The mystery of Danny Duffy and his high BABIP

Thru Monday's games, the average American League pitcher allowed hits on balls in play at a clip of .289. The Royals as a team have allowed hits at a .258 rate on balls in play. Leaving that gap aside for a moment (last year, the league pitchers allowed a BABIP of .299 and the Royals allowed them at a .294 rate).

There's 3 Royals pitchers to have BABIPs above .282:

Jason Frasor: .290
Jeremy Guthrie: .312 (he has never topped a .298 BABIP as a Royal), andddddd...
Danny Duffy: .371

So, why in the Wide Wide World of Sports does Danny Duffy have such a high BABIP so far? let's explore the possibilities and maybe approach some answers to this question.

The conventional wisdom about pitchers on this team is that you just need to put the ball in play and the defense will save you at a very consistent rate.

The Royals have had the same defenders (Perez, Hosmer, Escobar, Moustakas, Gordon and Cain) playing regularly at six of the eight positions since 2013 (at least five of those six started in 131 games last year). The relative consistency and quality of the defense can be taken into account in checking the 2013/2014 BABIP numbers for the regular pitchers.

Back in 2013, the regulars that checked in above the league average for BABIP were Wade Davis (.362), Luis Mendoza (.313) and James Shields (.301). Relievers with BABIPs over .300 included Aaron Crow, Tim Collins and the JC Gutierrez experience.

In 2014, the regular pitchers checking in above the league average for BABIP was just Jason Vargas (.300). Several relievers also had a BABIP above .299 (Liam Hendriks at .388, Bruce Chen, Casey Coleman, seven innings of Brandon Finnegan, Michael Mariot and Louis Coleman). Royals starting pitchers don't typically concede such a high BABIP with this defense, outside of Wade Davis, more on that later.

So, what's the problem with Danny Duffy and his BABIP? to the numbers...

BrooksBaseball says that Duffy has thrown 269 four-seam fastballs, 140 sliders, 108 sinkers, 72 change-ups, and 14 curveballs. You may notice from the MLB.com pitch-by-pitch that they credit Duffy with more curveballs. Brooks has those pitches listed as "sliders". Take it up with them.

When at-bats end on his fastball, opposing hitters are hitting .407 (24 for 59) with 12 walks. Slider? 7 for 37 (.189) with 2 doubles and 2 walks. Sinker? 9 for 24 (.375) with 2 HRs and 0 walks. Change 2 for 13 (.154) with 1 walk. No plate appearance against Duffy has ended on a curveball so far.

If you look at just the hitting results, Duffy's slider is the most effective pitch to end an at-bat so far. But there's a catch... 42% of the time he throws a slider, it's called a ball. Changeup? 44%. Curve? 50% (of 14).

On one side, Duffy is throwing fastballs for strikes, but nobody swings and misses at his fastball. 5% of the fastballs were whiffs. Chris Young, who throws an 87 mph fastball, gets nearly 8% whiffs on his fastball. (My apologies to Chris Young, but if Bruce Chen were still on the team, he'd be the comparison point for fastball whiffs today.)

What about Duffy's pitch usage? On the first pitch to lefties, Duffy typically throws a fastball (68% fastball, 21% slider, 12% sinker). On his first pitch to righties, Duffy typically throws a sinker (39% sinker, 31% fastball, 15% slider, 12% change, 3% curve).

On the first pitch against Duffy, hitters are hitting .533 (8 for 15). That's 15 times out of 153 plate appearances hitters are putting the ball in play on the first pitch (small sample size alert). Duffy threw a first pitch strike of some form to 96 of 153 hitters, so that's 15 times of 96 strikes a hitter put it in play. Make of that what you wish. Of those 153 plate appearances, 54 ended with Duffy ahead in the count, 50 with him behind in the count and 49 even counts.

