Minimax play in Armageddon chess

Wed 18.11 11:30 - 12:30

abstract: We use data on bids and game outcomes from Armageddon chess in high-stakes professional chess tournaments to undertake the first test using field data of the Minimax Hypothesis, and its generalization to Nash equilibrium, when the solution is in pure strategies. Under Armageddon rules, Black wins either by winning or drawing, while White must win outright. In the game we study, players bid time for the right to play Black. The player who makes the smaller bid plays Black, with the starting time on their clock equal to their bid; their opponent plays White with a full clock. We find that bids and game outcomes are consistent with minimax and hence Nash play: each player is equally likely to play Black or White, and a player’s probability of winning the game is the same whether playing Black or White. In this latter respect, Armageddon bidding fairly divides the probability of winning a match. We show that players sacrifice time, and hence playing accuracy, in exchange for the advantage of playing Black. In the data, these two factors exactly balance, as predicted by theory.   joint work with Miguel A. Fonseca, Andrew Wooders, John Wooders

Speaker

Todd Kaplan

Haifa University