San Francisco Giants v Pittsburgh Pirates

There isn’t just one way to have a great change piece. Some tumble straight down, some are more about having arm-side fade. Some get grounders, some get whiffs. Some are hard. Some disappear.

But it’s clear that you need to do one of these things to have a good changeup. Given the ubiquity of the slider, there are plenty of pitchers that have been told they need a slider to avoid the dreaded platoon split that might send them to the bullpen. Dan Straily was basically forced to try 17 changeup grips. Justin Masterson told me last week that “he’s tried every single one” and lamented the one grip that worked in 2007 for half a season and then disappeared forever.

So there’s a class of pitchers that are trying to ply their trade with iffy changeups. We want to know, as fantasy players, which ones are likely to succeed.

Now, thanks to a talk Harry Pavlidis of Brooks Baseball gave at the most recent SABR Seminar in Boston, we have some tools in our search.

The old school idea is that you need ten miles per hour difference between your fastball and changeup, as A.J. Burnett referenced in our conversation this week. But that hasn’t really been shown in the numbers, and that might be because of a key finding in Pavlidis’ report: changeups are good for either changeups *or* whiffs, and depending on your goal, you need to take a different approach. Here are some key findings from his presentation.

Whiffs on Changeups:

  • Swing rate on changeups had the strongest correlation with whiff rate (whiffs per swing) — more swings, more whiffs
  • Next up were a litany of gaps in vertical movement and angle at home plate — bigger gaps, more whiffs.
  • Faster fastballs lead to more whiffs on the changeup — throw faster fastballs
  • The bigger the gap between fastball speed and changeup speed, the bettter — separate the speeds
  • Changeup speed itself had a near zero correlation with whiffs.

Ground Balls on Changeups:

  • Changeup speed had the strongest (positive) correlation with ground ball rate — firm it up
  • The smaller the gap between fastball and changeup speed the better – keep the speeds together

So what we’re looking for is for a changeup to be decisive. It needs to either go for whiffs — and therefore be way slower than the fastball — or it needs to go for grounders — and be about as fast as the fastball. You can see now why a ten mile per hour gap isn’t that important overall. And why this is tough to turn into a prescriptive stat or leaderboard approach.

But we can look at some specific cases fairly easily. In order to find our test cases, let’s look at the starters under 25 with less than 200 innings (more than 20) over the last three years that throw the changeup less than 10% of the time. That seems like a young pitcher with an iffy changeup, no?

The list is 33 pitchers long, so we won’t reproduce it in it’s entirety, but it pretty much nails our target audience. Tyson Ross, Chris Archer, Garrett Richards, Jarred Cosart, Shelby Miller, Michael Pineda, Jacob Turner, Julio Teheran and Jose Fernandez are the names that leap off the page. All of them would be better off with a strong changeup, although the curveball guys (Cosart and Miller, in particular) might survive and even thrive without.

Sort the list by fastball velocity minus changeup velocity, and the top five include Chris Archer, Jarred Cosart, Wily Peralta, Julio Teheran and Garrett Richards. That’s good news for them if they use the pitch for whiffs. Archer deserves an entire writeup, or even a whole series — he has the biggest separation on this velocities, gets batters to swing at the changeup often, and yet he gets below-average whiffs on it. This analysis might suggest his changeup is primed for improvement. Cosart, on the other hand, doesn’t get nearly as many swings, has three inches less vertical movement, and gets even fewer whiffs. he’s got the separation, but he’s missing the rest of the package. Here are the key percentages for this top five in table form (average whiff rate is around 15%).

Pitcher CH% FBv-CHv CH Vert CH Swing% CH Whiff%
Chris Archer 8.3% 11.3 8.3 50.6% 9.6%
Jarred Cosart 7.1% 11 5.9 39.2% 7.8%
Wily Peralta 7.4% 10.6 8.4 39.7% 10.3%
Julio Teheran 6.8% 9.6 4.9 42.2% 14.6%
Garrett Richards 4.2% 8.6 6.2 40.0% 5.0%

Now let’s look at the ground-ball conditions. The average ground-ball rate for a changeup is around 49%.

Name CH% CHv FBv-CHv CH Vert CH GB% CH Whiff%
Gerrit Cole 6.9% 87.9 8.2 2 38.1% 20.0%
Kevin Gausman 4.6% 87.7 7.7 4.6 33.3% 22.8%
Chad Bettis 8.7% 87.7 5 0.9 66.4% 13.9%
Michael Pineda 6.3% 87.6 7.1 4 53.8% 6.4%
Zack Wheeler 3.4% 86.7 7.8 6 22.2% 11.4%

You’ll see that these guys generally have harder, straighter changeups. Other than Kevin Gausman, who pairs high nineties heat with a hard changeup that drops as much as some tumbling changeups. He already gets great whiffs and has the potential to get a great grounder rate with his changeup, and that’s part of why he has so much promise. It’s absolutely a great time to buy low on him in dynasty leagues.

I’m glad Michael Pineada showed up on one of these lists because he’s been working on a changeup forever, seemingly. And it doesn’t look sweet. But it does tick the ground-ball box. If he can use it to get grounders from lefties while striking righties out with the slider, he could be as great as it seemed he might be in Seattle. Provided he’s healthy again.

Chad Bettis’ changeup may not have the look, the drop of a typical changeup, but it gets the job done, at least so far, as a hard change good for ground balls. Zack Wheeler’s changeup? Could go either way, but it certainly has some interesting statistical aspects. Might be one of the few hard changeups that gets whiffs and grounders eventually. Like Stephen Strasburg‘s changeup. Well, if it really had that sort of promise, he’d probably use it more than three percent of the time. So Strasburg-lite?

Ed. Note: Alas, this is the final edition of Roto-Relevant Research. Getting Blanked would like to thank Eno for all his great work for us over the years. Make sure you follow him on twitter and read his work daily at Fangraphs.