Picking up the pace

Since the All Star Break, Joba Chamberlain has had two fairly impressive starts. Well, impressive when compared to his previous starts. We’re not exactly talking vintage Maddux here.

Here’s his line from his last three starts before the Break:

Date       Opponent    IP  H  R  ER  BB  SO  HR  Pitches  Game Score
June 30    Seattle     5.1 9  3  3   3   4   1   96       39
July 5     Toronto     3.2 9  8  3   1   1   2   86       21
July 10    Los Angeles 4.1 9  5  4   1   4   1   94       30

And his last two:

Date       Opponent    IP  H  R  ER  BB  SO  HR  Pitches  Game Score
July 19    Detroit     6.2 3  1  1   3   8   1   107      69
July 24    Oakland     7.0 2  1  1   3   6   0   100      72

Chamberlain has given up fewer hits, struck out more, and consequently has allowed fewer runs and pitched deeper into the game.

Now there’s nothing that special about these last two starts, and normally it wouldn’t even be worth studying. But reading the recaps after the Oakland game, something caught my eye.

Numerous Yankees were attributing Chamberlain’s success to improving his pacing.

Apparently the team had noticed Joba taking a lot of time between pitches, to the point where it was affecting his concentration. Jorge Posada provided an illustration of just how bad it had become:

“He was wandering. I’d look up and half the time he wasn’t on the mound, especially with men on base.”

Manager Joe Girardi and pitching coach Dave Eiland agreed with that assessment. Girardi stressed the importance of pace, explaining

“It’s understanding what you have to do very quickly. Knowing exactly what you want to do with the baseball and doing it.”

Even the Captain, Derek Jeter, weighed in:

“He did a lot better job of working quick. The last thing you want to do when you’re out there playing defense is have someone working slow. And offensively, pitchers that work quick give you a problem too. His pace of game was a lot better.”

That’s all well and good and sounds completely plausible. But did it really happen?

Luckily for us, the PITCHf/x data gives us the exact moment (well, precise to the second anyway) when each pitch was thrown. If there’s anything to what the Yankees are claiming, we’d expect to see the time between Chamberlain’s pitches go down in his most recent starts.

The methodology I used is pretty simple. Look at each one of Joba’s appearances, count the time between pitches and figure out the average for that appearance. Following in Mike Fast’s footsteps, I’ve thrown out any set of pitches with more than one minute between them under the assumption that some non-pitcher related happened (a coach’s visit, for example).

Let’s start off with a graph showing Joba’s average time between pitches for all his appearances since 2008 (when the pitch time became available).

The red dotted line is the average time between pitches across all measured pitches. For this set of data, it’s just over 26 seconds.

Chamberlain has been all over the map in terms of his pacing. This graph is a little misleading though. Many of his appearances over this time frame were relief appearances, and suffer from small sample size concerns and well as possibly a difference in approach.

Here’s another chart showing the same information, but only for appearances with more than 30 pitches.

Many of the peaks and valleys from the previous graph have disappeared in this view. The line representing his average time between pitches has stayed at roughly 26 seconds.

It is true that Chamberlain had slowed down in June, and has gotten faster in July. Of course the average difference is only somewhere between two and three seconds per pitch. I’m not sure how noticeable that really is on the field. It is five to 10 percent faster, but does it really make that much of a difference to the game play?

One way to find out is to compare Chamberlain’s game score with the time between pitches. Although game scores aren’t that scientific, they do give a pretty good indication of a good start versus a bad start. So if there’s anything to be found, game scores might give an indication.

It doesn’t look like much of relationship, does it? He has both good and bad starts at both ends of the time spectrum. The red line represents the best-fit equation, which tries to mathematically describe the relationship. As you can see from the R² value in the upper right corner, the best-fit line doesn’t do a very good job of explaining the data.

Jeter gave us a couple of other places to explore for a noticeable effect on the game. He suggested that the fielders might play better when a pitcher pitches faster, and that batters tend not to like pitchers who work quickly.

