The Frightening Prospect of a Good Matt Harvey Curveball

Matt Harvey hasn’t been featured in these digital pages since the illustrious David Temple pointed out in early August that the Mets right-hander was throwing again after 2013 Tommy John surgery. With his coming start on Friday against the Tigers in the Grapefruit League, Harvey is about to be pitching again to major league hitters; that is not only cause for much celebration, but also cause for his goodly reintroduction into the halls of our analyses and postulations.

During his bullpen sessions in preparation for his start this week, it has been reported that Harvey has “rediscovered” his curveball. Harvey’s curveball usage has been low (under 15%) since he broke into professional baseball, but the pitch was supposedly something he leaned on quite often in college at UNC, as Harvey referred to himself as a “fastball-curveball” pitcher while playing for the Tar Heels. Referring to getting the curve back, Harvey said:

“’I always threw sliders and I don’t know where this curveball came from, so it’s nice having that develop,’ Harvey said Monday. ‘I don’t know if I figured something out in my mechanics or it just magically appeared, but it’s nice having that and it felt good out there.’”

That statement is a little strange, as Harvey has always thrown a curveball during his tenure in the majors — he just hasn’t thrown it very often. There’s little chance 367 sliders were misclassified as curveballs for Harvey in 2013, as the speed and break are quite different, so either Harvey is just talking, or he actually thinks there is a whole new curve in the mix. Indeed, by the eye test, this example from April of 2013 looks like a fine curveball with a nice amount of drop on it, and was considered as such by a sometimes error-prone robot tasked with the sole job of identifying pitches:

We don’t just rely on the eye test, however. Something might be brewing here, and while we don’t yet have video or PITCHf/x data for a new Matt Harvey curveball, we can figure out just how good his phantom curve was in 2013, and what that might mean moving into this season.

First, the rest of his arsenal (this is important, as we’ll see): by standardized pitch value, Harvey had the most successful fastball in the majors out of qualified starters during 2013 at 1.95 wFB/C, the 24^th-ranked slider (0.88 wSL/C), and the 11^th ranked changeup (1.44 wCH/C). While pitch type linear weights have their limitations, this tells us what we already know: Matt Harvey was really, really effective in 2013, and his pitch offerings were all elite or close to it. Everything fed off of the incredible fastball, and why not — when you have a pitch that you can successfully throw in the middle of the plate whenever you want, everything else gets better, no matter the singular quality of the particular pitch.

With the rest of his arsenal established, how good was Harvey’s curveball in 2013? By standardized run value, it ranked 12^th in 2013, at 1.04 wCB/C. That’s impressive, but his curve’s effectiveness was certainly influenced by how great his other pitches were, which is one of the drawbacks of using pitch values – they don’t capture the fact that pitching is a series of interconnected events. When someone can throw a 97 MPH fastball, it makes sense that their curveball, change, and slider would look like cartoon pitches in comparison right afterward.

For fun, and to put his 2013 curveball into terms we may better understand, let’s see how it compares to the curve of other pitchers. I’ve taken a page out of Jeff’s book and compiled a z-score comparison of starting pitchers who threw at least 100 curveballs in 2013, based on velocity, vertical movement, and horizontal movement.

There’s one problem with this, however: no one threw their curve as hard as Matt Harvey in 2013. It’s not close, either. Gerrit Cole was the nearest in terms of velocity, and he came in .5 MPH under Harvey’s average: 84 to 84.5 MPH. That’s not a tiny difference, and it illustrates just how hard Harvey was throwing in 2013. It’s a power curve, if there ever was one.

