betaFeedback

WorksMathematics

An improved lower bound for piercing the chessboard by lines

Abstract

Let pnp_n be the least number of lines meeting the interior of every cell of an n×nn\times n board. We prove that lim inf⁡n→∞pn/n≥0.7520517>34\liminf_{n\to\infty}p_n/n\ge0.7520517>\frac34. The previous bound was 0.70.7. The proof is computer-assisted. It identifies a non-negative polynomial on the unit square and bounds its integral along every line.

Setup and results

Let Bn=[0,n]2B_n=[0,n]^2. Its cells are the open unit squares Ci,j=(i−1,i)×(j−1,j)C_{i,j}=(i-1,i)\times(j-1,j), 1≤i,j≤n1\le i,j\le n. A line pierces a cell if it meets its interior. A family of lines pierces the board if every cell is pierced. Let pnp_n be the least size of a piercing family.

The problem is Keller’s Problem E2862 [4]. Ackerman and Pinchasi [1] proved pn≥⌈23n⌉p_n\ge\lceil\frac{2}{3}n\rceil. Ambrus, Bárány, Frankl and Varga [3] proved 0.7n≤pn≤n−10.7n\le p_n\le n-1 for large nn and conjectured pn=n−1p_n=n-1. Alkauskas [2] proved pn≥n−1p_n\ge n-1 for families of lines whose slopes all lie in (−∞,−1]∪(0,1](-\infty,-1]\cup(0,1]. For unrestricted lines the conjecture is open. The lower bound of [3] is the method of Theorem 1.1 below with a quadratic polynomial. We use a polynomial of degree 28.

For a line ℓ\ell, let γ\gamma be the measure on ℓ\ell with dγ=∣dx∣+∣dy∣d\gamma=|dx|+|dy|. Explicitly, if ℓ\ell is neither horizontal nor vertical,

∫ℓ∩[0,1]2f dγ=∫f(x,ℓ(x)) dx+∫f(ℓ−1(y),y) dy,\int_{\ell\cap[0,1]^2} f\,d\gamma=\int f(x,\ell(x))\,dx+\int f(\ell^{-1}(y),y)\,dy,

the two integrals taken over the projections of ℓ∩[0,1]2\ell\cap[0,1]^2 to the xx-axis and to the yy-axis. For a horizontal line only the first term is present, and for a vertical line only the second.

Theorem 1.1. Let WW be a polynomial with W≥0W\ge0 on [0,1]2[0,1]^2. Suppose ∬[0,1]2W dx dy=α\iint_{[0,1]^2}W\,dx\,dy=\alpha and ∫ℓ∩[0,1]2W dγ≤β\int_{\ell\cap[0,1]^2}W\,d\gamma\le\beta for every line ℓ\ell. Then

lim inf⁡n→∞pnn≥αβ.\liminf_{n\to\infty}\frac{p_n}{n}\ge\frac{\alpha}{\beta}.

Theorem 1.2. There is an explicit polynomial W14(x,y)W_{14}(x,y) with W14≥0W_{14}\ge0 on [0,1]2[0,1]^2,

∬[0,1]2W14 dx dy=662695081102818845329881156433506484375000,∫ℓ∩[0,1]2W14 dγ≤100003100000 for every line ℓ.\iint_{[0,1]^2}W_{14}\,dx\,dy=\frac{662695081102818845329}{881156433506484375000},\qquad\int_{\ell\cap[0,1]^2}W_{14}\,d\gamma\le\frac{100003}{100000}\ \text{for every line }\ell.

Consequently

lim inf⁡n→∞pnn≥26507803244112753813163524731472797958278125=0.75205170…>34.\liminf_{n\to\infty}\frac{p_n}{n}\ge\frac{2650780324411275381316}{3524731472797958278125}=0.75205170\ldots>\frac{3}{4}.

§ 2 proves Theorem 1.1. § 3 reduces Theorem 1.2 to a finite computation, Theorem 3.2, which is carried out by the program described in § 4.

