Julio Urías, (Breaking Ball) Fusion Scientist

Julio Urías is only 24, but it feels like he’s been in the big leagues for a decade. Called to the majors at only 19 in the 2016 season, he’s been a part of the Dodgers’ future and present for a half-decade. When you start that young, much of your development happens at the major league level. In Urías’ case, that means all kinds of changes. Today, though, I want to focus on one: a curveball that has shape-shifted over time before arriving at a tremendously interesting final form.

When Urías came up, he threw a curve with two-plane break, something between a curve and a slurve. As you can see on our handy Pitch Type Splits, it featured 7.4 inches of horizontal break and only 2.9 inches of drop. In his next three seasons, all injury-affected, he turned the pitch into more of a classic curve — more drop than horizontal movement. 2020 saw a return to his original curveball shape. 2021? Well, it’s weird:

Is it a return to his old form? Is it an acceleration of his old form? Is it something else entirely? Let’s delve too deeply into some gifs and math and find out.

First, take a look at a curveball from 2019, the last year that he threw a “conventional” curve with more drop than run:

It’s a pretty pitch to look at, at least to Kevin Newman and me. Newman, in fact, liked watching it so much that he let it drop in for a middle-middle first pitch strike. That pitch wasn’t one I selected at random; more than half of the curves Urías threw in 2019 were on the first pitch of a plate appearance. He used it to get ahead, then went to it sparingly later in counts.

In 2020, that usage pattern changed. He still used the pitch to start batters off — more frequently, in fact. Here’s one of those:

You don’t need that table from up above to tell you that the shape is different. This curve is more slurve-y than the previous year’s model, left-to-right break with some drop mixed in. With that new shape, he used the pitch more in every count — though he faced 100 fewer batters in 2020 than 2019, he threw five times as many curves on 0-2, 15 times as many curves on 1-2, five times as many on 2-2, and so on.

At the same time, he drastically curtailed his slider usage. Given that the new curve behaves, well, slider-ishly, here’s a thesis: Urías blended his slider and curve and used the new pitch in counts where either pitch would previously have been appropriate. The net effect, from a linear pitch weights standpoint, was positive: he was 0.7 runs below average with his two breaking balls combined in 2019, then 2.2 runs above average with them in 2020.

You could, of course, advance an alternate theory: maybe pitch classification simply merged the two pitches, and Urías wasn’t really blending them at all. More curves, with more average horizontal break and faster average velocity? Sounds like half curves and half sliders to me.

But that’s not the case. Per Baseball Savant, he didn’t throw a single curveball in 2020 with fewer than 10 inches of horizontal break, as measured by their PFX_X calculation. In 2019, 85% of his curveballs had less than 10 inches of horizontal break. This wasn’t a statistical merge of two unchanged pitches; it was a new hybrid pitch that combined their characteristics, something Urías confirmed this spring.

This year, things have taken a further turn — a sideways turn. That “curveball” now basically looks like an 81 mph slider. Here, take a look at this 0-1 bender he threw to Chris Owings:

The movement is wholly left-to-right, a slider if I’ve ever seen one. And though we’re still classifying the pitch as a curveball, consider this: Zack Greinke throws a slider with roughly the same average velocity as Urías’ curveball. Greinke’s slider has roughly average drop. It also drops more than Urías’ curve, while breaking less horizontally. In other words, this new pitch is behaving like a slider, whatever he wants to call it.

There’s some evidence that the slider-y bent of the pitch has paid off. Imagine dividing his curveball into two halves: curves in the upper half of vertical movement and curves in the bottom half of vertical movement. In an admittedly small sample (32 pitches per half), batters are swinging and missing twice as frequently in the upper half. When they swing, they’re coming up empty twice as frequently as well — in other words, they swing at the two halves equally often but do much worse hitting the one that break down less.

What does that look like to the naked eye? Eh, not much, to be honest with you. Here’s the highest-positive-break curve that drew a whiff:

Got it? Okay, here’s the one that broke down the most while coaxing a swing and miss:

Those, uh, look pretty similar to me. I’ll grant that Ryan McMahon swung over the second pitch while Raimel Tapia swung under the first, but Rockies batters missing pitched baseballs is hardly evidence of anything given their struggles this year. Their vertical break (excluding gravity) differed by about three inches. That’s not a huge gap. Maybe the fact that Urías is getting slightly better results when giving it less drop is just a sample size artifact rather than anything unique about pitch shape. The overall pitch hasn’t been gangbusters, either; it’s drawing whiffs at a lower rate than it did last year and rarely finishing batters off.

That said, it’s far too soon to draw any conclusions about new pitches, and there are positive signs too. Batters haven’t been able to make hard contact on the pitch at all, so he’s been able to use it in the zone; he’s throwing it for a strike 57.8% of the time, a huge rate for a breaking ball. It also comes out of his hand with roughly the opposite spin of his other two pitches, a four-seam fastball and a changeup, which helps the whole package play up. The individual numbers might not jump off the page, but the three pitches tie together nicely.

More important than any early-season pitch-level results, Urías seems to trust the pitch. After throwing only 18% curves in his first start, he’s used it 27% of the time in his most recent two starts. He’s already thrown it a whopping 17 times on 0-1, a key count when it comes to turning the at-bat in his favor. It’s a well-suited pitch for that; as I mentioned, he commands it well and peppers the strike zone. Down 0-1? It’s your choice whether you’d prefer to watch a breaking ball for a strike or take a hack at a pitch that no one does any damage on.

For now, it’s more a curiosity than anything else. Julio Urías throws a “curveball” that looks like a frisbee of a slider. He’s used it to replace both his slider and his curveball, and he’s taken up the old curveball usage pattern — early in counts to steal strikes — while also using it in counts where he previously favored sliders. Pitchers love adding new pitches. For Urías, subtracting and simplifying is the new rage.