Bayes Goes to Bat: using baseball to introduce Bayesian estimation

Abstract

Based on an activity by John Spurrier, we use a baseball example to introduce students to Bayesian estimation. Students use prior information to determine prior distributions which lead to different estimators of the probability of a hit in baseball. We also compare our different Bayesian estimators and different frequentist estimators using bias, variability, and mean squared error. We can see the effect that sample size and dispersion of the prior distribution have on the estimator.