Proof of Theorem 1.1

We scale BnB_n by 1n\frac{1}{n} onto [0,1]2[0,1]^2. The scaled cells Ci,j/nC_{i,j}/n have side 1n\frac{1}{n}, and the interior grid lines are x=inx=\frac{i}{n} and y=jny=\frac{j}{n} for 1≤i,j≤n−11\le i,j\le n-1. For a polynomial VV on [0,1]2[0,1]^2 let

wi,j=n2∬Ci,j/nV dx dyw_{i,j}=n^2\iint_{C_{i,j}/n} V\,dx\,dy

the average of VV over the scaled cell, and for a line ℓ\ell let

S(ℓ)=∑(i,j): ℓ pierces Ci,jwi,j.S(\ell)=\sum_{(i,j):\,\ell\text{ pierces }C_{i,j}} w_{i,j}.

We write ∥P−P′∥1=∣x−x′∣+∣y−y′∣\lVert P-P'\rVert_1=|x-x'|+|y-y'| for the ℓ1\ell_1 distance. If ∣∂xV∣,∣∂yV∣≤Λ|\partial_xV|,|\partial_yV|\le\Lambda on [0,1]2[0,1]^2 then ∣V(P)−V(P′)∣≤Λ∥P−P′∥1|V(P)-V(P')|\le\Lambda\lVert P-P'\rVert_1 for P,P′∈[0,1]2P,P'\in[0,1]^2.

Lemma 2.1. Let VV be a polynomial with 0≤V≤M0\le V\le M and ∣∂xV∣,∣∂yV∣≤Λ|\partial_xV|,|\partial_yV|\le\Lambda on [0,1]2[0,1]^2. Then for every line ℓ\ell,

S(ℓ)≤n∫ℓ∩[0,1]2V dγ+3M+8Λ.S(\ell)\le n\int_{\ell\cap[0,1]^2}V\,d\gamma+3M+8\Lambda.

Proof. Let ℓ∘=ℓ∩(0,1)2\ell^\circ=\ell\cap(0,1)^2. If ℓ∘\ell^\circ is empty, or ℓ\ell is contained in a grid line, then ℓ\ell pierces no cell and S(ℓ)=0S(\ell)=0. So assume neither.

Call a point of ℓ∘\ell^\circ a crossing if it lies on a grid line. Call the crossing vertical if its xx-coordinate is a multiple of 1n\frac{1}{n}, and horizontal otherwise. A horizontal crossing has yy-coordinate a multiple of 1n\frac{1}{n}. There are at most n−1n-1 vertical crossings, at most one on each line x=inx=\frac{i}{n}, and at most n−1n-1 horizontal ones. Fix a direction along ℓ\ell. The crossings cut ℓ∘\ell^\circ into open pieces. Each piece lies in the interior of one cell, and ℓ\ell meets each cell interior in at most one segment. So the pieces correspond one to one to the pierced cells.

Every piece except the first begins at a crossing XX, and XX lies in the closure of the piece’s cell CC. So every point of CC is within ℓ1\ell_1-distance 2n\frac{2}{n} of XX, and the average of VV over CC is at most V(X)+2ΛnV(X)+\frac{2\Lambda}{n}. The cell of the first piece contributes at most MM. There are at most 2n2n crossings. Hence

S(ℓ)≤M+∑X verticalV(X)+∑X horizontalV(X)+4Λ.(1)S(\ell)\le M+\sum_{X\ \mathrm{vertical}}V(X)+\sum_{X\ \mathrm{horizontal}}V(X)+4\Lambda. \tag*{(1)}

If ℓ\ell is vertical, then there are no vertical crossings, and (2) below holds with the integral equal to 00. Otherwise write y=ℓ(x)y=\ell(x), let tt be the slope, and let [x−,x+][x_-,x_+] be the projection of ℓ∩[0,1]2\ell\cap[0,1]^2 to the xx-axis. Put r=x+−x−≤1r=x_+-x_-\le1 and s=∣t∣r≤1s=|t|r\le1, the run and the rise of the chord. The vertical crossings are the points Xi=(in,ℓ(in))X_i=\left(\frac{i}{n},\ell\left(\frac{i}{n}\right)\right) for the integers ii with x−<in<x+x_-<\frac{i}{n}<x_+, say i=a,…,bi=a,\ldots,b. Bound V(Xa)≤MV(X_a)\le M. For a<i≤ba<i\le b the part of ℓ\ell over [i−1n,in]\left[\frac{i-1}{n},\frac{i}{n}\right] lies in [0,1]2[0,1]^2, and each of its points is within ℓ1\ell_1-distance 1+∣t∣n\frac{1+|t|}{n} of XiX_i. So

V(Xi)≤n∫(i−1)/ni/nV(x,ℓ(x)) dx+(1+∣t∣)Λn.V(X_i)\le n\int_{(i-1)/n}^{i/n}V(x,\ell(x))\,dx+\frac{(1+|t|)\Lambda}{n}.

