REVIEW 1 major objections 4 minor 19 references
Bounding Radon numbers via Betti numbers
T0 review · 1 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Bounded Betti numbers force bounded Radon numbers in Euclidean space and on surfaces.
desk verdict Solid new Radon bound via Betti numbers; the optimal surface result leans on a companion theorem that a referee must check. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central mechanism is the constrained chain map: a nontrivial chain map from the simplicial chains of a complex $K$ to singular chains of the ambient space whose support over each simplex lies inside the relative convex hull of a prescribed set of points, together with a map $\Phi$ that preserves intersections and emptiness. Proposition 13 shows that a family with bounded topological complexity admits such a constrained map from any complex $K$ of dimension at most that level, via an induction that uses a Ramsey-type proposition to homogenize boundary homology classes. The Radon bounds then follow by combining this map with the non-embeddability of the $k$-skeleton of the $(2k+2)$-simplex in $\mathbb{R}^{2k}$ (for Euclidean space) or with the nonexistence of almost-embeddings of suitable graphs in a surface. For the sharp surface bound, an additional ingredient (Theorem 17) supplies a linear inequality relating the face numbers of nerves of open sets, which drives the bootstrapping that lowers the fractional Helly number.
What would settle it
Take a finite family of open sets on the $2$-sphere whose every intersection is either empty or path-connected ($TC_1=0$), with a constant fraction of intersecting triples but where no point belongs to more than $o(n)$ sets; Theorem 4's $b=0$ case forbids this, so such a construction would refute it. Likewise, an explicit family in $\mathbb{R}^d$ with $TC_{\lceil d/2\rceil}\le b$ but with Radon number exceeding any prescribed bound would refute Theorem 1.
Extended reading notes
Core claim
The central claim is Theorem 1: for each $b$ and $d$ there is a number $r(b,d)$ such that any finite family $\mathcal{F}$ in $\mathbb{R}^d$ with $TC_{\lceil d/2\rceil}(\mathcal{F}) \le b$ satisfies $r(\mathcal{F}) \le r(b,d)$. The paper also proves the surface analogue (Theorem 2): if all intersections of subfamilies have at most $b$ connected components, the Radon number is bounded by a function of $b$ and the surface alone. For open sets on a surface, Theorem 4 improves the fractional Helly number from the general bound to $2b+4$, and to $3$ when $b=0$; this is optimal and yields the conjectured $(p,q)$-theorem for such families.
Load-bearing premise
The optimal surface bound and the resolved conjecture depend on a linear inequality from a companion paper (Theorem 17) that is stated but not proved in this manuscript; if that inequality fails for open subsets of a surface with bounded first Betti number, the sharp fractional Helly bound and the conjecture resolution do not follow.
Editorial extensions
If this is right
- Every family in $\mathbb{R}^d$ whose first $\lceil d/2\rceil$ intersection Betti numbers are bounded by $b$ has Radon number bounded in terms of $b$ and $d$; consequently its Helly, Tverberg, colorful Helly, and fractional Helly numbers are also bounded.
- On any compact surface, bounding the number of connected components of all intersections bounds the Radon number, with no control needed on higher Betti numbers.
- For open sets on a surface, the fractional Helly number is at most $2b+4$ when intersection components are at most $b$, and at most $3$ for $b=0$; the latter is best possible.
- The fractional Helly results deliver weak $\varepsilon$-nets and a $(p,q)$-theorem for these families, including the previously conjectured $(p,q)$-theorem for open subsets of a surface.
Reading between the lines
- The companion inequality used for surfaces is quoted rather than proved here; if it generalizes, the same bootstrapping might give fractional Helly numbers for open sets in higher-dimensional manifolds, where the paper leaves bounded Radon numbers open.
- The author conjectures that the surface fractional Helly number is $3$ for every $b$; the linear bound $2b+4$ in Theorem 4 may be an artifact of the proof rather than the true threshold.
- The constrained chain-map method is formulated for any ambient space with a non-embeddability result, so it may yield Radon bounds beyond the listed spaces (e.g., for $\mathbb{Z}_2$-acyclic spaces or spaces with a fixed topological type), though no such extension is claimed in the paper.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper proves general topological bounds on Radon numbers. The main result (Theorem 1) states that if a finite family F of subsets of R^d has all reduced Betti numbers beta_i(∩G; Z2) for 0 ≤ i < ceil(d/2) bounded by b, then the Radon number r(F) is bounded by a function of b and d. Analogous statements are given for simplicial complexes, smooth manifolds, and surfaces (Theorem 2), where on a surface it suffices to bound the number of connected components of all intersections. Via the Holmsen–Lee theorem, these Radon bounds imply fractional Helly theorems (Theorem 3) and hence weak epsilon-nets and (p,q)-theorems. In Section 4, for open subsets of a surface with TC_1 ≤ b, the fractional Helly number is improved to 3 for b=0 and to 2b+4 for b≥1, with the b=0 case resolving a conjecture of Holmsen–Kim–Lee. The proof is built on a constrained-chain-map construction (Proposition 13), a Ramsey-type combinatorial lemma (Proposition 15), and known non-embeddability results.
