Pitchf/x Thoughts: Joba Chamberlain

*Administrative Note:  This is going to be a weekly or so column on Fridays about PitchF/X odds and ends.  I'd like to have a name for the column, but I can't think of a good one.  "PitchF/X Thoughts" is the placeholder name...please give me suggestions in the comments.  

It's a common thought that pitcher performance changes when they switch roles from Starter to Reliever, and vice versa.  The thought is that the velocity of a pitcher's pitches goes up as a reliever (as he doesn't need to conserve energy as he might as a starter) and that a reliever thus doesn't need to use as many pitches as a starter. 

With the New York Yankees, this idea is observable and testable in the form of Phil Hughes and Joba Chamberlain.  Both pitchers were known as starters in the minors, and very promising ones at that.  Yet Both have spent at least a season as a reliever and a season as a starter, so I thought  it'd be interesting to look how the role they played on the Yankees affected their stuff.  

In this article, I'm gonna look at Joba in particular (Hughes will be discussed next week).

Table 1:  The Movement and Velocity of the pitches of Joba Chamberlain over the last three years as a reliever and as a starter.  Results as a reliever are highlighted in light blue. 
To Read:
Horizontal Spin Deflection:  The amount in inches that the pitch breaks horizontally due to the spin of the ball.  A Positive Horizontal Spin Deflection means that the pitch goes in on Left-Handed Batters (LHBs) while a Negative Horizontal Spin Deflection means that the pitch goes in on Right-Handed Batters.
Vertical Spin Deflection:  The amount of inches the ball drops/"rises" as compared to how we would expect gravity to make a pitch drop.  So a Fastball with 10 inches of Vertical Spin Deflection "RISES" 10 inches more than it should if gravity was the only force acting on it and a curveball with -10 inches of Vertical Spin Deflection drops 10 inches more than a pitch thrown that is just acted on by gravity.
NOTE:  Table Cells marked with asterisks (*s) shouldn't be taken too seriously due to small sample size.

A couple things are notice-able from this chart:  First, as a starter, Joba relies upon 4 different pitches: A Curve Ball, a Change-Up, A Slider, and a Fastball.  As a reliever, the change-up disappears completely and Joba is a 3 pitch pitcher. 

Second, there does seem to be some velocity difference in his pitches as a reliever compared to a starter....in 2010 as a reliever, all of his pitches appear to be moving faster than they did in previous years as a starter.  In particular, Joba's fastball velocity decreased in 2009 but has picked up again in 2010 now that he's back to being a reliever.  It is important to note however that the biggest increase in velocity for all of his pitches came this year as a reliever.  Thus it's possible the velocity increase is due to him being stronger for this season rather than a conversion of role. 

So do Joba's pitches obtain different results as a reliever compared to as a starter?

Well obviously one result that changes is the loss of Joba's change-up.  Joba as a starter throws the change-up as his fourth pitch against left-handed batters (7.6% of the time) and the pitch is about average for a change-up at getting swinging strikes, just below average at obtaining ground balls, and is definitely below average at landing in a wide strike zone (For an out pitch, that's not necessarily a bad thing, of course).  Thus while it's not a bad weapon to have, and probably not a bad pitch, it's not one which couldn't be replaced if Joba's other pitches could handle the slack.  

Against Left-Handed Batters, that slack is picked up mainly by increased fastball use.  The results of this pitch are seen below:

Table 2:  The Results of Joba's Fastball over the last 3 years
Whiff rate = Swinging strikes/swings
Swing rate = Swings/total pitches
Swinging strike rate = Swinging strikes/total pitches
In Wide Zone percentage = Percentage of pitches located within a wide (two foot) strike zone
RV100 = Run value per 100 pitches (A measure of pitch effectiveness).  Negative Values are GOOD for pitchers, while positive values are BAD. 
RVe100 = Expected run value per 100 pitches (A measure of pitch effectiveness, controlling for luck on certain batted ball types.)

The Results of Joba's fastball use are very odd to say the least and are especially inconsistent.  Against Left-Handed Batters, as a reliever or a starter, Joba's fastball is terrible at getting swinging strikes (though the result in 2010 is merely below average, rather than terrible completely, so that's looking up).  As a reliever however, Joba seems to be able to get ground balls with the pitch, something he's not particularly good at as a starter.  The end result was that in 2010, this pitch was a positive because of the fact that it got a decent amount of ground balls and was no longer completely horrible at missing bats.

Against Right-Handed Batters, the pitch is more normal. In fact in 2008, the pitch was above average at missing bats (Swinging Strikes) by a good bit as a reliever (where it was really really good) AND as a starter.  In 2009 and 2010 on the other hand, the pitch has fallen back to average at missing bats.

However, there is one dramatic change shown this year by Joba with his fastball: It's ability to hit the strike zone. Against Right-Handed bats this year,  the pitch is in the strike zone over 77% of the time, resulting in roughly 25% fewer balls.  Meanwhile, despite this, batters are putting the ball in play against Joba at roughly the same rate as before (The real change is that the number of balls that batters are fouling off has increased).  The end result is that Joba's fastball is much more successful than last year against Right-Handed Batters this year and was a positive this year, whereas it was a negative (worse than average) pitch last year.

But are these changes in the results of the fastball this year clearly linked to Joba being a reliever?  Well, I would guess the improved ground ball rate against Left-Handed Batters is indeed such, but I'm not sure how being a reliever has improved the fastball against right-handed batters...since the main change was his in-zone rate, and that rate was never this high when he was previously a reliever).

Joba's Other Two Pitches:

Joba has two other pitches, a slider and a curve ball.  The Curve-ball is not particularly good, while the slider is amazing, though unlucky this year.  However, there really isn't much of an discernible effect of the role of Joba on the results of these pitches.  Being a reliever or a starter doesn't seem to affect these pitches much at all.

CONCLUSION:

Joba Chamberlain has clearly had more success as a reliever this year than he did as a starter last year.  But well, it's hard to see why the role change has really been the driving force behind his better performance this year (bad-luck notwithstanding).  The main pitch that has changed has been his fastball, which as a reliever is better at getting ground balls from left-handed batters.  But It's hard to see why the fastball's improved accuracy against right-handed batters would be caused by the change in role, especially as that improved accuracy was non-existent when Joba was a reliever in 2008.  In fact, while the fastball was great in 2008 when Joba was a reliever, that was because of completely different results (That was caused by a much better ability to get swinging strikes against right-handed batters and a super high ground ball rate against left-handed batters). So the fact that in 08 and 2010 as a reliever, Joba was a clearly better pitcher seems almost more coincidental than anything.

It should be interesting to watch Joba next year, to see if his fastball results in 2010 can continue next year.  He has managed to get increased velocity on the pitch in 2010 as a reliever...but it's not clear that this is the reason for his improved results. 

Next Week:  I'll take a look at Phil Hughes.