Looking for Justin Verlander’s Curveball
Maybe it was all a setup for the headline, but in Anthony Fenech’s piece about Justin Verlander and his effort to return to glory — titled “Tigers’ Verlander ‘way ahead of the curve’ early this spring” — the pitcher points to the current state of his curveball as a sign of early success.
“I’ve seen a pretty dramatic difference,” he said. “The curveball seems to be a lot better already than it was at any point last year.”
Catcher Bryan Holaday, who stood in for the final few pitches of the session, agreed, saying the spin was nice and tight.
Verlander notices the difference, especially in the break, from last year, when his breaking pitches, “Neither one of those pitches was good at all last year. They didn’t have the same bite.”
Much of the previous analysis of Verlander’s poor year focused on his fastball velocity and fading release points. And the pitcher himself referenced those factors a bit when, later in the piece, he admits that he was underweight due to last year’s offseason surgery on his core.
But this might be the first time we’ve heard about the curve missing tightness.
Tight spin is a bit of a blind spot for PITCHf/x, in that spin is not observed, it’s calculated from the movement and velocity. But there is a calculated number for spin in the database, and we can also look at the movement and velocity of his curveballs over time as well. Maybe he was having some … trouble with the curve.
If you switch to the movement tabs, you might be unimpressed. Last year, Verlander’s curve was only about a half inch off when it came to horizontal movement compared to his career values. And his vertical movement and curveball velocity matched his career averages on the dot.
But even that cursory look is flawed — the curve had been better from 2011-2013, averaging seven and a half inches of drop. In 2014, he lost an inch more of drop off that average. It’s hard to see with the naked eye, but on average, his curve didn’t drop as much (2013 on the left, 2014 on the right).
That “little” bit of drop actually meant a great deal to Verlander’s curveball, it seems. That spin he references was off by about 200 revolutions off his average, and off 400 revolutions off his peak in 2011 and 2012.
We might be excused from not seeing this earlier, because his spin rates and outcomes on the pitch do not link up all that well. With the outcomes from Brooks Baseball, it looks like he’s had good spin rates and bad outcomes in the past.
| Horizontal Move | Vertical Move | Velocity | Spin Rate | swSTR% | GB% | |
|---|---|---|---|---|---|---|
| 2008 | 0.81 | -3.18 | 80.3 | 1404.5 | 12.6 | 48.9 |
| 2009 | 5.71 | -5.61 | 80.1 | 1399.5 | 8.5 | 44.7 |
| 2010 | 6.89 | -6.50 | 79.6 | 1632.6 | 6.4 | 59.8 |
| 2011 | 7.67 | -6.79 | 79.4 | 1758.0 | 9.2 | 51.5 |
| 2012 | 6.27 | -8.55 | 79.4 | 1769.5 | 9.3 | 58.0 |
| 2013 | 4.88 | -7.19 | 78.9 | 1516.9 | 11.6 | 55.0 |
| 2014 | 4.55 | -6.35 | 79.2 | 1354.2 | 9.6 | 56.9 |
| Total | 5.43 | -6.30 | 79.6 | 1575.8 | 9.8 | 53.3 |
This shouldn’t be a surprise, actually. There’s no correlation between curveball spin rate and swinging strike rate on the pitch (minimum 100 curves thrown since 2009, n=275, p=.81, r2=.0002). There’s barely a correlation between curve spin rate and ground-ball rate (p< .0015, r2=.0365), but that's not much to hang your hat on. When we looked at all curves, we found a general agreement between curveball drop and ground-ball rate, but Verlander’s grounder rate on the pitch was fine last year.
So that’s the weird thing here. Verlander is correct to say that the spin on his fastball wasn’t as tight last year. But it’s unclear how important that spin rate is. And his curveball has only twice been above-average for whiffs. It’s comfortably above-average for grounders, but it was also last year, when he didn’t have the same drop or spin as he did during that three-year stretch that came before 2014.
