We return to our best/worst miniseries on the best and worst pitches of 2013 after a bit of a delay. As such, it's only fitting we come back to discuss baseball's delayed pitch, the changeup. Similar to what we have done with four seam fastballs and sliders, we will use FanGraphs data on pitch type linear weights (for changeups, wCH/C) as well as vertical movement (CH-X), and horizontal movement (CH-Z) data to create a z-score for each in order to compare the variables to one another. A summary score -- the Sum Z Score -- will then be calculated to give a final, single value for measuring the value of a given pitcher's changeup. The ten percent criteria will also be used, with only pitchers who used the changeup for at least ten percent of their pitches being considered for our best./worst list. However, we will throw a wrinkle into the methods for determining best changeup by departing from our previous way of handling pitch velocity by taking a z-score based on the difference between the velocities of a pitcher's changeup and his fastball, rather than just from the average changeup velocity. This is done with the idea being that a good changeup isn't necessarily one that has good absolute velocity, but one that has significant velocity difference in its velocity compared to a given pitcher's fastball.
With that methodological quirk pointed out, let's have a quick look at the lay of the land, only looking at at linear weights; these are the best and worst changeups of 2013, broken down by starters (SP) and relievers (RP):
|SP, First||Justin Masterson||9.87|
|SP, Worst||Edwin Jackson||-7.84|
|RP, First||Jason Grilli||13.25|
|RP, Worst||Drew Smyly||-71.35|
Like our other sections of the series, we know that this isn't actually the truth; digging deeper, we find that the top changeups -- Masterson and Grilli -- aren't even thrown one percent of the time for each (0.1 and 0.2 percent of each pitcher's pitches thrown, respectively). Include EJax throwing his changeup 1.8 percent of the time and Smyly at 3.2 percent and we see the ground truth is probably not reflected in this table.
OK, so let's add that ten percent criteria and then work our z-score magic -- who survives the cut, so to speak?
|Name||Sum Z Score||Pct Use|
|SP, First||Hyun-Jin Ryu||4.75||22.3|
|SP, Worst||Tim Lincecum||-3.44||23.5|
|RP, First||Cesar Ramos||3.14||16.7|
|RP, Worst||Jared Burton||-2.23||32.8|
With our summed z-scores to go along with pitch usage data, we find that Los Angeles Dodgers lefty Hyun-Jin Ryu takes the changeup crown for the starters, with Tampa Bay Rays Cesar Ramos doing the honours for the bullpen.
For those interested, here are the components of the summed scores; the final column is the z-score for the fastball-changeup velocity difference, while the other columns should be familiar:
...and a little supporting data in the form of changeup swing, contact, and swinging strike rates for our quartet, compared against their fastballs (or in Ramos' case, their sinker):
Here, we find that in terms of swings and contact rates, there isn't a whole lot of difference between the best and worst changeups; each gets more swings in and out of the zone and more swings and misses compared to their heaters. Burton is probably the oddball out of these four, in that he uses his changeup more than his fastball -- a 32.8% changeup rate compared to a 28.7% fastball rate.
One last table, this one looking at OPS against for the four and their changeup and fastball in 2013:
Once again, we see the fickleness of the pitch. While all four have respectable OPS splits against their fastball-changeup combos, we also see how little difference there is between good and so-so changeups, courtesy of our methods and the inclusion of some surrogates for what might be considered talent in the form of movement and velocity differences. Overall, we see that in 2013, the changeup is much like a pizza -- even when it's bad, it's still pretty good, at least in terms of getting hitters out.