個案資料

In 2017, a sports journalist based in Houston, Texas was confused about his editor’s latest idea. His editor had emailed him the night before proposing an article on the best (and worst) Major League Baseball hitters of all time. The first step was to provide a list of the players that would be profiled. This meant finding the 10 best (and worst) batters. However, after ranking the players, could he quantify the probability that he ranked them incorrectly? Because the article was based on data, he had to justify his choices and he wanted his decisions to be based on sound statistical grounds.

This case is intended for use in an advanced undergraduate or graduate business course. The content requires a basic familiarity with the Bayes theorem and related terms such as “prior” and “posterior.” However, outside of a basic familiarity, this case is best used as an introduction to the practical ways these terms manifest themselves in business. In a course, relevant readings can be assigned with the case to introduce Bayesian updating. This case is intended to provide students with an opportunity to think through how to empirically estimate a “Bayesian prior” and how to update it with new data. Students should emerge from this case having a better understanding of how Bayesian methods can be used in sports and entertainment environments and how Bayesian methods can help quantify uncertainty in low sample sizes. After completion of this case, students will be able to<ul><li>understand the parameters of a beta distribution and how they affect the mean/variance;</li><li>fit a beta distribution to empirical data using Microsoft Excel; and</li><li>update a beta distribution prior using new information.</li><ul>

• Bayes theorem , baseball, sports, data