The Yankees' bullpen candidates, part 2: David Robertson

In this post, I would like to profile David Robertson. Personally, I find him to be the most intriguing of the Yankees' bullpen pitchers due to his unconventional method of obtaining a superb strikeout rate.

Drafted (17th round) as a 21 year old in 2006, Robertson shot up through the Yankees' minor league system, dominating at every level:

League	IP	ERA	FIP	K/9	BB/9
A	47	0.77	1.31	12.83	2.87
A+	33 1/3	1.08	2.33	9.99	4.05
AA	22 2/3	1.19	1.17	13.90	3.18
AAA	49 2/3	1.99	1.79	13.77	4.17
MiL	152 2/3	1.30	1.67	12.67	3.60

Robertson made his major league debut against the Mets in June of 2008, and he stayed with the big club for two months before being optioned back to Triple A (he would make another cameo as a September call-up). He had an ugly 5.34 ERA over his 30 1/3 innings of work (25 appearances), but his 36 strikeouts (10.38 K/9) impressed; he also posted a much better 3.53/3.72 FIP/xFIP. Though Robertson was left off of the Opening Day roster in 2009, by the end of May, he was in the Yankee bullpen for good. For the 2009 regular season, Robertson put up a solid 3.30 ERA in 43 2/3 innings (45 appearances). His FIP/xFIP were both excellent at 3.05/3.20. What was most remarkable about Robertson in 2009 was his strikeout rate. His 12.98 K/9 for the year was second in all of major league baseball to Jonathan Broxton. Interestingly, nearly 15% of all at-bats against Robertson resulted in a called strike three, which was far and away the highest mark in the majors (the league average for called strikeout% is about 4.5%). This is helped by his 24.4% called strike rate (called strikes/total pitches), which was also best in the major leagues (~16% league average for relievers). Robertson relied on deception in the minor leagues, as well; in his Triple A stint from 2008-2009 (844 pitches overall), he had a similar called strike rate of 23.7%. But what is so special about Robertson that causes opposing batters to watch so many pitches? Can some PITCHf/x data help us out here?

Pitch	Average Speed	Max. Speed	pfx_x	pfx_z
FF	91.75	95.5	-0.01	11.14
CU	79.34	83.3	3.98	-8.69
CH	85.52	87.9	-8.08	9.85
SL	83.46	85.8	6.87	2.95

(click graphs for larger image)

Pitch	Pitch#	Pitch%	Swing%	Whiff%	Wide Zone%	Chase%	Watch%
FF	654	75.9%	35.8%	26%	59%	21.4%	52.3%
CU	172	20%	50.6%	27.6%	42.4%	40.2%	28.8%
CH	25	2.9%	24%	33.3%	32%	16.7%	37.5%
SL	11	1.3%	36.4%	0%	27.3%	75%	66.7%
	862	99.5%	38.4%	26.3%	54.5%	26.9%	48.5%

Pitch	SLGCON	RSv/100
FF	.543	0.88
CU	.773	-0.26
	.596	0.48

A lot of both called and swinging strikes from Robertson's fastball, it looks like. The pitch is pretty much straight, sometimes cutting slightly. He also utilizes a hard curveball with a significant amount of vertical drop. I also found a few tailing changeups in the data, and some other pitches that looked like sliders. But for the most part, Robertson is all about the fastball and curveball. And particularly, the fastball. The strikeout breakdown below shows the third strike for each of Robertson's 63 strikeouts in 2009.

Pitch	Total	Called	Swinging	Total%	Called%	Swinging%
FF	38	20	18	60.3%	31.7%	28.6%
CU	24	7	17	38.1%	11.1%	27%
SL	1	1	0	1.6%	1.6%	0%
	63	28	35	100%	44.4%	55.6%

The pitch that earned Robertson the most strikeouts in 2009 was a "watched" fastball. It's pretty ridiculous that he could consistently get batters to watch the straight fastball - moreover, one that averaged under 92 mph for the season. There is clearly something pretty special going on here. From the outside, Robertson's motion itself doesn't seem too deceptive, but there are all sorts of ways that a pitcher can hide the ball and make it appear that it's being thrown harder. There has been some work at Baseball Prospectus regarding perceived velocity, and I'm sure that there will be plenty more research done on the phenomenon. I can't say that I would have much to offer on the topic, so I will leave that as something of a mystery for now.

Another thing worth noting about Robertson's 2009 season was his dramatic velocity increase. In April and May, he was often working in the upper 80s and sat around 90 mph, but in August, he was averaging over 93 mph. The increase is illustrated clearly by Robertson's velocity graph over the course of the season (all pitches):

If you look closely, you can see a dip right near the end of the season. That marks Robertson's return from a three-week long hiatus due to a shoulder injury. On River Ave. Blues, there's an interesting post that suggested that increased velocity itself might have had something to do with Robertson's injury. Mid season, Robertson tweaked his delivery at Dave Eiland's heed (in order to add velocity), and it's not unreasonable to believe that this adjustment wound up hurting Robertson. Since Robertson's velocity remained strong into the beginning of September, I'm having a hard time believing that Robertson's injury was that serious. I don't claim to know a whole lot about injuries, but would I believe that the drop following his return had more to do with a lack of rehab time (the Yankees needed him for the playoff bullpen) than actual harm to the shoulder. (EDIT: According to the YES Network gun, Robertson sat around 93 mph and hit 95 mph in his first Spring Training appearance on Friday. The radar gun has seemed pretty accurate so far, so that is definitely an encouraging sign.)

So, what can we look for in Robertson's 2010 campaign? Sustaining a ~13 K/9 is obviously unrealistic, but it's clear that there's something special about his stuff. His xFIP has been higher than his FIP during both years, and since he's a flyball pitcher (41.3% FB rate for both years), home runs are something to watch out for. Manager Joe Girardi seemed to gain confidence in Robertson as the season wore on - Robertson picked up his first "hold" of the year against Seattle on August 15th, and picked up four more right before his shoulder injury. I fully expect Phil Hughes or Joba Chamberlain (personally, I think that it will be Hughes) to be the full time "set up guy" this season, but Robertson has shown that he should be in the mix for some important innings.

Gameday PITCHf/x data is from MLB Advanced Media; it can be easily accessed via this tool. Other statistics are from Fangraphs, Baseball-Reference, and Stat Corner.