Let’s consider the fielders first. This isn’t new ground. In his article “Short work,” Mike Fast looked for any effects of “bored” fielders, breaking it down a variety of ways. He concluded an effect might exist, but it is small, and visible mostly at the extremes. I’m not going to rehash his work here. Instead, I’m going to see if Joba breaks away from the norm and demonstrates some noticeable effect.

This next graph shows the relationship between the time Joba takes between pitches and the defensive efficiency (DER) of the fielders behind him. I limited my data to only those appearances in which Chamberlain allowed at least 10 balls in play.

There are definitely sample size concerns here, so it’s impossible to make any firm conclusions, but there’s nothing that looks very promising. Some of the worst fielding performances happened when Joba took the most time between pitches, but some of the others happened when he pitched his quickest.

It doesn’t look very likely that the fielders perform any better or worse for Chamberlain depending on how quickly he gets back to the rubber and releases the next pitch.

What about batters? How is their performance affected by Chamberlain pitching faster?

I’ll be honest and say I’m not entirely certain where to look. What stats would indicate that batters have trouble with pitchers who pitch fast? It would have helped if Jeter had been a little more clear as to why pitchers who work fast cause problems for batters. Oh well.

Also, it’s very tough to disentangle the batter’s performance from Chamberlain’s. Anything indication of something going on could be the result of Joba pacing himself better, or it could be the result of Chamberlain just plain pitching better. So it’s almost certain we won’t be able to come to any sort of conclusion on this one.

A couple of places to look might be swing rate and contact rate. If a batter is a little out of sorts he might get a little jumpy and swing at more pitches. Of those pitches he swings at, perhaps he misses a higher percentage of them. Of course I could see it working the other way too. A batter who’s thrown off his game by the pitcher’s pacing might not feel ready to swing and take more pitches. Let’s see if there’s a noticeable difference when Joba is pitching.

First, we’ll look at the percentage of pitches a batter swings at for different levels of pacing.

Again, there’s really not a whole lot there. Definitely nothing to indicate that batters are more or less likely to swing based on the time Joba takes between pitches.

What about contact rate? Are batters swinging and missing more when Chamberlain works faster?

Still a whole lot of nothing in the data.

There’s absolutely no indication from these two measures that batters are at all affected by how much time a pitcher takes between his pitches.

We haven’t found a lot of evidence to support the Yankees’ theory that Joba Chamberlain is a better pitcher when he takes less time between pitches. We did see that he has spent less time in between pitches over his last few starts, especially compared to his starts just before the All Star break. But the total amount of time over 100 pitches is something like five minutes, which probably isn’t all that noticeable.

We looked at his overall performance in his starts, as measured by game score, and found little to no relationship to his pacing there. We also struck out while examining Derek Jeter’s claims that it’s better for the defense and harder on the hitters when a pitcher speeds up his game.

It’s getting pretty hard to lend any credence to the idea that a pitcher’s pacing has anything to do with his performance in a game.

There are some reasons to think we might not have the whole story though. First off, we’re working with a very small sample size – a handful of pitches in a handful of appearances. Any effects might be hidden by random noise in our data set.

Second, our measurement tool might not accurately affect Chamberlain’s pace. The batter may step in and out of the box causing a lot of time to pass between pitches even if Chamberlain is ready to throw. Also, Posada mentioned that Chamberlain was often wandering around the mound between pitches. Perhaps the improvement noticed by the team is mental focus and faster preparation as opposed to faster delivery.

Finally, maybe we’re not looking at the right metrics to detect an improvement. Maybe overall defensive efficiency isn’t the right measure because we should only be looking at tough plays. Or maybe batters aren’t more likely to swing at bad pitches, they’re instead more likely to hit the ball weakly.

So there are definitely some further avenues to explore. But based on our initial look at this, I’m going to have to call this one busted.