The Impact of Pitch Level Tracking Data and Catcher Receiving Skills on Strike Calls
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractFrom the full season debut of PITCHf/x in 2007, advanced radar and optical tracking systems have significantly changed the way that Major League Baseball players have been evaluated. With that information, managers and front offices have ample amounts of information to make decisions in all aspects of the game. While baseball largely is focused on the batter and pitcher matchup, there is one aspect of a pitch where neither the batter nor pitcher have control, the decision of the umpire to call a ball or strike. In this thesis, we investigate the impact of the catcher on a pitch call. Our input variables are a variety of pitch information including pitch velocity, spin rate, and movement. To start, a baseline strike called probability is calculated using a logistic general additive model using the pitch location, batter hand, and pitcher hand. Using this information, we isolate the edge of the strike zone where the umpire decision is most uncertain. Using a generalized mixed effects binary model, we analyze the impact of the baseline strike called probability using pitch data as fixed effects and catcher and umpire as random effects. Originally, our pitch level data included a categorical pitch type variable, which classified the data set’s fourteen pitch types into fastballs, sliders, curveballs, and changeups. Collinearity issues of pitch type against pitch velocity, spin rate, horizontal movement, and vertical movement led us to remove these categorical variables to create a simpler but still effective model. From this model, each pitch level variable has a statistically significant impact on the probability of a strike call, including a negative relationship with pitch velocity. Catchers and umpires have similar impacts on the probability of a called strike. Lastly, this model was assessed on pitches that land at each of the four edges to investigate differences in model estimates. Horizontal break changes with each side and the top of the zone leads to higher variability in umpire performance while the bottom has the opposite effect. The limitations of the methods re- strict how well we can understand the scale of the individual performances of each catcher and umpire. For future research, I would consider alternative and additional methodologies and explore other situational factors like score and count.
Degree ProgramGraduate College