scientific article; zbMATH DE number 2211587

DOI10.1155/IMRN.2005.1601MaRDI QIDQ5694432zbMATH OpenFDO

Authors C. Sinan Güntürk

Publication date 30 September 2005

Published in IMRN. International Mathematics Research Notices (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/math/0503373

zbMATH Keywords

bias functionfair duel problemfairness of ordering sequencepower series approximation

Mathematics Subject Classification ID

Approximation in the complex plane (30E10) Combinatorial probability (60C05) Inequalities; stochastic orderings (60E15) Asymptotic distribution theory in statistics (62E20) Numerical summation of series (65B10)

Review text

The aim of this paper is to provide a generalized solution to the so-called ``fair duel problem, described as follows: there are two duelists \(X\) and \(Y\) who will shoot at each other, one at a time, using a given \(\pm 1\) sequence \(q = (q_n)_{n \geq 0}\) to specify whose turn is to shoot at time \(n\). The shots are i.i.d. (independent, identically distributed) random variables whose outcome is hit or miss. Each shot hits (and kills) its target with the probability \(\varepsilon\), which is arbitrary but fixed throughout the duel. Now, the ``fair duel problem consists in finding an ordering \(q\), independent of the hit probability \(\varepsilon\), such that the probability of survival for each duelist to be as close as possible to 1/2. The fairness of the ordering sequence \(q\) is measured by its bias function \(B_q(\varepsilon)\) defined as \(B_q(\varepsilon) = {\mathbf P}(X \text{ survives}) -{\mathbf P}(Y \text{ survives})\), under the condition \(B_q(\varepsilon) \rightarrow 0\) as fast as possible, when \(\varepsilon\rightarrow 0.\) The bias function takes various shapes depending on the ordering sequence \(q_n\), from the simple alternation between the duelists \(X\) and \(Y\), i.e. \(q_n = (-1)^n\), to other much more complicated (periodic) forms. The main result of the paper is the construction of universal orderings \(q\) for which \(B_q(\varepsilon) = {\mathbf O}(e^{- c / \varepsilon})\), where \(c\) is an absolute constant. In the proposed framework, the bias function \(B_q(\varepsilon)\) of the fair duel problem is proved to be approximated by power series with \(\pm 1\) coefficients.

Recommendations

• Greedy Galois games
• On the approximation of positive functions by power series. II
• scientific article; zbMATH DE number 1374999
• The stochastic duel with time-dependent hit probabilities
• Best approximation of alternating power series

Cited in

(8)

• Best approximation of alternating power series
• Greedy Galois games
• scientific article; zbMATH DE number 3855675 (Why is no real title available?)
• George Lorentz and inequalities in approximation
• EXTRAPOLATION OF POWER SERIES BY SELF-SIMILAR FACTOR AND ROOT APPROXIMANTS
• On the Multiplicity of the Zeros of Polynomials with Constrained Coefficients
• The size of exponential sums on intervals of the real line
• Pseudo-Boolean functions and the multiplicity of the zeros of polynomials

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5694432)