With the velocity caveat established, I’ve compiled two different comparisons – one with velocity considered and one based solely on the different curve’s movement profiles. If you didn’t read Jeff’s pitch comparison articles (which you should), the closer the comparison score is to zero, the better match it is, and each increment of 1 is a standard deviation away from Harvey’s curve. As we’ll see, the results are pretty interesting, and not as we might expect them to be. Let’s start with the overall comparison, which includes velocity:

Pitcher	Velocity	Horizontal	Vertical	Rating
Matt Harvey	84.53	1.32	-3.23
Andrew Cashner	81.39	3.29	-3.2	2.3
David Phelps	80.65	2.4	-4.3	2.4
Lucas Harrell	81.95	3	-4.7	2.5
Gerrit Cole	84.03	5.53	-3.68	2.7

Andrew Cashner is the “best” comp, but it’s not at all a good one. There isn’t a great overall comparison, as Harvey threw his curve harder than Barry Zito’s average fastball in 2013, and because the pitch doesn’t move very much (especially horizontally). In fact, Harvey’s curve had the lowest horizontal movement of qualified pitchers by a pretty wide margin. Let’s look at Harvey and Cashner, in that order:

The comp is also a little problematic because Cashner threw his curveball even less often than Harvey did in 2013: 160 of them to Harvey’s 367. It was also a pretty awful pitch, and Cashner almost completely ditched it in 2014, only throwing it 27 times. That’s normally what happens when you only get 16% whiffs on your curveball. We’re dealing with small sample sizes, and we’re also in the territory of players that aren’t really known for their curveballs: it’s strange company for Harvey to be keeping.

Now let’s take a look at just the movement comparison:

Pitcher	Horizontal	Vertical	Rating
Matt Harvey	1.32	-3.23
Ryan Vogelsong	2.15	-4.09	0.8
David Phelps	2.4	-4.3	1.0
Jacob Turner	2.96	-2.91	1.1
Andrew Cashner	3.29	-3.2	1.1

That’s a much better comp, as Harvey is really at the extreme end of the velocity spectrum when it comes to curveballs — that obviously makes comparison inherently difficult.  I’ll just go ahead and say this about the movement comparison table: Ryan Vogelsong is a lot of things – World Series champion, seemingly good guy – but he is not known for his Uncle Charlie. Let’s see Harvey and Vogelsong:

We can see that yes, they both move about the same, with not much drop or horizontal movement. Vogelsong’s curve is obviously much slower and is part of the reason it got hit harder in 2013, with a below-average whiff rate of 21% and poor ground ball rate of 30% In short, Vogelsong’s curveball isn’t one that you say “I want that one!” to.

The bottom line is that Matt Harvey’s curveball doesn’t have a good comparison. No right-handed starter in 2013 threw one that was nearly as hard, and no one threw one that had less horizontal movement. For a pitch that had comparatively little horizontal or vertical movement in respect to good curveballs, it had a slightly above-average 30% whiff rate and yielded above-average grounders (57%). On it’s own, that’s a bit crazy. Considering Harvey’s fastball, changeup, and slider in 2013, it’s not so crazy, and shows that there’s a lot that goes into the effectiveness of a pitch (as we see by the run values).

Currently, all we have is the rumor of Matt Harvey “rediscovering” his curveball and the small sample size of curves from 2013. What we see is a curve that doesn’t do much in terms of movement, yet still was the 12^th most effective curve in the majors by run value during that year. There isn’t a lot of future predictive value in linear weights, but we can see what Harvey accomplished with a curve that didn’t really do much. Right now, we don’t have any idea if Harvey is going to make it back to the overpowering pitcher he was in 2013: elbow surgery forces that terrible question into our heads. However, if Harvey does get back to who he was in 2013, and also improves the movement of his curveball, we don’t need to think too hard to realize the possibilities.

In a vacuum, Matt Harvey threw a curveball in 2013 similar to guys who don’t throw very good curveballs. That was good enough to save over a run per 100 pitches in 2013, because pitching isn’t a vacuum, especially when someone has the potential to throw a fastball like Harvey did in 2013. There’s still a lot we don’t know yet, but the fact that we’re talking about the possibility of a healthy, new-and-improved Matt Harvey means it’s going to be a good year.