REVIEW 2 major objections 4 minor 34 references
On the volume of non-central sections of a cube
T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Every off-center hyperplane slice of the unit cube has volume at least 1/17, in every dimension.
desk verdict Good paper, true-looking results, but a typo in Proposition 5.4 breaks the printed proof of Theorem 1.2; it is fixable and worth refereeing. 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 load-bearing machinery is a probabilistic representation of section volumes. With independent random vectors $U_j$ uniformly distributed on $S^2$ in the real case and on $S^3$ in the complex case, the volume of the hyperplane section is $A_{\mathbb R}(a,t)=\int_{|\sum a_jU_j|\ge t} dP/|\sum a_jU_j|$ and $A_{\mathbb C}(a,t)=\int_{|\sum a_jU_j|\ge t} dP/|\sum a_jU_j|^2$. The identity $|\sum a_jU_j|^2=1+2S$, with $S=\sum_{i<j}a_ia_j\langle U_i,U_j\rangle$, turns the threshold event into $S\ge0$; the paper controls $P(S\ge0)$ from below via a Laplace-transform estimate and duality in Orlicz spaces, and controls the upper tail of $|\sum a_jU_j|$ via a Khintchine-type inequality for rotationally invariant vectors. For general codimension $d$, the section volume is interpreted as the density of a projected uniform random point, and the classical lower bound for central sections together with the known upper bound for central sections complete the reduction.
What would settle it
Check Proposition 5.4 numerically for $k=3$, $t=1.1$: the printed bound is $1.1^3\exp(3/2-3/(2\cdot1.1^2))\approx1.73$, which is greater than 1 and therefore cannot be a probability tail; the integration in Section 6 needs a genuine exponential tail near $t=1$, so replacing the printed bound by any valid tail $\le C_k e^{-c_k t^2}$ would either restore the proof with adjusted constants or reveal a genuine obstruction.
Extended reading notes
Core claim
The central claim is that dimension does not enter any lower bound for non-central sections of the cube. Specifically, Theorem 1.2 asserts $A_{\mathbb R}(a,1)>1/17$ for every unit $a\in\mathbb{R}^n$ and $A_{\mathbb C}(a,1)>1/27$ for the polydisc in $\mathbb{C}^n$, while Theorem 1.1 asserts that for each $d$, all $(n-d)$-dimensional affine sections at distance at most $1/2$ have volume $\ge\varepsilon(d)>0$. The proof derives integral representations for these volumes over products of unit spheres and bounds the resulting probabilities from below and above. The constants are not claimed to be optimal: Remark 6.1 shows they cannot be improved by more than factors $\simeq5.1$ (real) and $\simeq7.1$ (complex), since the diagonal-direction limits are $\sqrt6/(\pi e^3)\simeq0.3084$ and $2/e^2\simeq0.2707$.
Load-bearing premise
The numerical content of Theorem 1.2 depends on the upper-tail estimate in Proposition 5.4; as printed, that estimate gives a bound larger than 1 for all $t>1$, so it cannot support the integration that produces $1/17$ and $1/27$.
Editorial extensions
If this is right
- For every dimension $n$ and every unit direction $a$, the slice $Q_n\cap\{x:\langle x,a\rangle=1/2\}$ has $(n-1)$-volume strictly larger than $1/17$, so high-dimensional cubes have uniformly thick off-center hyperplane sections.
- Because the bound is independent of $n$, the result controls all non-central hyperplane sections simultaneously as $n$ grows, complementing the known upper bounds by the same dimension-free character.
- The distance $1/2$ is the entire non-empty range: for any subspace $E$ and any $v\in E^\perp$ with $|v|\le1/2$, the parallel section has volume at least $\varepsilon(d)$, while for $|v|>1/2$ some choices give empty sections.
- For every codimension $d$ a constant $\varepsilon(d)$ exists, so the phenomenon is not special to hyperplanes.
- In the complex case, the same argument gives $A_{\mathbb C}(a,1)>1/27$ for polydisc hyperplane sections, a dimension-free lower bound in $\mathbb{C}^n$.
Reading between the lines
- The sphere-integral method should extend to other rotationally symmetric bodies whose section volumes admit Bessel-function or Fourier representations, giving dimension-free lower bounds for non-central sections of such bodies in their guaranteed non-empty distance range.
- The true infimum of $A_{\mathbb R}(a,1)$ is probably larger than $1/17$; the diagonal-direction limit $\sqrt6/(\pi e^3)$ and the exact small-dimension values suggest that the extremal direction may vary with $n$.
