On saddlepoints of two-person zero-sum games with applications to data verification tests (Q1972599)

The authors prove the following: Consider a two-person zero-sum game in which the set of pure strategies of one of the players is a compact metric space, the other player has a best answer to each randomized strategy and the payoff function satisfies a continuity condition concerning the weak convergence of probability measures. Under these conditions the game has a saddlepoint. This result is then applied to an important practical problem in which player 2 , an inspectee, makes \(n\) observations and reports these to player 1, the inspector, who proceeds to verify the claim by making independent observations. The inspectee can gain by cheating or reporting deliberately misleading values, provided the inspector fails to detect it. This situation is modelled as a zero-sum two-person game and the result of the paper is applied. See also \textit{R. Avenhaus, H. P. Battenberg} and \textit{B. J. Falkowski} [J. Oper. Res. 39, 341-348 (1991; Zbl 0793.68055)].