Significance. The main Euclidean bound is a substantial and credible advance: it strengthens the earlier Helly-number bounds of [GPP+17] to Radon numbers, and the proof cleanly separates the topological non-embeddability statement from the combinatorial Ramsey argument. Proposition 13 is proved in detail within the manuscript, and the application to Theorem 1 uses only the cited non-embeddability of the k-skeleton of the (2k+2)-simplex. The optimal fractional Helly number 3 for b=0 on surfaces would be a sharp conjecture resolution. However, the surface-optimality part (Theorem 4 and the HKL conjecture) relies on Theorem 17, which is only cited as a weaker reformulation of [KP19, Theorem 4] and is not proved or fully specified in this manuscript. This dependency is the main weakness.
major comments (1)
- [Section 4, Theorem 17] The proof of Proposition 16, and therefore Theorems 4 and 6, depends entirely on Theorem 17, which is stated as a reformulation of [KP19, Theorem 4] and explicitly acknowledged to be weaker than that result. The manuscript does not prove Theorem 17, nor does it specify the constants c1 and c2, the exact ranges of k and b, or how the reformulation follows for open subsets of a surface with TC_1(A) ≤ b. Because the bootstrapping in Proposition 16 starts from k0 = 3 (b=0) or k0 = 2b+4 (b≥1), the claimed optimal fractional Helly number and the resolution of the Holmsen–Kim–Lee conjecture cannot be audited from this text. Please add a self-contained proof of Theorem 17 (or of the weaker reformulation actually used), or state Theorem 17 with complete hypotheses and a derivation from [KP19], so that the k-range and constants can be verified.
minor comments (4)
- [Abstract] The sentence 'for b=1 we get that the fractional Helly number is at most three' is inconsistent with Theorem 4, which gives k=3 for b=0 and k=2b+4 for b≥1; this should be corrected to b=0.
- [Section 3.4, proof of Proposition 13] The phrase 'Let the cardinality of F be large enough' appears to be a typo for 'Let the cardinality of P be large enough'; the role of P in the induction hypothesis should be stated explicitly, and the definition of rK(b) in terms of the chosen s and the induction constants should be spelled out.
- [Abstract and Section 2] The statement that TC_1(F) ≤ b bounds 'the number of connected components' should say that it bounds the reduced Betti number beta_0, i.e., the number of connected components is at most b+1; the current wording has an off-by-one discrepancy.
- [Section 4, proof of Proposition 16] In the inequality verification, the expression 'tk+1' should be written as 't^{k+1}' to denote exponentiation; as typeset, it could be misread as a product.
Circularity Check
No circularity found; the main Radon bound is self-contained and the surface bound relies on an independent companion theorem.
full rationale
The derivation chain does not reduce to its own inputs. Theorem 1 is proved in-manuscript via Proposition 13 (with the induction proof given in Section 3.4) together with the external non-embeddability result [GPP+17, Cor. 13] that the k-skeleton of the (2k+2)-simplex has no homological almost-embedding into R^{2k}; that cited theorem is a separate published result whose stated assumptions do not include the Radon bound. Theorem 2 similarly uses the independent almost-embedding obstruction from [GMP+17]. The fractional Helly consequences invoke Holmsen and Lee's theorem [HL19], again an external input rather than a rename of the paper's output. The surface-optimal Theorem 4 relies on Theorem 17, a reformulation of [KP19, Thm 4] from the author and Kalai's companion paper; while this is a self-citation and is load-bearing for the optimal surface bound, it is not circular: the cited inequality is a separate theorem with its own proof and assumptions and is not derived from the present paper or from the conjecture it settles. No parameter is fitted and no prediction is an identity; the paper even flags its reformulation as slightly weaker, which is an auditability concern rather than a circularity. Hence score 0.
Assumptions & free parameters
assumptions (6)
- standard math Standard singular and simplicial homology theory over Z2, including the standard cohomological proof behind Theorem 10.
- standard math Ramsey's theorem (Proposition 15 uses a strengthened Ramsey-type statement) and the Erdos-Simonovits supersaturation theorem.
- domain assumption Holmsen-Lee theorem: bounded Radon number implies bounded fractional Helly number.
- domain assumption Theorem 17, reformulated from [KP19, Theorem 4]: for open subsets of a surface with TC_1 <= b, if f_{k+1}=0 then f_k <= c1 f_{k-1} + c2.
- domain assumption For each compact surface S there is a finite graph that does not almost embed into S, from [GMP+17] and improved in [PT19] and [FK18].
- domain assumption F is a finite family of sets in X; for surface results the sets are open and the surface is compact two-dimensional.