Sure, Justin Verlander’s curveball was lost, at least in terms of shape and feel. But in terms of results, it was about the same as it’s ever been. Funny how that works.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
This is good stuff. To my eyes, more often than not last year his curve was just a meatball that hung over the plate and got crushed, especially against right-handed hitters.
The lack of correlation between the spin rate and the whiff rate is interesting, but I’m curious if perhaps the taken strike rate went down? It seems to me that Verlander’s curve has always been more of a pitch to freeze batters than to induce whiffs, so I just wonder if his lower velocity, combined with the curve’s lower spin rate and decreased depth simply made it much easier to either lay off, or square up.
Lower fastball velocity, that is.
It could also mean that the curve spin was inconsistent, which would make it harder to command, no?
Help me out. You note “But in terms of results, it was about the same as it’s ever been. Funny how that works.” Yet his PitchFx pitch values tell a different story. His wCB from 2009 to 2004 goes 2.5, 5.2, 11.2, 11.4, -0.5, -11.7. That’s a pretty big dropoff from the 3rd best CB in MLB in 2012 to the 2nd worst in 2014 (amongst qualified pitchers).
I typically ignore the balls in play stuff, and that’s the big part of the pitch values. So that means it got hit harder when it went into play. Despite having the same whiff rate and ground ball rate as ever, and not much difference in movement. Strange thing.
i get why you’re focused on the outcomes when he throws the curve, but maybe it’s not that pitch he should be worried about.
his change-up and slider were thrown at similar speeds in 2011 and 2012. however, there was a pretty substantial difference in vertical movement. that gap closed toward the end of 2013.
if there’s no significant difference in vertical movement and velocity, those two may more or less act as the same pitch. if batters swing about halfway between where the change-up and the slider would end up, they might be able to put their bat on the ball, regardless.
if getting the spin right on his curveball leads to a better slider (i.e., different vertical movement from the changeup), batters will have to worry about 4 distinct pitches instead of 3 and a half.
Given what he has said about his injury, I’m hoping for a nice bounce back. He has to! Our World Series window is quickly closing! Any thoughts on where he ends up this year stats wise?
I like the article, but I really like the inforgraphic.
Love it.
I never get tired of hearing Justin Verlander identify a weakness. It was a foreign feeling to watch him struggle to resolve his issues last year, he really deserves a good season.
Verlander doesn’t throw a camel, so when he throws the curve high it is lazy. The second gif shows the curve down. It approaches the plate on a flat plane out of his hand–because his extension is better– and takes a later and sharper break. In addition, even the same break (which this is not) looks far greater to the hitter’s eye when it is down. The pitchers who I caught with great curves all had exceptional spin, but one. He had a poor fastball but a deadly curve, thrown in a modified knuckleball grip. It turned over exactly one rotation, the seams were very visible, and it always followed the direction of the rotation, small as it was, and broke sharply. That made no sense to me, but it is what it is.
I’ve got 3 theories:
1) Curve spin rate is correlated to Z-Swing%. You’ll keep your whiff rate up that way, but you’ll fall into worse results because you’ll have more balls batted into play.
2) The z-swing rate is up because it plays off fastball velocity. His curves could look like high fastballs for a while, prompting batters to lay off that 100 mph cheese when he had it. Now that he’s lost it, batters are more aggressive.
3) This is simply a pitch sequencing issue – a body of work we need to examine more closely as a sabermetrics community. Aside from the z-swing% on the pitch, the other big difference was the stark lack of K% on the pitch. All else aside, that points to sequencing to me.
This could largely be a psychological issue for him that affected sequencing and comfort. Even if he was getting similar results on his curve last year, if he was *feeling* like the curve was poor, he’d be a little wary of going to it when the time was right, and avoiding it would affect his other pitches in terms of sequencing and feel.
I buy what he’s saying here. (And, as a Tiger fan, I hope to god he’s right.)