These b−ab-a intervals are disjoint subintervals of [x−,x+][x_-,x_+], so b−a≤nrb-a\le nr. Summing, and using V≥0V\ge0 to extend the integral to [x−,x+][x_-,x_+],

∑X verticalV(X)≤M+n∫x−x+V(x,ℓ(x)) dx+r(1+∣t∣)Λ≤M+n∫x−x+V(x,ℓ(x)) dx+2Λ,(2)\begin{aligned} \sum_{X\ \mathrm{vertical}}V(X)\le M+n\int_{x_-}^{x_+}V(x,\ell(x))\,dx+r(1+|t|)\Lambda\\ &\le M+n\int_{x_-}^{x_+}V(x,\ell(x))\,dx+2\Lambda, \tag*{(2)} \end{aligned}

since r(1+∣t∣)=r+s≤2r(1+|t|)=r+s\le2. We now use the same argument with the roles of xx and yy exchanged. If ℓ\ell is horizontal, then there are no horizontal crossings. Otherwise, write x=ℓ−1(y)x=\ell^{-1}(y), with slope 1/∣t∣1/|t| in this parametrisation, and let [y−,y+][y_-,y_+] be the projection of ℓ∩[0,1]2\ell\cap[0,1]^2 to the yy-axis. The horizontal crossings are among the points Yj=(ℓ−1(j/n),j/n)Y_j=(\ell^{-1}(j/n),j/n) with y−<j/n<y+y_-<j/n<y_+, and omitting some of these only decreases the sum. The number of nodes after the first is at most nsns, and the error per node is (1+1/∣t∣)Λn\frac{(1+1/|t|)\Lambda}{n}. Since s(1+1/∣t∣)=s+r≤2s(1+1/|t|)=s+r\le2,

∑horizontal XV(X)≤M+n∫y−y+V(ℓ−1(y),y) dy+2Λ.(3)\sum_{\text{horizontal }X} V(X)\le M+n\int_{y_-}^{y_+}V(\ell^{-1}(y),y)\,\mathrm{d}y+2\Lambda. \tag*{(3)}

We now use (2) and (3) to get an upper bound for the right-hand side in the inequality (1). The two integrals add up to ∫ℓ∩[0,1]22V dγ\int_{\ell\cap[0,1]^2}2V\,\mathrm{d}\gamma by the definition of γ\gamma in Section 1. So S(ℓ)≤n∫ℓ∩[0,1]22V dγ+3M+8ΛS(\ell)\le n\int_{\ell\cap[0,1]^2}2V\,\mathrm{d}\gamma+3M+8\Lambda.

Proof of Theorem 1.1. Let M,ΛM,\Lambda be as in Lemma 2.1 with V=WV=W. The scaled cells tile [0,1]2[0,1]^2 up to a set of measure zero, so

∑i,jwi,j=n2∬[0,1]2W dx dy=αn2.\sum_{i,j}w_{i,j}=n^2\iint_{[0,1]^2}W\,\mathrm{d}x\,\mathrm{d}y=\alpha n^2.

Let ℓ1,…,ℓpn\ell_1,\ldots,\ell_{p_n} pierce BnB_n. Every cell is pierced by at least one ℓr\ell_r, and every wi,j≥0w_{i,j}\ge0. So by Lemma 2.1

αn2≤∑r=1pnS(ℓr)≤pn(βn+3M+8Λ).\alpha n^2\le\sum_{r=1}^{p_n}S(\ell_r)\le p_n(\beta n+3M+8\Lambda).

Divide by n(βn+3M+8Λ)n(\beta n+3M+8\Lambda) and let n→∞n\to\infty.

Reduction of Theorem 1.2 to a computation

Recall that the Bernstein polynomials of degree DD are

BiD(z)=(Di)zi(1−z)D−i,i=0,1,…,D.B_i^D(z)=\binom{D}{i}z^i(1-z)^{D-i},\qquad i=0,1,\ldots,D.

They are non-negative on [0,1][0,1], and they form a basis of the polynomials in zz of degree at most DD. The (D+1)2(D+1)^2 products BiD(z1)BjD(z2)B_i^D(z_1)B_j^D(z_2) form a basis of the polynomials in z1,z2z_1,z_2 of degree at most DD in each variable.

