Challenging Nostalgia and Performance Metrics in Baseball

• Activity Number: 327 - Stories of CHANCE: Frontiersman, Exoneree, Player, Spy
• Type: Invited
• Date/Time: Thursday, August 12, 2021 : 10:00 AM to 11:50 AM
• Sponsor: CHANCE
• Author(s): Daniel Eck*
• Companies: Department of Statistics, University of Illinois
• Keywords: bias; Sabermertrics; population dynamics; segregation in baseball; Binomial Distribution

Abstract

We show that the great baseball players that started their careers before 1950 are overrepresented among rankings of baseball's all time greatest players. The year 1950 coincides with the decennial US Census that is closest to when Major League Baseball (MLB) was integrated in 1947. We also show that performance metrics used to compare players have substantial era biases that favor players who started their careers before 1950. In showing that the these players are overrepresented, no individual statistics or era adjusted metrics are used. Instead, we argue that the eras in which players played are fundamentally different and are not comparable. In particular, there were significantly fewer MLB eligible players available at and before 1950. As a consequence of this and other differences across eras, we argue that popular opinion, performance metrics, and expert opinion over include players that started their careers before 1950 in their rankings of baseball's all time greatest players.