If a batter gets ahead, Duffy starts throwing more fastballs. 79% to left-handed bats and 60% to right-handed bats. With 2 strikes and/or Duffy ahead, he splits between fastballs (around 37-46%) and sliders (30% to 41%) without regard for handedness. Hitters hit better v. Duffy on even counts (0-0/1-1/2-2) than when they were ahead of Duffy in the count. Despite Duffy leaning harder on his fastball behind in counts than in even counts.

To put it in chart form.

BABIP by final count
		Strikes
		0	1	2
Balls	0	0.500	0.273	0.188
	1	0.300	0.330	0.375
	2	0.000	0.143	0.636
	3	0.000	0.250	0.571

OBP by final count
		Strikes
		0	1	2
Balls	0	0.533	0.273	0.188
	1	0.364	0.385	0.222
	2	0.000	0.143	0.400
	3	1.000	0.700	0.667

When do Duffy PA's end
		Strikes
		0	1	2
Balls	0	16	11	16
	1	11	13	27
	2	2	7	20
	3	2	10	18

What kind of contact is being made? According to Fangraphs, this year it's 52% medium-hit balls, 34% hard and 14% soft. That's good for 40% groundballs, 31% flyballs and 29% (!!!) line drives. Last year the percentages were 52% medium/30% hard/18% soft and 46% FB/36% GB/18% LD. In Duffy's only reasonably full season before 2014 (back in 2011), it was 51% medium/26% hard/23% soft and 40% FB/37.5% GB/22% LD. So there's a sliiight difference between 2014 and 2015. That's a strong candidate for a problem in Duffy's results so far.

Looking back at last year and this year..

Last year: 58% Fastballs, 14% Curveballs, 11% Sinkers, 9% Changeups, 8% Sliders.

This year: 45% Fastballs, 23% Sliders, 18% Sinkers, 12% Changeups, 2% Curves. More sliders (of varying control), sinkers (of varying BABIP), more changeups (slightly), less fastballs, way less curveballs (or maybe Brooks is messed up).

Last year, Duffy allowed a BABIP of .227 on his fastball, .250 on his sinker, .278 on his change, .333 on his slider.

Compare that to this year, BABIP of .480 on his fastball, .350 on his sinker, .143 on his change, .259 on his slider.

That leads me to Dave Eiland and his words of concern:

"There’s no major problems here. There’s no health issues. There’s no game plan issues. Well, there is a game plan issue. Because his first eight to 10 pitches, he wants to throw all fastballs, and he wants to throw them 110 mph. You can’t do that. You’ve got to go out there and you’ve got to pitch"

Last year, Duffy's fastball averaged 94.2 MPH and he threw it on the first pitch to lefties 79% of the time and 55% of the time to righties. More fastballs, by percentage, on the first pitch last year than this year. Duffy's fastball this year is at 95.2 MPH on average. So that's an increase of velocity as a starter as the breaking stuff is hindered, something that pitching coaches probably shake their heads at a little bit.

I won't go overly literal and look at Duffy's first ten pitches.

But, instead, I checked to see how many pitches it takes Duffy to (a) throw an offspeed pitch and (b) throw an offspeed pitch for some variant of a strike.

• 5/11: Slider third pitch (ball). Slider 11th pitch (called strike).
• 5/6: Slider seventh pitch (ball). Slider 19th pitch (groundball to 2nd base).
• 4/30: Changeup second pitch (ball). Slider 6th pitch (called strike).
• 4/24: Slider eleventh pitch (ball). Slider 16th pitch (swinging strike).
• 4/19: Slider tenth pitch (swinging strike 3).
• 4/13: Changeup fourth pitch (ball). Slider 8th pitch (foul ball)
• 4/8: Slider third pitch (grounder to 1st base)

So.. Duffy does attempt to establish the fastball, as you can see here. He did that heavily at times in 2013. I'm confident he did that at other times in 2014 as well. There's always gonna be a little bit of that in him, it's just a matter of keeping it within reason.

