Appendix B — Historical Notes on PITCHf/x Data
B.1 Introduction
PITCHf/x was a product by Sportvision, a company that produced broadcast effects for sports, such as the first-down virtual line for football and the FoxTrax hockey puck. Two cameras installed in each MLB park recorded the flight of the baseball between the pitcher’s mound and home plate, and advanced software calculated the position, the velocity, and the acceleration of the ball, giving sufficient information to estimate the full trajectory of the ball in its mound-to-plate trip.
From 2006 until fairly recently, Major League Baseball Advanced Media (MLBAM) maintained a publicly accessible Gameday web server that fed their online content with real-time data delivered in an XML format. Previous editions of this book used the pitchRx package to download this pitch-by-pitch information. Unfortunately, this server has been superseded by an API-based system located at https://statsapi.mlb.com/, and thus the pitchRx package no longer works. While this new system is also publicly accessible and delivers real-time data, it is not well documented and proper usage requires registration as a developer. Several developers have published API packages on GitHub, most notably the MLB-StatsAPI package for Python available at https://github.com/toddrob99/MLB-StatsAPI, but to the best of our knowledge no similar R package is in widespread use.
That is the bad news.
The good news is that most, if not all, of the useful data provided by the old Gameday server is available through the Statcast data served via Baseball Savant. Appendix C details the retrieval and use of these data.
In the remainder of this section, we list several resources that were vital in moving sabermetric research fueled by PITCHf/x data forward, mainly for posterity.
B.2 Online Resources
The following PITCHf/x resources were available on the World Wide Web. Note that due to site maintainers being hired by MLB front offices, the demise of PITCHf/x, or exclusive licensing contracts, these resources are subject to being removed or moved.
Mike Fast’s PITCHf/x glossary (https://fastballs.wordpress.com/2007/08/02/glossary-of-the-gameday-pitch-fields/): Detailed explanations for the PITCHf/x fields have been provided by Mike Fast.
Brooks Baseball (https://www.brooksbaseball.net/): Created and maintained by Dan Brooks, its main features are the Player Cards, consisting of tables and charts for every pitcher who has ever played in a ballpark with the PITCHf/x system installed. Tables and charts report information on characteristics of pitches, their usage (including sequencing), and the outcomes they produce. The classification of pitches used at Brooks Baseball is not the MLBAM one, as pitches are classified by Pitch Info LLC. Another useful resource of Brooks Baseball is the PitchFX Tool, which allows site visitors to select one pitcher for one game and obtain a pitch-by-pitch table.
Baseball Prospectus (https://www.baseballprospectus.com/): In its Statistics section, Baseball Prospectus offers PITCHf/x Hitters Profiles, PITCHf/x Pitchers Profiles, PITCHf/x Leaderboards, and PITCHf/x Matchups. The building blocks of these resources come from the previously mentioned Brooks Baseball.
FanGraphs (https://www.fangraphs.com/): FanGraphs has PITCHf/x tables and charts for individual players. For example, pitcher James Shields’s PITCHf/x page is available at https://www.fangraphs.com/pitchfx.aspx?playerid=7059&position=P.
F/X by Texas Leaguers (https://pitchfx.texasleaguers.com/): Allows one to set a time frame and find PITCHf/x pitching or batting data for one particular player. This site includes charts on trajectory and movement, tables on pitch characteristics, and outcomes and pitcher/batter match-ups.
Prof. Alan Nathan’s The Physics of Baseball (http://baseball.physics.illinois.edu/): Contains research on baseball physics and has a section dedicated to pitch tracking using video technology at http://baseball.physics.illinois.edu/pitchtracker.html.
Katron’s MLB Gameday BIP Location (https://katron.org/projects/baseball/hit-location/): Allows one to transpose hit location data of a given ballpark in another ballpark of choice. Keeping in mind all the caveats previously illustrated for batted ball data, it can be used to explore the effect moving to a new team can have on a player’s batting.
Sportvision: Sportvision was the company that has devised the PITCHf/x system. They have since been acquired by SMT.