Pig Data and Bayesian Inference on Multinomial Probabilities

Abstract

Bayesian inference on multinomial probabilities is conducted based on data collected from the game Pass the Pigs®. Prior information on these probabilities is readily available from the instruction manual, and is easily incorporated in a Dirichlet prior. Posterior analysis of the scoring probabilities quantifies the discrepancy between empirical and prior estimates, and yields posterior predictive simulations used to compare competing extreme strategies.

Acknowledgments

The author would like to acknowledge the Phillip H. and Betty L. Wimmer Family Foundation for contributing their generous support to this project. The author is grateful to Rachel Riberich for her contribution of 3000 pig rolls, and to the editor and referees for their comments and suggestions.