arXiv Science⌕ Search

arXiv · 2611.23308

Condorcet's nontransitive dice: A minimal realization of a complete ranking profile

Abstract

We construct three independent finite-support random variables whose six strict ordering probabilities reproduce the complete ranking-frequency profile in a classical example of Condorcet's paradox. The construction has total support size seven, and we prove that no realization of this profile has smaller total support size. Here dice are allowed to be weighted: different outcomes need not have equal probabilities. We then formulate the realization problem for an arbitrary vector of six ordering probabilities. We distinguish constraints on the pairwise comparison probabilities from constraints on the full ranking distribution and give an example showing that pairwise feasibility is not sufficient. For the fixed interlaced supports $\{1,4,7\}$, $\{2,5,8\}$, and $\{3,6,9\}$, we derive the polynomial realization system. On the branch where the smallest outcome of the second variable has probability zero, we give a compact polynomial characterization for profiles with their first two ordering probabilities positive. This yields an explicit semialgebraic sufficient condition for the general realization problem and a rational construction, both with elementary proofs.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Sándor Bozóki, László Csató, József Temesi. 2026-11-11. Condorcet's nontransitive dice: A minimal realization of a complete ranking profile. https://arxiv.science/abs/2611.23308

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

Inequalities involving Higher Degree Polynomial Functions in $π(x)$

Earlier research reflects that the two sides of Ramanujan's inequality for the prime-counting function \begin{align*} π(x)^2<\frac{ex}{\log x}π(x/e) \end{align*} happens to agree through four asymptotic orders. In this paper, we try to examine how this cancellation arises in higher-degree expressions and how shifts affect the size of the resulting gaps. In signed sums of equal powers of shifted prime counts, the first nonvanishing moment of the logarithmic shifts determines the leading term. A cubic comparison at five dilations exhibits four successive cancellations, leaving a gap of order $x^3/(\log x)^7$. A finite-difference argument proves its positivity for every real $x\geq e^{52}$ when the logarithmic shifts are spaced by $2$. When the shifts and nonnegative weights are fixed, the weighted Ramanujan gap retains order $x^2/(\log x)^6$. Its leading coefficient is $c^2(1+σ^2)$, where $c$ is the total rescaled weight and $σ^2$ is the weighted variance of the logarithmic shifts. We also give the explicit range $x\geq 85e$ for the logarithmically weighted five-shift inequality and determine the eventual sign of a polynomial family in every fixed integer degree $m\geq 3$. Exact counts and interval enclosures illustrate the expansions and the effect of variance at fixed rescaled weight and mean.

math.GM↗

Generalized Centralizers and Polynomial Equivalence of Matrices: Answers to Questions of Dolinar, Guterman, Kuzma, and Markova

For $n\times n$ matrices $A$ and $B$ over a field and a scalar $ω\ne0$, let $\mathcal C_ω(A)=\{X:AX=ωXA\}$. If $ω$ is a root of unity of order $q\ge2$, then $\mathcal C_{ω^j}(A)=\mathcal C_{ω^j}(B)$ for all integers $j$ if and only if each of $A$ and $B$ is a polynomial in the other with exponents $\equiv1\pmod q$. Over algebraically closed fields, for $ω\ne1$, all $A$, and scalars $\varepsilon\ne0$, we describe all $Y$ with $XY=\varepsilon YX$ for all $X\in\mathcal C_ω(A)$. Let $a(μ)$ be the size of the largest Jordan block of $A$ at $μ$, or $0$ if there is none. If $\varepsilon$ is a power of $ω$, all such $Y$ are polynomials in $A$ if and only if $a(ωμ)=a(μ)$ for all $μ\ne0$, or this fails only at some $ν$ and at $ω^{-1}ν$, where $A$ has a single largest Jordan block at $ν$ and $a(ων)=a(ω^{-1}ν)=a(ν)-1$. This answers two questions of Dolinar, Guterman, Kuzma, and Markova. If $ω\ne1$ has finite order, then for such $A$, $\mathcal C_ω(A)=\mathcal C_ω(B)$ exactly when $A$ and $B$ are polynomials in each other of this form. For all other $A$ some $B$ that is not a polynomial in $A$ satisfies $\mathcal C_ω(A)=\mathcal C_ω(B)$ if the field has more than $n^2+1$ elements. In the $ω$-commuting digraph of Raja and Vaezpour, with an edge from $A$ to $B$ if $AB=ωBA$, vertices with edges to and from the same vertices are twins. Over fields with more than $2n^2+1$ elements we determine the vertices $A$ whose twins are all polynomials in $A$, and describe these twins. We find all $A$, $B$, and $k\ge2$ for which the kernel of $(X\mapsto[A,X])^k$ is the set of $X$ with $[X,[X,\ldots,[X,B]\ldots]]=0$ ($k$ brackets), where $[X,Y]=XY-YX$. Apart from scalar $B$, this happens only in characteristic $\ne2$, for $k=2$ and $2\times2$ matrices that generate the same quadratic extension.

math.GM↗

Theta-Eta Mellin Transforms, Shifted Reflections, and Rigorous Approximation Theory for a Variant Theta Construction

We develop a rigorous analytic framework for investigating a proposed functional equation associated with a variant theta construction. We prove the classical Mellin identity connecting the Jacobi kernel $1-θ_4(0\mid ix^2)$ to the completed Dirichlet eta function, and extend its Mellin transform meromorphically to the whole plane with a single simple pole at the origin. Mellin uniqueness identifies $x^{3/8}(1-θ_4(0\mid ix^2))$ as the unique kernel in a natural weighted $L^1$ class whose Mellin transform is the shifted completed eta function. We distinguish this Gaussian kernel from the weighted exponential series $Θ^B(s)=\sum_{n\geq1}ne^{-ns}$, whose Mellin transform is $Γ(s)ζ(s-1)$ and obeys a different, explicitly derived reflection relation. We show that every affine reflection $s\mapsto c-s$ can be generated for a shifted completed eta or zeta function by choosing the shift appropriately; consequently, the occurrence of $\tfrac14-s$ alone cannot identify the original variant with a shifted eta function. The corrected analytic product defined here is an elementary meromorphic multiplier of the completed zeta function and therefore has a precise $1-s$ functional equation. We prove an explicit eta-tail bound, a sharper alternating-tail asymptotic, a uniformly convergent Euler-transformed eta series with a geometric error bound, a uniform zeta-approximation bound away from the line $\Re s=1$, and a combined truncation estimate when both the zeta and weighted-exponential factors are truncated. We also determine the principal Laurent data and the cancellation of the gamma poles at the trivial zeta zeros for the corrected product. Reproducible high-precision experiments verify the stated error scales. The exact identification of the original variant with a shifted completed eta function remains open.

math.GM↗