Gerrit Cole and curveballs

The Pirates' quiet winter means that if the Pirates are going to get back to the playoffs in 2014, they're going to have to do it by improving internally. The good news is that it's relatively easy to find young players on the Pirates capable of doing that, and the obvious player to start with is Gerrit Cole. I saw a couple of articles around the internet about Cole this week, which got me thinking about Cole and his approach and the way he ended 2013, and since we're still in the middle of the winter with no end (or actual baseball news) in sight, I figured that today would be a good day to do some digging on Cole. 

The first piece I read about Cole this week was this one at Beyond the Box Score, which in turn lead me to this one at FanGraphs. The gist of the two stories is the same, and it's something that you already know if you are a Pirate fan that watched Cole closely in 2013; much of Cole's success down the stretch came because he stopped leaning so heavily on his fastballs and started using his breaking ball, a curve in particular, much more heavily. 

The use of the curveball is interesting for a couple of reasons. The most notable one is that a curveball was not Cole's dominant breaking pitch in college. When the Pirates drafted Cole in 2011, Cole was primarily known for his 100 mph fastball and his dominant slider. The Pirates' method of developing pitchers (focus on fastballs to learn both command and how to throw two-seam fastballs with movement) is pretty well-known to us at this point, so it was no surprise that Cole's slider didn't show up much in the minors in 2012 or 2013. The emergence of a curve this year was a bit surprising, but it probably shouldn't have been given AJ Burnett's nasty knuckle-curve and his influence on the pitching staff. I would even go a step further than that and say that it's likely that the curveball is a team-wide philosophy. In researching this post, I came upon this old FanGraphs post. The post is nominally about about correlationg curveball and slider usage with injury, but the chart at the bottom kind of blew me away. That chart shows the three pitchers in 2011 that threw a higher percentage of curveballs than any other pitcher. First on the list is Wandy Rodriguez, who was not a Pirate in 2011 but became one in 2012. Second on the list is AJ Burnett, who was not a Pirate in 2011 but became one in 2012. Third on the list is Erik Bedard, who was one of the Pirates' main off-season acquisitions beforet he 2011 season. My hunch is that Gerrit Cole learned a curveball last year because the Pirates want him to have a curveball. 

Unrelated sidenote with a conclusion that now seems staggeringly obvious: Edinson Volquez's second favorite pitch is a curveball. We'll have to come back to this later, I suppose. 

Anyway, back on track: reading these stories about Gerrit Cole's newfound curveball domination reminded me of this one little thing from Cole's torrid stretch in September that didn't quite fit. Whenever I watched Gerrit Cole pitch last year, I was never entirely sure if his breaking pitch was a curveball or a slider. Sometimes it looked like one and sometimes it looked like the other, but the two pitches often seemed hard to me to distinguish. I remember joking on Twitter the night of his dominant start against Texas that Cole had broken the pitch identification algorithim on's Gameday, because it had no idea what to do with 90 mph changeups or what to call whatever his breaking ball was that night. If you read my recap of that game, I even mention that I wasn't sure what the pitch was

So what is Gerrit Cole's breaking ball? Let's turn to PitchFX (as always, from Brooks Baseball) to take a closer look. This first chart below is the typical horizontal vs. vertical break chart, with Cole's different pitches color-coded (remember that it's from the catcher's perspective, so positive vertical break just indicates a ball that drops less):

Instead of having a distinct curveball (yellow) and slider (red), Cole's sort of got a continuum that covers both pitches. If the circles weren't colored differently to distinguish pitch type, I'm not sure you'd know where to draw the line between the two pitches on this chart. Compare that to Matt Harvey, who also uses a slider and a curve with regularity: 

Harvey has a much more traditional looking slider and curve profile, in terms of movement. This is not a criticism of Harvey, mind you, because Matt Harvey is awesome. It's just a comparison point for what Cole did in 2013. In general when you talk about sliders and curveballs you talk about them in Matt Harvey terms; a slider is a hard pitch with some late horizontal break while curveballs tend to be slower with more vertical break. What Cole was doing in 2013 was something slightly different. 

Let's bring velocity into the equation now. First, we'll look at velocity (y-axis) plotted against horizontal movement: 

We can make a couple of quick observations here. The first is that the line that the Brooks algorithim is using to identify sliders vs. curveballs is probably velocity based, and the second is that there's not a huge difference in the horizontal movement of Cole's slider vs. his curve. Let's get velocity (y-axis again) plotted against vertical break (which is on the x-axis, which can be a little confusing): 

Now we can see that in addition to the velocity difference, Cole's curveball has a bit more drop on it than the slider, which is obviously something that you'd expect. Let's get one more chart up here before pulling everything together: release point. 

So what do these graphs tell us? Gerrit Cole's four-seam fastball and two-seam fastball tend to clock in with really similar velocities, but Cole's two-seamer bears down and in on righties in a slight, but noticeable fashion. In terms of his breaking ball, he can throw it anywhere from the low 80s to the low 90s. Occasionally the bottom will fall out of it, and occasionally it won't. When he's throwing his changeup well, it comes in at the same speed as his slider, only instead of breaking in on lefties, it bears slightly away from them. All of these pitches come from more or less the exact same release point.

The reason that Cole dominated down the stretch is that his four-seam fastball looks a lot like his two-seam fastball and his two-seamer kind of looks like his changeup and his changeup masks his slider really nicely and it's not really terribly easy to draw a line or distinction between his slider and his curveball, but all five of those pitches are different enough that if you can't find some way to distinguish between them, you probably aren't going to find a way to get a hit off of Cole. Early in the year Cole leaned on his fastballs (I think this was both because the Pirates wanted him throwing fastballs to save his arm, and because he was still getting a handle on the curve/slurve, but that's some conjecture at this point), but when the Pirates unshackled him and the breaking pitches, he suddenly leapt to another level. 

Of course, this leaves me with some questions. I think the obvious one is this: was the "slurve continuum" a result of the on-going process of Cole learning a curveball, or was it an intentional move towards the result we saw in September? What does the Pirates' idealized version of Cole look like? Is it someone like Harvey, with separate-but-dominant sliders and curves, or someone like AJ Burnett and his overwhelming curveball? How, exactly, does this affect Cole's arm? When David Price came out of Vanderbilt he was a bit like a left-handed version of Cole as a college prospect. Over time, the Rays have slowly been transitioning him away from his slider to a curveball, to the point that his curve his is only breaking ball. I've always wondered why that happened, because Price's slider was a nasty pitch in the early part of his career. Was it a health concern or a command concern or a philosophical concern or something that I haven't even considered? I'm wondering the same thing about the Pirates now, both because of the way they've tweaked Cole's breaking pitches and because of the way the organization seems to focus on curveballers. 

I don't have a lot of answers at this point, because these questions don't quite have answers yet. If last September was any indication, though, watching Cole develop and getting answers to those sorts of questions is going to be quite a privilege. 

Important source link: Gerrit Cole's Brooks Baseball player card, which is how I made all these fancy charts. Except the Matt Harvey one, obviously. 

About Pat Lackey

In 2005, I started a WHYGAVS instead of working on organic chemistry homework. Many years later, I've written about baseball and the Pirates for a number of sites all across the internet, but WHYGAVS is still my home. I still haven't finished that O-Chem homework, though.