The polynomial W14W_{14} of Theorem 1.2 will be written as

W14(x,y)=∑i,j=014bijBi14((2x−1)2)Bj14((2y−1)2).(4)W_{14}(x,y)=\sum_{i,j=0}^{14}b_{ij}B_i^{14}\left((2x-1)^2\right)B_j^{14}\left((2y-1)^2\right). \tag*{(4)}

It is determined by 225 rational numbers bijb_{ij}, 0≤i,j≤140\le i,j\le14, with denominator 10610^6. These coefficients were found by solving a linear program numerically. The 120 values with i≤ji\le j are listed in the file W14.txt of [5].

We choose the coefficients so that bij=bjib_{ij}=b_{ji}. Since W14W_{14} depends on xx and yy only through (2x−1)2(2x-1)^2 and (2y−1)2(2y-1)^2, it is invariant under the symmetries (x,y)↦(1−x,y)(x,y)\mapsto(1-x,y), (x,y)↦(x,1−y)(x,y)\mapsto(x,1-y) and (x,y)↦(1−x,1−y)(x,y)\mapsto(1-x,1-y) of the square, and since bij=bjib_{ij}=b_{ji}, it is invariant under the symmetry (x,y)↦(y,x)(x,y)\mapsto(y,x).

For a polynomial WW on [0,1]2[0,1]^2 and (a,b)∈[0,1]2(a,b)\in[0,1]^2 define

FA(a,b)=(1+b−a)∫01W(τ,(1−τ)a+τb) dτ.(5)F_A(a,b)=(1+b-a)\int_0^1W\left(\tau,(1-\tau)a+\tau b\right)\,\mathrm{d}\tau. \tag*{(5)}
FB(a,b)=(1−a+b)∫01W(bτ,(1−τ)a+τ) dτ.(6)F_B(a,b)=(1-a+b)\int_0^1W\left(b\tau,(1-\tau)a+\tau\right)\,\mathrm{d}\tau. \tag*{(6)}

Here FB(a,b)F_B(a,b) is the integral of WW with respect to γ\gamma along the segment from (0,a)(0,a) to (b,1)(b,1), and, when a≤ba \le b, FA(a,b)F_A(a,b) is the integral along the segment from (0,a)(0,a) to (1,b)(1,b); see Figure 1 and the proof below.

Figure 1. The segments along which FAF_A and FBF_B integrate WW.

Lemma 3.1. Let WW be a polynomial with W≥0W \ge0 on [0,1]2[0,1]^2 that is invariant under the four symmetries (x,y)↦(1−x,y)(x,y) \mapsto(1-x,y), (x,y)↦(x,1−y)(x,y) \mapsto(x,1-y), (x,y)↦(1−x,1−y)(x,y) \mapsto(1-x,1-y) and (x,y)↦(y,x)(x,y) \mapsto(y,x). Then for every line ℓ\ell,

∫ℓ∩[0,1]2W dγ≤max⁡(max⁡[0,1]2FA,max⁡[0,1]2FB).(7)\int_{\ell\cap[0,1]^2} W\,\mathrm{d}\gamma\le\max\left(\max_{[0,1]^2} F_A,\max_{[0,1]^2} F_B\right). \tag*{(7)}

Proof. The four symmetries map lines to lines and preserve γ\gamma and WW. If ℓ∩[0,1]2\ell\cap[0,1]^2 is empty or a single point, the integral on the left-hand side of (7) is 0. The right-hand side is non-negative, because W≥0W \ge0 and the arguments of WW in (5) and (6) lie in [0,1]2[0,1]^2. Otherwise ℓ∩[0,1]2\ell\cap[0,1]^2 is a segment whose endpoints lie on two different edges of the square.

Suppose the endpoints lie on the left and the right edge, say at (0,a)(0,a) and (1,b)(1,b). Applying the symmetry (x,y)↦(x,1−y)(x,y) \mapsto(x,1-y) if necessary, we may assume a≤ba \le b. The points of the segment are (τ,(1−τ)a+τb)(\tau,(1-\tau)a+\tau b), τ∈[0,1]\tau\in[0,1], so dx=dτ\mathrm{d}x=\mathrm{d}\tau, dy=(b−a) dτ\mathrm{d}y=(b-a)\,\mathrm{d}\tau, and dγ=(1+b−a) dτ\mathrm{d}\gamma=(1+b-a)\,\mathrm{d}\tau. Hence the left-hand side of (7) is FA(a,b)F_A(a,b). If the endpoints lie on the bottom and the top edge, the symmetry (x,y)↦(y,x)(x,y) \mapsto(y,x) reduces this to the previous case.