- Since the printed upper-tail estimate of Proposition 5.4 is vacuous for $t$ close to $1$, a necessary repair is a valid exponential tail bound with constants depending only on $k$; if such a bound holds, the qualitative theorem survives with possibly different constants.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies lower bounds for the volume of non-central sections of the unit-volume cube Q_n and of the complex polydisc. The first main result, Theorem 1.1, asserts that for every codimension d there is a constant ε(d) > 0 such that every affine (n−d)-plane at distance at most 1/2 from the origin cuts a section of volume at least ε(d), uniformly in the ambient dimension n and the direction of the section. The second main result, Theorem 1.2, gives explicit uniform bounds for hyperplane sections at distance exactly 1/2: A_R(a,1) > 1/17 in the real case and A_C(a,1) > 1/27 in the complex case, for every unit direction a. The proofs combine a probabilistic representation of section volumes as integrals over independent spherical vectors, exponential tail estimates obtained via Orlicz-space duality, and known results of Vaaler, Ball, and Veraar. The paper also contains a new method for bounding probabilities of the form P(|Σ a_j U_j| ≥ 1) for spherical vectors U_j, which may be of independent interest.
Significance. If the central claims hold, the paper provides the first explicit dimension-free lower bounds for non-central hyperplane sections of the cube and polydisc at the maximal distance where sections can be nonempty, complementing the classical central-section results of Vaaler and Ball. The constants are derived rather than numerically fitted, and the proof introduces a technically interesting Orlicz-space duality argument for spherical sums that goes beyond the standard Hanson-Wright framework. The lower bounds are not optimal, as the discussion in Remark 6.1 shows, but the qualitative uniformity in dimension and direction is the main contribution.
major comments (2)
- [§5, Proposition 5.4; §6, real and complex cases] The tail bound in Proposition 5.4 is printed with the wrong exponent: the statement P(|Σ a_j U_j| ≥ t) ≤ t^k exp(k/2 − k/(2t^2)) is vacuous for every t > 1, since the exponent k/2 − k/(2t^2) is positive and the right-hand side exceeds 1. The displayed minimization leading to the bound actually yields t^k exp(k/2 − k t^2/2); the factor t^{-2} in the exponent is a typographical error. This is not a cosmetic issue because Section 6 relies on the bound as a genuine tail estimate: the integral ∫_{t0}^∞ v exp(3/2 − 3/(2v^2)) dv, used in the real case, diverges, and the claimed equality with (1/3)exp(3/2 − 3/(2t0^2)) is false. The numerical constants 0.06011 and 0.03789, and the equations defining t0 and t1, correspond to the corrected exponent −(k/2)t^2, not to the printed inequality. The argument can likely be repaired by a local correction, but as written the proof of Theorem 1.2 does not go through.
- [§6, real and complex computations] The derivation of the explicit lower bounds uses a relationship between the threshold t0 (or t1) and the probability p0 (or p1) that is stated with the incorrect tail expression. The displayed equations t0^3 exp(3/2 − 3/(2t0^2)) = p0 and t1^4 exp(2 − 2t1^2) = p1 are inconsistent with the subsequent integration: the numerical values t0 ≈ 1.9182 and t1 ≈ 1.7657 solve the equations with the exponent −(3/2)t^2 and −2t^2, respectively. The printed versions make the constants p0 and p1 incompatible with the stated t0 and t1, so the numerical verification of the inequalities must be redone with the corrected tail bound.
minor comments (4)
- [§1, Introduction] The phrase 'probabiliy' in the reference [He] (Hensley, 'Slicing the cube in Rn and probabiliy') appears to be a misspelling of 'probability'; the reference should be checked against the original publication.
- [§4, Lemma 4.2] In the proof of Lemma 4.2, the inequality ∏_{j=2}^n (1+y_j) ≥ 1 + ∑_{j=2}^n y_j is stated for y_j ∈ (−1,1) 'having the same sign'; in the application all y_j are nonnegative, so the condition is satisfied, but this should be made explicit to avoid ambiguity.
- [§5, proof of Proposition 5.1, part (b)] The counting argument for the fourth moment E(S^4) is compressed: the sentence about '6 times' and the bound 24(∑ a_i^2 a_j^2 a_l^2 a_m^2) ≤ 4(∑ a_i^2 a_j^2)^2 is hard to follow. A short derivation of the displayed inequality would improve readability.
- [§6, real case] After correcting the tail exponent, the displayed intermediate expression 'p0/t0 (1 − 1/(3t0^3))' should read 'p0/t0 (1 − 1/(3t0^2))' in the real case; the printed form appears to be a typographical error.
