Brandon McCarthy Wants to Change It Up

Brandon McCarthy has famously embraced sabermetrics before. Learning about ground balls and efficiency helped save his career. That doesn’t mean that the struggle to learn and improve doesn’t continue. And, after we talked last week, it seems the pitcher is running up on the limits of sabermetric research. Or at least, he is identifying places where execution trumps theory.

Coming up in the minor leagues, McCarthy “didn’t walk anyone,” and he’s right: his walk rate was close to 4% before 2005, or half league average. He had great strikeout rates then, before he’d tasted the big leagues (29%), but he also gave up home runs (close to one a game). A few stints of struggle in the big leagues — “Without being a plus velocity or stuff guy, once you shrink those margins, everything goes the wrong way” — and McCarthy was looking for a change. His walk rates weren’t translating, his whiff rates weren’t either, and the home runs were leaving the yard.

So McCarthy read some Fire Joe Morgan. And some FanGraphs. He watched Roy Halladay pitch. He decided he needed to work with a good cutter and two-seamer and keep the ball on the ground.

Changing from a fly-ball arsenal to one more ground-ball friendly altered everything. “When I changed things over, I tried to work the mental side with it, but I also had way more confidence in the 2-1 count, of, here’s the sinker, here’s the cutter, swing as hard as you want, I don’t care, my goal is for you to hit it, just not well,” McCarthy said. Of course, McCarthy’s ground-ball rate went up with his sinker usage. But if you cut his career into pre- and post-sinker halves, the pitcher had an 8.9% walk rate before this discovery, and a 4.2% rate after. That’s what he means when he says that he “got comfortable again” with what he was doing.

Though he relishes getting outs with grounders — “However you get the outs, I don’t care. You can give up 27 hits and get them with 27 pickoffs, I don’t care. Just get all the outs you can” — McCarthy admits that strikeouts are important, and that he’s concerned that his strikeout numbers have decreased. “Some days it works, you get through nine innings and 90 pitches and two strikeouts and it feels fine because you’re just getting lots of early contact,” McCarthy said, “but then there are those situations where I need an out and I need it now and this guy needs to be done and I don’t want this guy to put it play, I need to end this.”

And there’s little theory that can help him. He knows what he needs. “The one thing that has bugged me the most and the biggest thing that has held me back is my lack of a changeup,” McCarthy admitted.

For the last two years, McCarthy has been struggling to find consistency with the pitch. He had his old-old changeup. Then he tried a new one. Then, after his latest DL stint, he tried the split-finger. And though that pitch had promise, “it wasn’t mechanically sound, it needed an offseason” the pitcher said. So he scrapped that and went back to the old-old changeup.

He’s admitted before that he uses PITCHf/x to look at his pitches, and so it wasn’t surprising that McCarthy turned to the system to describe his changeups. The ‘old-old changeup’ “keeps getting misclassified as a curveball,” because it goes 78mph and “there isn’t a lot of run, there isn’t a lot of depth.”

Looking at the PITCHf/x for his San Diego start on August 26th, there are three changeups listed, but McCarthy said he threw eight or nine of the pitch. There, in the curveball classifications, there are indeed five pitches that lack the horizontal bite of his curve. Though the pitch didn’t garner any swinging strikes, here’s one of the three that got called strikes.

The changeup is a day-by-day thing right now. “I tell my guys, if you see it in the pen and you like it, then we’ll call it in the game,” McCarthy says. But even then, he advises caution: “Once Robbie Cano is in the box, you can say, let’s throw a changeup, changeup sucks, he goes 600 feet with it.” He laughed when I reminded him that the hypothetical actually happened, but pointed out that they went with the wrong pitch that day. “You can follow scouting reports but success depends on execution,” McCarthy said, before pointing out that a good relationship with the catcher can give him an extra set of eyes on the field.

Some of this anxiety is inherent to the changeup, maybe. “I feel like it’s the most mental pitch of any pitch you throw,” said McCarthy, “It’s the one pitch you truly have to trust the grip. … You’re taking velocity off your fastball, which is scary: if it misses, it’s probably going to get clobbered.” He agreed that organizations that develop good pitching stress the pitch and have all sorts of ways to “force guys over the hump” by encouraging changeup usage. “But you can give me any grip, and if it doesn’t feel right, and I don’t trust it, you can make me throw it 70 times and all 70 times are going to suck,” he said.

And so, for McCarthy to make the next leap in his career, he’ll have to do so on the back of execution. An offseason spent refining the changeup, and maybe he’ll find some of the lost strikeout punch to complement his ground-ball approach. He can use the best available technology to track that process, but in the end it’ll be up to his ability to throw the right changeup.

This realization fits somewhat hand-in-glove with a few other realizations about the short-comings of current statistical research.

For example, when it comes to pitch sequencing, we can’t gather every piece of information from a hitter when they finish a plate appearance. We can know that he swung at a changeup outside the zone, but can we know why? McCarthy wondered: “what did you see there, was that a strike to you, was it because the ball spins the same, or his arm speed is the same, or is it because of something he did earlier in the game to set it up?” There’s pitch-to-pitch sequencing, “but then there’s something that can carry game to game,” he pointed out — Trevor Cahill and Ryan Vogelsong were about to pitch against each other for the second time in two weeks, which is “pretty much one long 18-inning game.”

And then there’s leadership, which McCarthy brought up at this year’s SABR Analytics Conference. “There are guys that are dinner party people, someone that, if you told anyone that he was coming to a dinner party, would anyone complain ‘aw that guy’s coming?’,” says McCarthy of a clique-busting clubhouse leader. But can a team build that without personnel? They have, some with “flawed execution,” but “it’s not hard to mess that up,” McCarthy pointed out — one trade could bring in bad seeds. “You can make a fantastic drink, but one drop of something that tastes terrible ruins the whole thing,” he said.

That said, he knows that there are limitations to focusing on leadership or clubhouse demeanor. Teams can identify guys that can help in the clubhouse, and look at them a little more, but “talent’s still going to be the one that wins out,” he admitted. “It’s still a good thing to add on to a team.”

Being aware of the latest research on baseball helped Brandon McCarthy find a niche in the big leagues. But he can still see challenges ahead that may depend solely on execution. “A big, big way to differentiate yourself from the pack — the guys that can really get strikeouts and *then* do everything else,” McCarthy said of his goals, “Until I get there.. that’s frustrating.” This time, all the numbers can do is help him track his progress as he works on the changeup.