PitchingBot and Stuff+ Pitch Modeling Is Now on FanGraphs!

I’m happy to announce that we now have two pitch quality models, PitchingBot and Stuff+, available for your perusal on the FanGraphs Leaderboards.

PitchingBot is the brainchild of Cameron Grove. We worked with Cameron to be able to run and maintain his model in-house at FanGraphs; he has since joined an MLB front office. You can read all about PitchingBot in the FanGraphs Library here.

In short, PitchingBot takes inputs such as pitcher handedness, batter handedness, strike zone height, count, velocity, spin rate, movement, release point, extension, and location to determine the quality of a pitch, as well as its possible outcomes. Those outcomes are then aggregated and normalized on a 20-80 scouting scale, which is what is displayed on the leaderboards.

There are three models used in PitchingBot: Overall (botOvr in the leaderboards – stuff and location), Stuff (botStf – stuff only) and Command (botCmd – location only). These are available for a pitcher’s entire combined arsenal and for each individual pitch type.

There’s also botxRV100, which is the expected run value per 100 pitches, and botERA, which is that same run value converted into an easy to understand ERA scale.

Eno Sarris and Max Bay created Pitching+, with inspiration from work by Ethan Moore, Harry Pavlidis, and Jeremy Greenhouse, among others. Eno and Owen McGrattan currently maintain and work to improve the model, with engineering support from Matt Dennewitz. Stuff+ has been written about extensively at The Athletic. Additionally, Stuff+ and its corresponding models — Pitching+ and Location+ — are detailed in the FanGraphs Library here.

Stuff+ only looks at the physical characteristics of a pitch, including but not limited to: release point, velocity, vertical and horizontal movement, and spin rate. Generally, the model aims to capture the “nastiest” pitches in baseball. Location+ is a count- and pitch type-adjusted judge of a pitcher’s ability to put pitches in the right place. It ignores a pitch’s physical characteristics and looks at count, pitch type and location. The overall model, Pitching+, is not just a weighted average of Stuff+ and Location+ across a pitcher’s arsenal. Rather, it is a third model that uses the physical characteristics, location, and count of each pitch to try to judge the overall quality of the pitcher’s process. Batter handedness is also included in Pitching+, capturing platoon splits on pitch movements and locations.

Stuff+, Location+, and Pitching+ are all on the familiar “+” scale (like wRC+), with 100 being average.

It’s worth noting that all pitch modeling data can be rolled up to the team level and can also be broken down by custom date ranges.