Cite this review
Pith. "Pith review of Bounding Radon numbers via Betti numbers." pith.science (2026). https://pith.science/paper/K5EIJDND
@misc{pith2026190801677,
author = {Pith},
title = {Pith review of: Bounding Radon numbers via Betti numbers},
year = {2026},
howpublished = {\url{https://pith.science/paper/K5EIJDND}},
note = {Machine review of arXiv:1908.01677}
}
abstract
We prove general topological Radon-type theorems for sets in $\mathbb R^d$ or on a surface. Combined with a recent result of Holmsen and Lee, we also obtain fractional Helly theorem, and consequently the existence of weak $\varepsilon$-nets as well as a $(p,q)$-theorem for those sets. More precisely, given a family $\mathcal F$ of subsets of $\mathbb R^d$, we will measure the homological complexity of $\mathcal F$ by the supremum of the first $\lceil d/2\rceil$ reduced Betti numbers of $\bigcap \mathcal G$ over all nonempty $\mathcal G \subseteq \mathcal F$. We show that if $\mathcal F$ has homological complexity at most $b$, the Radon number of $\mathcal F$ is bounded in terms of $b$ and $d$. In case that $\mathcal F$ lives on a surface and the number of connected components of $\bigcap \mathcal G$ is at most $b$ for any nonempty $\mathcal G \subseteq \mathcal F$, then the Radon number of $\mathcal F$ is bounded by a function depending only on $b$ and the surface itself. For surfaces, if we moreover assume the sets in $\mathcal F$ are open, we show that the fractional Helly number of $\mathcal F$ is linear in $b$. The improvement is based on a recent result of the author and Kalai. Specifically, for $b=1$ we get that the fractional Helly number is at most three, which is optimal. This case further leads to solving a conjecture of Holmsen, Kim, and Lee about an existence of a $(p,q)$-theorem for open subsets of a surface.
Figures
Reference graph
Works this paper leans on
-
[1]
N. Alon, G. Kalai, J. Matou s ek, and R. Meshulam. Transversal numbers for hypergraphs arising in geometry. Adv. Appl. Math. , 29:79--101, 2002
work page 2002
-
[2]
D. C. Kay and E. W. Womble. Axiomatic convexity theory and relationships between the C arathéodory, H elly, and R adon numbers. Pacific Journal of Mathematics , 38, 08 1971
work page 1971
-
[3]
J. De Loera, X. Goaoc, F. Meunier, and N. Mustafa. The discrete yet ubiquitous theorems of C arath\'eodory, H elly, S perner, T ucker, and T verberg. Bulletin of the American Mathematical Society , 06 2017
work page 2017
-
[4]
P. Erd o s and M. Simonovits. Supersaturated graphs and hypergraphs. Combinatorica , 3(2):181--192, 1983
work page 1983
-
[5]
R. Fulek and J. Kyn c l. The Z_2 -genus of K uratowski minors. In 34th I nternational S ymposium on C omputational G eometry , volume 99 of LIPIcs. Leibniz Int. Proc. Inform. Schloss Dagstuhl. Leibniz-Zent. Inform., Wadern, 2018
work page 2018
- [6]
- [7]
-
[8]
A. Holmsen, M. Kim, and S. Lee. Nerves, minors, and piercing numbers. Trans. Amer. Math. Soc. , 371:8755–8779, 2019
work page 2019
Show all 19 references
-
[9]
Holmsen and D
A. Holmsen and D. Lee. Radon numbers and the fractional Helly theorem . http://arxiv.org/abs/1903.01068, 2019
1903 arXiv
-
[10]
R. E. Jamison-Waldner. Partition numbers for trees and ordered sets. Pacific J. Math. , 96(1):115--140, 1981
1981
-
[11]
Kalai and Z
G. Kalai and Z. Pat \'a kov \'a . Intersection patterns of planar sets . http://arxiv.org/abs/1907.00885, 2019
1907 arXiv
-
[12]
F. W. Levi. On H elly's theorem and the axioms of convexity. The Journal of the Indian Mathematical Society , 15(0):65--76, 1951
1951
-
[13]
Matou s ek
J. Matou s ek. Lectures on discrete geometry , volume 212 of Graduate Texts in Mathematics . Springer-Verlag, New York, 2002
2002
-
[14]
Pat \'a k and M
P. Pat \'a k and M. Tancer. Embeddings of k -complexes into 2k -manifolds . http://arxiv.org/abs/1904.02404, 2019
1904 arXiv
-
[15]
J. Radon. Mengen konvexer K \"o rper, die einen gemeinsamen P unkt enthalten. Mathematische Annalen , 83(1):113--115, Mar 1921
1921
-
[16]
F. P. Ramsey. On a problem in formal logic. Proc. London Math. Soc. , 30:264---286, 1929
1929
-
[17]
V. P. Soltan. Vvedenie v aksiomaticheskuyu teoriyu vypuklosti . ``Shtiintsa'', Kishinev, 1984. In Russian, with English and French summaries
1984
-
[18]
M. L. J. van de Vel. Theory of convex structures , volume 50 of North-Holland Mathematical Library . North-Holland Publishing Co., Amsterdam, 1993
1993
-
[19]
H. Whitney. The self-intersections of a smooth n -manifold in 2n -space. Annals of Mathematics , 45(2):220--246, 1944
1944
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.