Abstract
We present a nonparametric statistical test for determining whether an agentis following a given mixed strategy in a repeated strategic-form game givensamples of the agent's play. This involves two components: determining whetherthe agent's frequencies of pure strategies are sufficiently close to the targetfrequencies, and determining whether the pure strategies selected areindependent between different game iterations. Our integrated test involvesapplying a chi-squared goodness of fit test for the first component and ageneralized Wald-Wolfowitz runs test for the second component. The results fromboth tests are combined using Bonferroni correction to produce a complete testfor a given significance level $\alpha.$ We applied the test to publiclyavailable data of human rock-paper-scissors play. The data consists of 50iterations of play for 500 human players. We test with a null hypothesis thatthe players are following a uniform random strategy independently at each gameiteration. Using a significance level of $\alpha = 0.05$, we conclude that 305(61%) of the subjects are following the target strategy.