Suppose the endpoints lie on two adjacent edges. One of the symmetries (x,y)↦(1−x,y)(x,y) \mapsto(1-x,y), (x,y)↦(x,1−y)(x,y) \mapsto(x,1-y), (x,y)↦(1−x,1−y)(x,y) \mapsto(1-x,1-y), or no symmetry, maps the segment onto a segment from the left edge to the top edge, say from (0,a)(0,a) to (b,1)(b,1). Its points are (bτ,(1−τ)a+τ)(b\tau,(1-\tau)a+\tau), τ∈[0,1]\tau\in[0,1], so dx=b dτ\mathrm{d}x=b\,\mathrm{d}\tau, dy=(1−a) dτ\mathrm{d}y=(1-a)\,\mathrm{d}\tau, and dγ=(1−a+b) dτ\mathrm{d}\gamma=(1-a+b)\,\mathrm{d}\tau. Hence the left-hand side of (7) is FB(a,b)F_B(a,b). □

Denote by FA14F_A^{14} and FB14F_B^{14} the polynomials (5) and (6) for W=W14W=W_{14}.

Theorem 3.2. The polynomial W14W_{14} has the following properties.

(a) W14≥0W_{14} \ge0 on [0,1]2[0,1]^2.

(b) ∬[0,1]2W14 dx dy=662695081102818845329881156433506484375000\iint_{[0,1]^2} W_{14}\,\mathrm{d}x\,\mathrm{d}y = \frac{662695081102818845329}{881156433506484375000}.

(c) FA14≤100003100000F_A^{14} \le\frac{100003}{100000} and FB14≤100003100000F_B^{14} \le\frac{100003}{100000} on [0,1]2[0,1]^2.

Theorem 3.2 is proved by computer in § 4.

Proof of Theorem 1.2. By Theorem 3.2(a), Lemma 3.1 applies to W14W_{14}, and with Theorem 3.2(c) it gives

∫ℓ∩[0,1]2W14 dγ≤100003100000\int_{\ell\cap[0,1]^2} W_{14}\,\mathrm{d}\gamma\le\frac{100003}{100000}

for every line ℓ\ell. Together with Theorem 3.2(a) and (b) these are the hypotheses of Theorem 1.1 with β=100003100000\beta=\frac{100003}{100000}, and α/β\alpha/\beta is the fraction in Theorem 1.2. □

The verifier

Theorem 3.2 is proved by the program [5]. It is written in C++, and all its arithmetic is exact. Integers are held in fixed-width words of 1024 bits from the Boost Multiprecision library, and every operation is checked for overflow. An overflow stops the verification. The program reads the 120 coefficients bijb_{ij} with i≤ji \le j, sets bji=bijb_{ji}=b_{ij}, forms W14W_{14} by (4), and verifies the following three statements, where β=100003100000\beta=\frac{100003}{100000}:

W14≥0 on [0,1]2,(8)W_{14} \ge0 \text{ on } [0,1]^2, \tag*{(8)}
∬[0,1]2W14 dx dy=662695081102818845329881156433506484375000,(9)\iint_{[0,1]^2} W_{14}\,dx\,dy = \frac{662695081102818845329}{881156433506484375000}, \tag*{(9)}
β−FA14≥0 and β−FB14≥0 on [0,1]2.(10)\beta-F_A^{14} \ge0 \text{ and } \beta-F_B^{14} \ge0 \text{ on } [0,1]^2. \tag*{(10)}

These are parts (a), (b) and (c) of Theorem 3.2.

Put S=(2x−1)2S=(2x-1)^2 and U=(2y−1)2U=(2y-1)^2. By (4), W14W_{14} is a polynomial in (S,U)(S,U) of degree 14 in each variable, and the map (x,y)↦(S,U)(x,y)\mapsto(S,U) sends [0,1]2[0,1]^2 onto itself. So (V1) says that this polynomial is ≥0\ge0 on [0,1]2[0,1]^2.