The first quick possibility for a high BABIP could be a very hittable fastball. I won't assume that a half-inch difference in Pitch f/x horizontal/vertical movement on fastballs changes much for Duffy. One of the pitch f/x scientists can tell us if anything differs there. Same disclaimer is possible for a difference in horizontal release point between early 2015 Duffy and late 2014 Duffy.

The second quick possibility is that his control hinders his aspirations of a good BABIP. Looking over the top 10 list of pitchers with high BABIPs, it's a list of Stephen Strasburg, Clay Buchholz, Alex Wood, Jeremy Hellickson, Mark Buehrle, Gio Gonzalez, Corey Kluber, Duffy, Clayton Kershaw, and Lance Lynn. First glance says "not a lot of high walk rates". Second says "you know, those pitchers with bad control may not accumulate innings to get on the qualified list on Baseball-Reference". Removing that checkmark, I see Jorge de la Rosa, Zach McAllister, Jared Hughes, Mike Fiers, Bryan Morris, and Aroldis Chapman on the list of high BABIPs. So not a huge change on that list.

So.. that collection suggests it's probably a variety of misfortunes that trapped Mark Buehrle and Aroldis Chapman in a land of high BABIPs.

Last year, Colby Lewis and Brandon McCarthy led the band of pitchers with high BABIPs. Chris Young had a .240 BABIP, the same as Duffy's BABIP. So 2014 Duffy defied BABIP and 2015 Duffy is a BABIP victim?

Third quick possibility is that it's just hard contact. Something that might be fixable or might lead to a bullpen stint.

That's where Wade Davis checks back in to this story. From Opening Day 2013 to August 2013, Wade Davis allowed a BABIP of .377 as a starting pitcher. Last year as a reliever, his BABIP was .264.

As a Royals starter, Wade threw 45% fastballs, 21% curves, 19% cutters, 11% sinkers and 3% changeups.

His BABIP by pitch? .320 on fastballs. .383 on curves. .444 on Cutters. .385 on sinkers. .556 on changeups.

Wade allowed 11 hits on 68 changeups. 16% of his changeups became hits, if his fastballs were hits at this rate, he would have allowed 169 hits on fastballs instead of 55. I'm only typing that because it's astounding

As a Royals reliever, Wade threw 57% fastballs, 22% cutters, 20% curves, and 1% sinkers/2 sliders (that are probably a BrooksBaseball error).

His BABIP by pitch? .237 on fastballs, .172 on cutters, .317 on curves.

The differences: no more change-ups (duh). No more sinkers. And huge drops in BABIP on fastballs/cutters. Wade also added 3-4 mph of velocity on all his pitches as a reliever for obvious reasons.

The eye test from what I still remember from Wade Davis' starts in Kansas City is that he got crushed by balls in play but it wasn't exactly a bunch of check swings to right field. Wade also posted seven starts with Game Scores under 30 in 2013. So, that was a bunch of pain. Nobody can say that Wade didn't get enough time as a starter in KC.

I'm confident that there are people with alternate theories on this matter, such as the possibility of injury (diagnosed through the television) or that Duffy is being harmed by his beard (he's not letting a possum latch onto his face).

Like a bunch of guys giving up high BABIPs so far, it could be an assortment of bad fortune at some stop in this season. Heck, Zack Greinke gave up a BABIP of .364 thru seven starts in 2007 before he ventured to the bullpen for three months, and I'm sure he had stretches of high BABIPs after he became good too.

Is there an answer to Duffy's BABIP? No, unless you found an answer between the lines. Sometimes you just have some sort of flaw and a bunch of misfortune to magnify those flaws. Sometimes you just get beat. Sometimes they just hit the fastball.

To be fair to Duffy, he posted four consecutive pretty solid starts before the calendar turned to May. He allowed a ERA of 2.28 from April 13th to 30th and a BABIP of .314. All things might pass and he'll shut down the Yankees despite all of the efforts to boot him from the rotation for two consecutive bad starts. Weirder things have happened.