Query Complexity of Mastermind Variants
Abstract
We study variants of Mastermind, a popular board game in which the objective is sequence reconstruction. In this two-player game, the so-called \textit{codemaker} constructs a hidden sequence of colors selected from an alphabet (\textit{i.e.,} for all ). The game then proceeds in turns, each of which consists of two parts: in turn , the second player (the \textit{codebreaker}) first submits a query sequence with for all , and second receives feedback , where is some agreed-upon function of distance between two sequences with components. The game terminates when , and the codebreaker seeks to end the game in as few turns as possible. Throughout we let denote the smallest integer such that the codebreaker can determine any in turns. We prove three main results: First, when is known to be a permutation of , we prove that for all sufficiently large . Second, we show that Knuth's Minimax algorithm identifies any in at most queries. Third, when feedback is not received until all queries have been submitted, we show that .
Cite
@article{arxiv.1607.04597,
title = {Query Complexity of Mastermind Variants},
author = {Aaron Berger and Christopher Chute and Matthew Stone},
journal= {arXiv preprint arXiv:1607.04597},
year = {2017}
}
Comments
Revised and trimmed- 17 pages