Stuff: What I've Learned About Pitch Quality From Creating a Burner Account on Twitter (Part 1)
And what Nick Lodolo's curveball can teach us about the process
I’d like to preface this column by saying a couple of things:
Almost everything I’ve learned about Pitch Quality models has been secondhand from listening to the incredible Rates & Barrels podcast and following the show’s hosts, Derek VanRiper & Eno Sarris, on Twitter; or by tinkering around with Cameron Grove’s Pitching_Bot model on his website
My understanding of this topic is by no means perfect or complete, but I feel more confident in what I’ve learned thanks to extremely helpful & insightful Twitter interactions with Eno, Cameron, and DVR
Stuff is like lightning, Command is a bottle.
If you’d like, you can stop right now, avoid a long and wordy column and fully understand what this image means:
However, if you’re like me, this image may have piqued your interest.
You may have found yourself asking, “wait, what exactly is Stuff?” or, “who/what is the Pitching_Bot?” and possibly, “oh man, can I afford to learn deeply about a new baseball metric and not have it impact my social life and ability to talk about baseball in a normal, coherent way with the people I love and respect?”
Lucky for you, in this column, I’m going to try my best to answer the first two questions. The third one… I’m still trying to find out, myself.
Part A: What Exactly is Stuff?
Great question. Stuff is exactly what baseball scouts have been trying to define for the past 140 years.
Stuff is what makes batters flail at pitches that bounce a foot in front of home plate or trip over themselves swinging at a pitch in the opposite batter’s box. It’s what creates moments that leave you muttering to the person next to you, “I can’t believe that bozo swung at that.”
In simpler terms, though, Stuff is deception.
For example, the Cincinnati Reds’ young and up-and-coming left-hander, Nick Lodolo, throws a nasty curveball. He’s made everyone from grizzled vets to up-and-coming stars look like fools trying to make contact with the pitch.
But what, exactly, makes Lodolo’s curveball so nasty?
A quick jump over to Baseball Savant will tell you that Lodolo’s curveball has 3.5 more inches of horizontal break & 8.9 less inches of vertical break than a normal curveball:
…all while travelling significantly faster than almost every other curveball in the game:
If you plug all of those variables (along with a few other ones that are a little more difficult to define) into the Pitching_Bot, it will tell you that Nick Lodolo has a truly one-of-a-kind, exceptional curveball:
Part B: Great, but who/what is the @Pitching_Bot?
The @Pitching_Bot is a Pitch Quality model created by Cameron Grove, a PhD student who has publicly posted his model on a website that is free to anyone to use.
Cameron’s model takes the pitch characteristic data of every pitch thrown in Major League Baseball and grades each one based on how successful it should or will be at suppressing runs. His model even spans over time and can be used to look at any pitcher’s individual pitch grades going all the way back to 2015.
But Cameron’s model isn’t alone in the Pitch Quality world.
Eno Sarris — the patron saint of Twitter analytics junkies — has also helped create a Pitch Quality model of his own, named Pitching+. Unsurprisingly, his model loves Nick Lodolo’s curveball too.
While it’s my understanding that Cameron & Eno’s models differ in some ways, it’s important to note that their processes lead to a common, shared observation: “Seriously, Nick Lodolo’s Stuff is nasty.”
Part C: So what does all that actually mean?
This is the part where it’s probably better if I defer to the experts. Cameron & Eno have a much better understanding of the exact intricacies of their models — and why, specifically, one pitch may grade out better than others.
But, there are some general tropes that Cameron & Eno have preached about their models:
Generally, increased velocity is better for all pitches
Four-seam fastballs benefit from increased spin
And — most critically in the case of Nick Lodolo’s Curveball — Stuff ratings for off-speed pitches rely on the pitch’s relationship to a pitcher’s primary fastball
Let’s pause there.
Pitching itself is a bit of a delicate balance between over-powering hitters and disrupting their timing. Each individual pitch in a pitcher’s arsenal plays its own unique role in either one or the other.
Fastballs travel to the plate quicker, while offspeed pitches travel slower, but move drastically more. The batter has the luxury of having to guess which one of these pitches is coming his way.
In Nick Lodolo’s case, because his curveball drops significantly less than an average MLB pitcher’s, it shares a beneficial relationship with his primary fastball — a 94mph sinker that drops significantly more.
These two pitches are tailored in a way that makes them looks vertically more similar, despite having polar opposite horizontal movement — a relationship probably best exemplified in this PitchingNinja overlay:
Lodolo’s curve & sinker remain on an almost identical plane as they approach the plate. However, at the very last second, his sinker breaks outward to the middle of the plate, while his curveball falls of the table and darts for the batter’s backfoot.
That’s Stuff.
While I can’t claim that every good Stuff grade that I’ve seen is as straightforward as Nick Lodolo’s curveball and sinker, I have found that Cameron & Eno’s work has drastically changed the way I’ve thought about pitching.
Instead of basing opinions strictly on a pitcher’s results, these models allow you to analyze a pitcher’s process.
At a time where there has never been more readily available data on every moment that takes place in a baseball game, these Pitch Quality models provide an essential service — the ability to see the forest through the trees in a never-ending (and endlessly growing) thicket of data.
Eno & Cameron, hopefully I didn’t butcher my attempt to make sense of your models. Everyone else, I appreciate you following along and I hope this made somewhat reasonable sense.
If you have more questions after reading this than you did when you started… welcome to the club. But, feel free to send them my way @RedsInFour on Twiter; I always love talking through these things (when I’m not rambling about potential Reds free agent signings).
Stay tuned for Part II, where I’ll try to explain how Stuff relates to the second part of the Pitching_Bot screenshots: Command!
@RedsInFour