Sacrifice bunting and attempting stolen bases are popular strategies in baseball. We set out to determine if using these strategies help collegiate teams win games, and if so, what situations they can be used to maximize probability of winning. We found that, in most cases, sacrifice bunts and stolen bases decrease win probability. However, especially in late innings and in some specific base/out situations, sacrifice bunts and stolen bases can be beneficial. In 2016 SEC play, using our optimal strategy as opposed to the current usage, win probability is increased by 1.7%.
We took play-by-play data from 2016 SEC baseball games, compiling batting order, plate appearance results and future runs, to develop average expected runs tables along with probability of scoring tables for each base/out situation. In our initial data we had a glimpse into the optimal strategy but the sample size was too small to be statistically significant. Our next step was to create a simulation in Python to model SEC games. We used the simulation to create win probability charts based on every inning, and we compared simulations using various bunting and stealing strategies to develop an optimal strategy. Our in-depth optimal strategy tables are in the final pages of the report.
Kail, Jacob T., "Smallball: A Statistical Analysis on the 2016 SEC Baseball Season" (2017). Senior Independent Study Theses. Paper 7507.
Bachelor of Arts
Senior Independent Study Thesis
© Copyright 2017 Jacob T. Kail