The Changeup and BABIP: does throwing changeups help you lower your BABIP?

While a large part of our readers here on Bluebird Banter believes that pitchers have no influence over BABIP at all, there are a few pitchers who do their best to make us believe otherwise. Take former Jays pitcher Ted Lilly: he's got a .271 BABIP in 1904 career innings. That's not a small sample size by any means, and the BABIP is significantly lower than league average (.290-.300 range). He has managed to keep his career ERA at 4.17 even though his career FIP sits at 4.42. Another former Jay: Shaun Marcum. Career BABIP: .269. Career FIP: 4.28, but career ERA: 3.77. That's over 792 2/3 innings, so the sample isn't as large. Matt Cain (1317 1/3 IP) has a career .265 BABIP, a career 3.69 FIP and a career ERA of 3.35. So, obviously these guys are doing something right, right? While looking for explanations (the use of the cutter, quality of the fastball, horizontal movement on pitches) I believe to have found a relatively strong link between the use of the changeup and the ability to defy the league average BABIP.

What I'll do here is list the starting pitchers with over 700 innings pitched since 2006 who have a career BABIP of .287 or lower. (I wanted to include John Lannan who has 4.62 FIP but a 4.01 ERA over that period, and it seemed like a good cutoff point)

What you'll see in the table below is a pitcher's career BABIP, his most valuable pitches in order of value (Fangraphs' pitch values), his career changeup% (of pitches thrown), and the whiff% for the changeup since 2009 (league average for changeups is around 12%).

*=splitter, **=did not qualify, a "bonus" if you will

(analysis after the jump)

Chris Young is an interesting case, worthy of an article of his own, because he throws mostly a unique rising fastball that gets him a lot of popups and (probably) a lot of weak flyball outs. It is unfortunate that he has been injured so much the past couple of years, which is why he didn't get to the innings cutoff. Others who didn't make it to the cutoff but have limited BABIP are John Maine (FB/SL), Armando Galarraga (SL), Roger Clemens (FB/CH*), Ian Kennedy (FB/CH), Trevor Cahill (FB/CH), David Price (FB/CH), Colby Lewis (SL), Rich Harden (CH*/FB), Mat Latos (FB/SL), J.A. Happ (FB), Tommy Hanson (SL/FB/CB), Bruce Chen (SL/CB) and Jeff Niemann (FB). Of these, Niemann, Hanson, Latos, Lewis, Price and Galarraga throw their changeups less than 10% of the time, a percentage that is remarkably higher than the group where the sample size is bigger. Niemann (cutting action) and Price (running action) have pretty unique fastballs in terms of horizontal movement, and in Price's case, velocity. Hanson's fastball is a "rising" one. Latos's fastball is, like Niemann's more of a cutter in terms of movement, but it's thrown harder than most cutters. Lewis' fastball both cuts a bit and "rises" a bit. Galarraga is more of a mystery, his slider has a lot less cutting and a lot less downward movement than those of the other slider specialists, and he doesn't throw a four-seamer at all, just a "sinker". Perhaps he limited his BABIP because he gave up so many homers?

Back to the pitchers who did make the cutoff. Not counting Wakefield (who is obviously a very different kind of pitcher), only Kershaw threw the changeup less than 10% of time. At just 3.6%, one can presume the hitters won't be sitting changeup against Kershaw. Of the 20 non-knuckleballers, 7 had the changeup as their best pitch, 12 the fastball and just one (Arroyo) the slider. None were curveball specialists. Only 4 got less than 10% whiffs on their changeup. For Wolf, Arroyo and Washburn, it seems quite strange to throw the changeup so often, though perhaps it made their other offerings that much better? Zambrano, Marcum and Hamels throw cutters, while Weaver throws a fastball that moves like a (rising) cutter. The only groundballers of the bunch are Webb and Hudson, and both have pretty extreme sinkers.

I think it's time to take a look at 20 guys who underperformed the league average BABIP the most, using the same criteria as for the other group (>700 innings since 2006):

*=splitter

First thing to note here is that it's harder to find pitchers who have worse than league average BABIP than it is to find those who have lower BABIPs. This isn't too strange, as you'd expect most pitchers with bad BABIPs to not last long in the major leagues as a starting pitcher. Pitchers like Kevin Millwood, Joe Blanton, Joel Pineiro, Jeff Suppan and Jake Westbrook have had inflated BABIPs since 2006, but not too much over their careers. Most of them have seen their fastball effectiveness sharply decline.

Now another thing you will have noted is the large number of breaking ball specialists on this list. 7 pitchers are slider specialists, as Doug Davis and Andy Pettitte throw a "slutter" which is seen as cutter by PitchFX, but is, in my opinion, too different from the regular fastballs used by these gentlemen to be called a cutter. Whether you count 7 or 5 slider specialists, the group is a lot bigger than the lone slider specialist (Arroyo) in the group that limited hits on balls in play. In addition, there are 3 curveball specialists, which means it's 10 (BABIP underperformers) to 1 (BABIP overperformers) in breaking ball junkies. Only one (Duke) is a changeup specialist, compared to 7 in the other group. Six or seven (depending on if you count Blackburn) fastball specialists compared to 12 in the group of low BABIPers. Compared to just one in the other group, half of those who have suffered from bad BABIPs threw less than 10% changeups. Thirteen got less than 10% whiffs with their changeups, compared to four in the other group.

Two interesting cases here are Zach Duke and Edwin Jackson. Zach Duke has a bad fastball, and I mean very bad. Epically bad.  He has a -101.2 run value on the fastball over his career, which is quite an achievement, really. Edwin Jackson is no slouch either, he has -75.6 run value on the fastball over his career, and (despite the whiffs) -10.4 on the changeup. Throwing in the mid-90s isn't everything, huh, Edwin?

Conclusion

I do not want to make the conclusion that a good changeup will make a pitcher have lower than usual BABIP. Zach Duke is plenty of proof, and Edwin Jackson and James Shields are, too, that a bad fastball can mean a normal or even bad BABIP despite a good changeup. Along with a good, often used changeup, throwing a good four-seamer or cutter might also work towards a low BABIP. Sinkerballers and breaking ball pitchers who lack a good changeup might find themselves struggling to keep their BABIPs low.

What does this mean for Jays' pitchers? Ricky Romero now has a .285 career BABIP, which he might even improve upon, given his strong changeup/fastball (sometimes with some cutting action) combination. But I wouldn't expect him to put up anywhere near a .242 mark again, so I'd expect him to be a 3.40-3.70 ERA pitcher. Brett Cecil is also a changeup guy, but his fastball is a liability, I don't expect him to last long if he doesn't improve on his fastball. For Brandon Morrow, his BABIP hasn't been that problematic this year at .300, although last year it was .342, it's mostly the LOB% that needs to come down. As a pitcher who relies heavily on the slider, BABIP could still be a problem if his new cutter doesn't work out. Henderson Alvarez is probably safe from BABIP troubles, and time will tell if he's got the ability to pitch to a low BABIP with that power sinker of his, in the footsteps of Webb and Hudson, but although he throws harder, the pitch does not have as much sink. McGowan actually has a very good changeup, but he could use it more, and of course, he needs to command it. The command issue is of course also the worry with Drabek's change, which he has thrown for swings and misses but rarely for strikes (like his other pitches).