Will Joba Chamberlain make a good starter?by David Gassko
June 07, 2008
When a man posts a 0.38 ERA in his first major league season, people tend to notice; when that man is a New York Yankee, he becomes an overnight superstar. That is exactly what happened to Joba Chamberlain last season, who, after an early August call-up, appeared in 19 games, allowing just one earned run and racking up 34 strikeouts to just six walks.
Already a top prospect, Chamberlain’s performance cemented him as the next big Yankees star, and during the offseason, the Yankees repeatedly stressed that even though he had been startlingly successful in the bullpen, Chamberlain’s ultimate role with the team would be as a starter. Even so, come opening day Chamberlain had reprised his role as Mariano Rivera’s super setup man.
On Tuesday, that all changed. After pitching out of the bullpen the first two months of the season, Chamberlain made his first appearance this season as a starter, giving up two runs (one earned) in 2.1 innings and striking out three hitters while walking four. It was only 62 pitches, but that start has been the biggest topic of conversation among most baseball fans since.
Are the Yankees right to try to move Chamberlain to the rotation in the middle of the season like this? Will he continue to struggle, or will he justify the move by pitching like an ace? While these questions are difficult to answer, I thought I would take a look at the Pitch f/x data on Chamberlain this season to see how his stuff on Tuesday compared to what he’s been throwing all season.
First, let’s answer the most basic question there is: What does Chamberlain throw? Chamberlain has three pitches: A mid-90s fastball, a mid-80s slider and a high-70s curveball, and you can see a graph of how each of those pitches moves to the right. What’s most notable about this graph is the vertical movement on Chamberlain’s pitches.
Chamberlain’s fastball comes in 10.8 inches higher than a theoretical spin-less pitch thrown at the same speed—almost two inches more than the average fastball. His curve, on the other hand, has more than four extra inches of drop to it in comparison to the average curveball. His slider does the same: Whereas the average slider has about 4 inches of vertical movement on it, Chamberlain’s slider actually drops a bit more in comparison to that theoretical spin-less pitch thrown at the same speed. The nasty downward movement makes Chamberlain’s slider basically unhittable.
As a reliever, Chamberlain throws his fastball 68 percent of the time, his slider 25 percent of the time and his curveball the other seven. That mix didn’t change too much Tuesday, though he did throw a few more fastballs (3.7, to be exact) and a few less curveballs than usual.
Despite the mediocre numbers, Chamberlain was actually the same unhittable pitcher in his first start as he has been all season. He struck out three in 2.1 innings and allowed just one hit. Those four walks, however, were a killer, so let’s take a look at where Chamberlain located his 62 pitches on Tuesday, from the catcher’s perspective.
As you can see, Chamberlain’s pitches were really kind of all over the place. Twenty-nine of his pitches were outside of the strike zone—that’s 47 percent—and 48 percent of his pitches were called balls, compared to a league average of 38 percent. There isn’t really any surefire way of knowing if Chamberlain’s control suffered because he was forced to go longer or just due to random luck, but we can try to break the results down by pitch to see what exactly was missing the strike zone.
Here are the outcomes of Chamberlain’s pitches as a reliever this season:
pitch_type Ball Strike In-Play Curveball 0.47 0.43 0.10 Fastball 0.40 0.46 0.14 Slider 0.35 0.56 0.10
And here are the same numbers for Tuesday:
pitch_type Ball Strike In-Play Curveball 0.50 0.50 0.00 Fastball 0.48 0.41 0.11 Slider 0.50 0.50 0.00
Though the sample sizes are tenuously small (two curveballs, 14 sliders, and 46 fastballs), the results for the latter two categories are at least intriguing. Chamberlain’s fastball was called a ball 20 percent more often than usual, and his slider was a ball 43 percent more. Both those results are a bit more than a standard deviation away from our expectation—not a statistically significant difference by any means, but perhaps something to watch out for.
Maybe we can learn more about why Chamberlain’s pitches were missing so much by comparing their basic properties (speed, horizontal movement, vertical movement) to how Chamberlain’s pitches acted while he was pitching out of the pen.
I’ve created a graph to the right showing how each of Chamberlain’s pitches acted in relief (red) and as a starter (blue). As you can see, each of them seems to have lost some vertical and horizontal movement, which frankly makes me a bit suspicious as to how accurately the Yankee Stadium Pitch f/x camera is placed.
At the same time, there shouldn’t be a problem with the pitch speed data, and indeed we see that Chamberlain’s fastball and slider were coming in slower than usual, with about a .7 mile-per-hour deviation for the fastball and .4 for the slider. That’s to be expected when a pitcher is forced to go from throwing 20 pitches and game to 60 (and eventually, of course, around 100).
In fact, if we look at Chamberlain’s first 10 fastballs of the game and his last 10, we see a drop in pitch speed from 95.5 mph (about .4 mph less than his fastball as a reliever) to 94.9 mph (a full mile per hour worse). So while Chamberlain may have tried to pace himself from the get-go, he still seems to have tired a bit as the game went on. And while all these differences may seem small, if we graph Chamberlain’s distribution of fastball by speed as a starter and as a reliever, I think its importance becomes much clearer:
As you can see, while Chamberlain throws more fastballs at 96 mph than any other in speed when coming out of the pen, in his start, Chamberlain’s most common velocity was 94—still very good, but not quite the same. Indeed, a full 39 percent of Chamberlain’s pitches came in at 94 mph or less on Tuesday, versus just 21 percent in all his previous appearances this season.
Tom Tango suggested after the start that Chamberlain may have showed the ability to dial it up when needed, noting that he thought the variation in Chamberlain’s pitch speed was greater than he would have expected. However, if we look at all of Chamberlain’s fastballs, we find the standard deviation of his velocities on Tuesday was 1.74 mph, which is not statistically distinguishable from his standard deviation as a reliever, which is 1.64.
So, with all that said, what have we learned about Chamberlain’s abilities as a starter? Honestly, not much. His pitches came in a little slower on Tuesday, he threw his fastball a little more often and his control was definitely off.
Still, it’s difficult to make any concrete conclusions from a sample size of 62 pitches, so for now—since Chamberlain’s start was far from devastatingly bad—my best guess would be that he will fare just fine as a starter. His ERA will probably jump about a run, as is normal when a pitcher goes from the bullpen to the rotation, but even with a run added to his ERA, Chamberlain is exactly the kind of starting pitcher the Yankees sorely need.
David Gassko is a former consultant to a major league team. He welcomes comments via e-mail.