Circularity Check
No significant circularity: the lower bounds are derived from independent external theorems and the self-citations are not load-bearing.
full rationale
The proofs of Theorems 1.1 and 1.2 do not reduce by construction to their inputs. The volume formulas in Proposition 3.1 are cited to Ball, Pólya, and Oleszkiewicz–Pelczynski, and Proposition 3.2 derives the spherical integral representations from those standard identities rather than assuming the desired lower bound. Proposition 5.1 uses Veraar's independent centered-variable inequality together with an explicit moment computation, and Proposition 5.4 invokes the König–Kwapien Khintchine inequality, an external published result whose stated assumptions do not include the cube-section conclusion. The upper bound cited from [KK3] is used only for context, not as an input to the lower-bound argument. The printed Proposition 5.4 tail estimate is indeed vacuous for t>1, and the Section 6 integration as printed would diverge; however, this is a correctness or typographical defect in an inequality whose intended exponent is derived, not a case of the target claim being assumed through a self-citation or a fitted parameter. No step in the derivation chain is equivalent to its own inputs by definition.
Assumptions & free parameters
assumptions (5)
- standard math Every central hyperplane section of the unit cube has (n-1)-volume at least 1 (Vaaler's theorem).
- standard math Every k-codimensional central section of the unit cube has volume at most 2^(k/2) (Ball's theorem).
- standard math For centered real-valued X, P(X >= 0) >= (2*sqrt(3)-3) E[X^2]^2 / E[X^4] (Veraar's inequality).
- standard math Khintchine inequality for independent uniform vectors on S^(k-1): ||sum a_j U_j||_p <= b_{p,k} |a| with best constants.
- standard math The density of the projection of uniform measure on a convex body is log-concave (Brunn-Minkowski).
Cite this review
Pith. "Pith review of On the volume of non-central sections of a cube." pith.science (2026). https://pith.science/paper/NUHKNX7L
@misc{pith2026190809358,
author = {Pith},
title = {Pith review of: On the volume of non-central sections of a cube},
year = {2026},
howpublished = {\url{https://pith.science/paper/NUHKNX7L}},
note = {Machine review of arXiv:1908.09358}
}
abstract
Let $Q_n$ be the cube of side length one centered at the origin in $\mathbb{R}^n$, and let $F$ be an affine $(n-d)$-dimensional subspace of $\mathbb{R}^n$ having distance to the origin less than or equal to $\frac 1 2$, where $0<d<n$. We show that the $(n-d)$-dimensional volume of the section $Q_n \cap F$ is bounded below by a value $c(d)$ depending only on the codimension $d$ but not on the ambient dimension $n$ or a particular subspace $F$. In the case of hyperplanes, $d=1$, we show that $c(1) = \frac{1}{17}$ is a possible choice. We also consider a complex analogue of this problem for a hyperplane section of the polydisc.
Reference graph
Works this paper leans on
-
[1]
Ball; Cube slicing in ^n , Proc
K. Ball; Cube slicing in ^n , Proc. Amer. Math. Soc. 97 (1986), 465-473
work page 1986
-
[2]
K. Ball; Volumes of sections of cubes and related problems, Geometric aspects of functional analysis, Lecture Notes in Math., 1376, 251-260, Springer, Berlin, 1989
work page 1989
-
[3]
Ball; Volume ratios and a reverse isoperimetric inequality, J
K. Ball; Volume ratios and a reverse isoperimetric inequality, J. London Math. Soc. 44 (1991), 351-359
work page 1991
- [4]
-
[5]
J. Bourgain; On the distribution of polynomials in high-dimensional convex sets, Geometric aspects of functional analysis, Lecture Notes in Math., 1469, 127-137, Springer, Berlin, 1991
work page 1991
-
[6]
A. Ben-Tal, A. Nemirovski, and C. Roos; Robust solutions of uncertain quadratic and conic-quadratic problems, SIAM Journal on Optimization, 13(2): 535-560, 2002
work page 2002
-
[7]
R. B. Boppana, R. Holzman; Tomaszewski's problem on randomly signed sums: breaking the 3/8 barrier, Electron. J. Combin. 24 (2017), no. 3, Paper 3.40, 10 pp
work page 2017
-
[8]
S. Brazitikos, A. Giannopoulos, P. Valettas, B. H. Vritsiou, Geometry of isotropic convex bodies. Mathematical Surveys and Monographs, 196. American Mathematical Society, Providence, RI, 2014