The functions FA14F_A^{14} and FB14F_B^{14} are defined by integrals, one for each (a,b)(a,b). Substituting the linear expressions of (5) and (6) into W14W_{14} gives a polynomial in aa, bb and τ\tau, and integrating each power τk\tau^k over [0,1][0,1] gives 1k+1\frac{1}{k+1}. Hence FA14F_A^{14} and FB14F_B^{14} are polynomials in (a,b)(a,b), of degree at most 29 in each variable: W14W_{14} has degree 28 in each of xx and yy, the arguments of WW in (5) and (6) are of degree 1 in each of aa and bb, and the factors in front add 1. The program computes them exactly, so (V3) says that two explicit polynomials are ≥0\ge0 on [0,1]2[0,1]^2.

For (V2), the program writes W14W_{14} in the monomials SiUjS^iU^j and uses

∬[0,1]2SiUj dx dy=1(2i+1)(2j+1).\iint_{[0,1]^2} S^iU^j\,dx\,dy=\frac{1}{(2i+1)(2j+1)}.

Let gg be a polynomial of degree at most D1D_1 in z1z_1 and at most D2D_2 in z2z_2. It can be written as

g(z1,z2)=∑i=0D1∑j=0D2cijBiD1(z1)BjD2(z2).(11)g(z_1,z_2)=\sum_{i=0}^{D_1}\sum_{j=0}^{D_2}c_{ij}B_i^{D_1}(z_1)B_j^{D_2}(z_2). \tag*{(11)}

If all cij≥0c_{ij}\ge0, then g≥0g\ge0 on [0,1]2[0,1]^2, because the Bernstein polynomials are non-negative on [0,1][0,1]. For W14W_{14} as a polynomial in (S,U)(S,U), the coefficients in (11) are the bijb_{ij}. They are not all non-negative, so (V1) is not immediate, and the same holds for the two polynomials in (V3).

The program therefore subdivides. It divides the square into four squares of half the side, maps each of them onto [0,1]2[0,1]^2 by an affine change of variables, and computes the coefficients (11) of gg on each. On a square where all coefficients are non-negative, g≥0g\ge0 is proved. A square where some coefficient is negative is divided again. If g>0g>0 on [0,1]2[0,1]^2, this terminates, since the coefficients of gg on a shrinking square converge to values of gg. The result is a finite list of squares covering [0,1]2[0,1]^2, with non-negative coefficients on each, and hence a proof that g≥0g\ge0.

For (V1) the final list has 160 squares, and for the two polynomials in (V3) it has 175 and 448 squares. The whole run takes less than a second on a laptop.

The maximum of FA14F_A^{14} and FB14F_B^{14} on [0,1]2[0,1]^2 exceeds 1 by about 10−510^{-5}, which is why β\beta is slightly larger than 1.

Acknowledgements

The authors used Anthropic’s Claude models (Claude Fable 5 and 5.1 and earlier Claude Opus models, accessed through claude.ai and Claude Code) between July and November 2026, in developing the ideas of the proofs, in writing the verifier, and in editing the text. The authors take full responsibility for the correctness of the results.

References

Department of Mathematics, Baruch College, City University of New York, One Bernard Baruch Way, New York, NY 10010, USA

Email address: [email protected]

Department of Mathematics, Baruch College, City University of New York, One Bernard Baruch Way, New York, NY 10010, USA

Email address: [email protected]

References

  1. [1]E. Ackerman and R. Pinchasi, Covering the chessboard with staircase walks, Discrete Math. 313 (2013), 2547–2551.DOI
  2. [2]G. Alkauskas, Piercing a chessboard with restricted slopes, arXiv:2611.12329.
  3. [3]G. Ambrus, I. Bárány, P. Frankl and D. Varga, Piercing the chessboard, SIAM J. Discrete Math. 37 (2023), 1457–1471, arXiv:2111.09702.DOI
  4. [4]T. Keller, Problem E2862, Amer. Math. Monthly 87 (1980).
  5. [5]I. Matić and R. Radoičić, Computer-assisted proof of a lower bound for piercing the chessboard (computer program), https://github.com/maticivan/piercing (2026), https://doi.org/10.5281/zenodo.24239645.

Paper details

vlnc:W_bwn7a7q7
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Claude Fable 5, 5.1, Claude Opus (multiple versions)

View statement
Looking for related papers…
Contents