
Figure 3
Figure 3. A set of points and their supporting lines. The next result shows that, indeed, in any point configuration with sublinearly-scaling dot products, the number of supporting lines in the configuration must grow with n. 4 [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: If the number of dot products scales ≪ n α , there exists some popular line with ≫ n 2−2α points.. 4. Circle configurations We now turn to proving bounds on the number of dot products we can expect when all points in a configuration are placed on one circle. Similarly …
Figure 5
Figure 5. Figure 5: Any n points equally spaced on a circle form only ∼ n dot products. We now show the above arrangement of points on a circle is optimal. Lemma 4.2 (Bounding dot products on a circle). If we have a set C ⊂ R 2 consisting of n points placed anywhere on a circle centered a…
Figure 6
Figure 6. Figure 6: Any n points on a circle form ≳ n dot products. 7 [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: A set of points and their supporting circles. The next result shows that in any point configuration with sublinearly-scaling dot prod￾ucts, the number of supporting circles in the configuration must grow with n. Lemma 4.4 (Lower bound on # of supporting circles). Let P…
Figure 8
Figure 8. Figure 8: If the number of dot products scales ≪ n α , there exists some popular circle with ≫ n 1−α points. 5. Line-and-circle configurations As we have shown, a slowly-scaling point configuration’s “popular circle” and “popular line” both have many points. It turns out that co…
Figure 9
Figure 9. Figure 9: A point configuration with points equally spaced on a circle, and points in geometric progression on a line. For convenience, we will assume r > 1, but an analogous argument holds for 0 < r < 1. It will suffice to ignore the dot products generated between any two point…
Figure 10
Figure 10. Figure 10: The set of complex dot products C ⋆ L. For each i ∈ [M], define the i th bucket Bi to be the line segment ari−1 < x < ari (except for B0 which will be the line segment 0 < x < ar0 ). To prove there are ≳ NM unique projections, it will suffice to prove that for each of…
Figure 11
Figure 11. Figure 11: The real line partitioned into buckets Bi . Furthermore, it will suffice to consider only the points in C in the top-right quadrant of the plane, of which there are still ∼ N. Let Ci ⊂ C ⋆ L be the subset of complex dot products C ⋆ L which lie on the circle with radi…
Figure 12
Figure 12. Figure 12: Only certain points in Ci project into Bi . That is, points in the set Ci with angle 0 < θ < cos−1 ( 1 r ) will have distinct real projections into the bucket Bi . Since points are equally distributed along the circle, every set Ci contains kN distinct points whose re…
Figure 13
Figure 13. Figure 13: The proportion of points in Ci projecting down to Bi is constant. So, there are ≳ N distinct real projections in each of the ∼ M buckets, and therefore ≳ NM dot products. Note that an analogous argument holds for 0 < r < 1, where the bucket Bi is defined to be the lin…
Figure 14
Figure 14. Figure 14: All points p ∈ Ci with angle 0 ≤ θ ≤ cos−1 (b) will project into Bi . Thus, every set Ci contains ≥ kN points whose real projection falls in bucket Bi , where k = cos−1 (b) 2π . Note that k does not depend on N, so that means that ∼ N points are projected from Ci onto…
Figure 15
Figure 15. Figure 15: All complex dot products C ⋆ L on the upper half-plane have distinct real projections. We then define sets Ci and buckets Bi as we did previously, and note that all ∼ N points in the “popular part” of set Ci have distinct real projections into the bucket Bi . So, ther…
Figure 16
Figure 16. Figure 16: Pθ contains points from P in one wedge of the plane. Let Pmax be the largest such set Pθ, i.e.: |Pmax| := max θ∈[0,2π) |Pθ|. Note that |Pmax| ∼ n, since by an averaging argument |Pmax| ≥ cos−1 b 2π n ∼ n. In this set of ∼ n points, we must still have a popular circle …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

10 extracted references · 4 linked inside Pith

  1. [1]

    Bounds on distinct and re- peated dot product trees

    Aaron Autry, Slade Gunter, Christopher Housholder, and Steven Senger. Bounds on distinct and re- peated dot product trees. InCombinatorial and Additive Number Theory, New York Number Theory Seminar, pages 21–37. Springer, 2022

    2022

  2. [2]

    Upper bounds on pairs of dot products.arXiv preprint arXiv:1502.01729, 2015

    Daniel Barker and Steven Senger. Upper bounds on pairs of dot products.arXiv preprint arXiv:1502.01729, 2015

    arXiv 2015

  3. [3]

    On the number of dot product chains in finite fields and rings

    Vincent Blevins, David Crosby, Ethan Lynch, and Steven Senger. On the number of dot product chains in finite fields and rings. InCombinatorial and Additive Number Theory, New York Number Theory Seminar, pages 1–20. Springer, 2021

    2021

  4. [4]

    Pinned dot product set estimates.arXiv preprint arXiv:2412.17985, 2024

    Paige Bright, Caleb Marshall, and Steven Senger. Pinned dot product set estimates.arXiv preprint arXiv:2412.17985, 2024

    arXiv 2024

  5. [5]

    Pairs of dot products in finite fields and rings

    David Covert and Steven Senger. Pairs of dot products in finite fields and rings. InCombinatorial and Additive Number Theory, New York Number Theory Seminar, pages 129–138. Springer, 2015

    2015

  6. [6]

    Dot product bounds in galois rings.arXiv preprint arXiv:2207.10051, 2022

    David Crosby. Dot product bounds in galois rings.arXiv preprint arXiv:2207.10051, 2022

    arXiv 2022

  7. [7]

    Convexity, superquadratic growth, and dot products.Journal of the London Mathematical Society, 107(5):1900–1923, 2023

    Brandon Hanson, Oliver Roche-Newton, and Steven Senger. Convexity, superquadratic growth, and dot products.Journal of the London Mathematical Society, 107(5):1900–1923, 2023

    1900

  8. [8]

    Dot product chains

    Shelby Kilmer, Caleb Marshall, and Steven Senger. Dot product chains. InCombinatorial Number Theory: Proceedings of the Integers Conference 2023, page 183. Walter de Gruyter GmbH & Co KG, 2024

    2023

Show all 10 references
  1. [9]

    An improved bound on the number of dot products determined by a finite point set in the plane.arXiv preprint arXiv:2502.12727, 2025

    Michalis Kokkinos. An improved bound on the number of dot products determined by a finite point set in the plane.arXiv preprint arXiv:2502.12727, 2025

    2025

  2. [10]

    Incidences and pairs of dot products.arXiv preprint arXiv:1509.01072, 2015

    Ben Lund. Incidences and pairs of dot products.arXiv preprint arXiv:1509.01072, 2015. Department of Pure Mathematics and Mathematical Statistics, Centre for Mathematical Sciences, University of Cambridge, Cambridge, UK Email address:ag2163@cam.ac.uk 18

    2015 arXiv

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.