work page 2014
Show all 34 references
-
[9]
D. L. Burkholder; Independent sequences with the Stein property, Ann. Math. Stat. 39 (1968), 1282-1288
1968
-
[10]
Gradstein, I
I, S. Gradstein, I. M. Ryshik; Table of Series, Products and Integrals, VEB Deutscher Verlag der Wissensch., 1957
1957
-
[11]
Hadwiger; Gitterperiodische Punktmengen und Isoperimetrie, Monatsh
H. Hadwiger; Gitterperiodische Punktmengen und Isoperimetrie, Monatsh. Math. 76 (1972), 410-418
1972
-
[12]
Hensley; Slicing the cube in ^n and probabiliy, Proc
D. Hensley; Slicing the cube in ^n and probabiliy, Proc. Amer. Math. Soc. 73 (1979), 95-100
1979
-
[13]
Holzman and D
R. Holzman and D. J. Kleitman; On the product of sign vectors and unit vectors, Combinatorica, 12(3): 303-316, 1992
1992
-
[14]
D. L. Hanson, E. T. Wright; A bound on tail probabilities for quadratic forms in independent random variables, Ann. Math. Statist. 42 (1971), 1079-1083
1971
-
[15]
K\"onig, A
H. K\"onig, A. Koldobsky; Volumes of low-dimensional slabs and sections in the cube, Adv. in Appl. Math. 47 (2011), 894-907
2011
-
[16]
K\"onig, A
H. K\"onig, A. Koldobsky; Minimal volume of slabs in the complex cube, Proc. Amer. Math. Soc. 140 (2012), 1709-1717
2012
-
[17]
K\"onig, A
H. K\"onig, A. Koldobsky; On the maximal perimeter of sections of the cube, Advances in Math. 346 (2019), 773-804
2019
-
[18]
K\"onig, S
H. K\"onig, S. Kwapie\'n; Best Khintchine type inequalities for sums of independent, rotationally invariant random vectors, Positivity 5 (2001), 115-152
2001
-
[19]
Klartag; An isomorphic version of the slicing problem, J
B. Klartag; An isomorphic version of the slicing problem, J. Funct. Anal. 218 (2005), 372-394
2005
-
[20]
Koldobsky; Fourier Analysis in Convex Geometry, Amer
A. Koldobsky; Fourier Analysis in Convex Geometry, Amer. Math Soc., Providence, 2005
2005
-
[21]
Koldobsky; An application of the Fourier transform to sections of star bodies, Israel J
A. Koldobsky; An application of the Fourier transform to sections of star bodies, Israel J. Math 106 (1998), 157-164
1998
-
[22]
Kwapie\'n; Personal communication
S. Kwapie\'n; Personal communication
-
[23]
Meyer, A
M. Meyer, A. Pajor; Sections of the unit ball of l_q^n , J. Funct. Anal. 80 (1988), 109-123
1988
-
[24]
Milman, A
V. Milman, A. Pajor; Isotropic position and inertia ellipsoids and zonoids of the unit ball of a normed n -dimensional space, Geometric aspects of functional analysis, Lecture Notes in Math. 1376, 64-104, Springer, Berlin, 1989
1989
-
[25]
Moody, C
J. Moody, C. Stone, D. Zach, A. Zvavitch; A remark on extremal non-central sections of the unit cube, Asympt. Geometr. Anal., Fields Inst. Comm. 68, 211-228, Springer, New York, 2013
2013
-
[26]
Oleszkiewicz; On the Stein property of Rademacher sequences, Prob
K. Oleszkiewicz; On the Stein property of Rademacher sequences, Prob. and Math. Stat. 16 (1996), 127-130
1996
-
[27]
K. Oleszkiewicz; On p-pseudostable random variables, Rosenthal spaces and l_p^n ball slicing, Geometric aspects of functional analysis, Lecture Notes in Math., 1807, 188-210, Springer, Berlin, 2003
2003
-
[28]
Oleszkiewicz, A
K. Oleszkiewicz, A. Pelczy\'nski; Polydisc slicing in ^n , Studia Math. 142 (2000), 281-294
2000
-
[29]
P\'olya; Berechnung eines bestimmten Integrals, Math
G. P\'olya; Berechnung eines bestimmten Integrals, Math. Ann. 74 (1913) 204-212
1913
-
[30]
Rudelson, R
M. Rudelson, R. Vershynin; Hanson-Wright inequality and sub-Gaussian concentration, Electron. Commun. Probab. 18 (2013), no. 82, 9 pp
2013
-
[31]
J. D. Vaaler; A geometric inequality with applications to linear forms, Pacific J. Math. 83 (1979), 543-553
1979
-
[32]
Veraar; A note on optimal probability lower bounds for centered random variables, Colloq
M. Veraar; A note on optimal probability lower bounds for centered random variables, Colloq. Math 113 (2008), 231-240
2008
-
[33]
Watson; A treatise on the theory of Bessel functions, second edition, Cambridge Univ
G. Watson; A treatise on the theory of Bessel functions, second edition, Cambridge Univ. Press, 1952
1952
-
[34]
Webb; Central slices of the regular simplex, Geom
S. Webb; Central slices of the regular simplex, Geom. Dedic. 61 (1996), 19-28
1996
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.