Joe Sheehan Gets it Wrong on AJ Burnett

Dear New York Yankees,

Please start AJ Burnett as much as you can from now until the end of the year.

Signed, Every single fan of the Red Sox, Rangers, Tigers, Phillies, Braves, Brewers, and either the Diamondbacks or Giants.

It's been discussed in depth already, but I wanted to put in my two cents regarding Joe Sheehan's article on SI.com today regarding A.J. Burnett. In the article, Sheehan argues that, despite Yankee fans frustrations with him, Burnett is one of the five best starters the Yankees have. Sheehan is normally very well-researched and in depth, so the shallowness of his article was striking. Most notably, to me, was the following passage:

There's an argument to be made that Burnett isn't one of the Yankees' five best starters, but you have to work very hard to make it, essentially evaluating Nova, Garcia and Colon based just on 2011 while giving Hughes a pass for his injury-plagued season. Hughes's work since coming off the DL in July has been acceptable, with a 4.28 ERA; he's struck out just 15% of batters. Burnett has a better strikeout rate (20%), the primary evidence that he's still the better starter.

Nah. Calling Burnett a better pitcher than Phil Hughes because of a higher strikeout rate is like calling 2011 Adam Dunn a better hitter than 2011 Vlad Guerrero because Dunn walks more. Unfortunately for Dunn there's more than walks to hitting, and unfortunately for Burnett, there's more to pitching than striking dudes out.

Strikeouts have NEVER been Burnett's problem. In his career, he has finished in the top 10 in his league in K/9 five times, and finished in the top 10 in ERA zero times. High strikeout rates are great predictors of success for young pitchers, even struggling ones. They are poor predictors of success for 34 year old pitchers who have not had success despite high strikeout rates. Also, while Burnett's strikeout rates are up a tick from last year, they would still be his second lowest rate since 2002. Most people would see that Burnett had been an average pitcher with high strikeout rates at his career rates, and ask what happens to that pitcher when those strikeouts decline, but Sheehan is missing the forest for the trees here.

No, the problem with Burnett is all of those other things pitchers are supposed to do. Specifically, not walk people and not give up home runs. Burnett is in the bottom five in the American League in both of those categories, which nobody else can say. He also has hit 7 batters and thrown 17 wild pitches. If you add up these four bad things (and forgive the crude measurement), Burnett gets to 117 on the season, 18 more than Trevor Cahill and Francisco Liriano, who are tied for second place. Per nine innings, Burnett is second (behind Liriano) at 6.75 "bad things" per nine innings. No wonder Yankee fans are frustrated. They watch the pitcher who hurts his team by himself more than anyone else in baseball. He can't blame the passed balls on Jorge Posada anymore, either.

The criticism of basing Colon, Garcia and Nova as better based "only on 2011" is a little bit silly too, since it hasn't exactly been close this year. He then seems to be accusing those who want to use 2011 as a basis for those three but not consider it for Hughes as disingenuous. However, isn't this a pretty sensible way to look at this situation? Hughes really was hurt. Now he's not and is pitching better than Burnett. He was better than Burnett last year. He's been better than Burnett the last two months. Isn't it more disingenuous to rate Burnett as better based on the fact Hughes was banged up in the first half and that Burnett's recent strikeout rate (like I said, the only thing he does well) has been higher? It really seems like Sheehan is cherry picking his arguments here.

To prove a point, the numbers of the five competing starters since the start of 2009:

Year	ERA	IP	H	ER	HR	BB	SO	HBP	WP	BF	ERA+	BF/K	BF/BB+HBP	BF/HR
Burnett	4.72	549.2	548	288	73	245	472	36	50	2407	93	5.10	8.57	32.97
Nova	4.10	166.2	173	76	16	60	100	4	10	723	103	7.23	11.30	45.19
Garcia	4.05	335.1	350	151	36	94	207	6	9	1417	107	6.85	14.17	39.36
Colon	3.75	187	192	78	28	55	145	4	1	797	117	5.50	13.51	28.46
Hughes	4.13	314	288	144	40	104	270	8	14	1310	106	4.85	11.70	32.75

Among this group, the only thing Burnett seems to do obviously better is pitch frequently, and durability seems to be one of Sheehan's arguments for keeping Burnett on rotation. Being durable is a nice trait, but lets think about this - Garcia, Colon and Hughes are healthy *right now.* The Yankees are trying to win a World Series this year. Maybe Burnett gives them a better chance to win games in 2013 than Bartolo Colon, because at some point Colon is going to blow back up, but so what? If Colon is pitching better, and has a history of pitching better when healthy, shouldn't he pitch instead of Burnett WHEN he's healthy? As far as Hughes is concerned, Sheehan seems to be talking out of both sides of his mouth. He's younger and proven to be better. If we're going to give Burnett credit for the higher strikeout rate over the past two months (when Hughes ERA is less than half of Burnett's), doesn't Hughes get credit for having the higher one for the last two and a half years (again, with the lower ERA)? And if Sheehan thinks Burnett should be starting over Nova, he's more foolish than I realize. Nova's low K rates might come back to haunt him, but he clearly does everything else a lot better than Burnett.

Sheehan then uses advanced metrics FIP and xFIP to continue the point in favor of Burnett. He's treading on dangerous ground here, doing something that those of us in the statistical community are often accused of - letting that statistics drive the narrative rather than be used as a part of it. With just a little more looking, Sheehan would realize that Burnett has ALWAYS underperformed his advanced statistics. If a pitcher underperforms his stat line for a year or two, it's fair to believe that it's a fluke, destined to be corrected by the law of averages. If he underperforms it for 13 years, it's fair to say that, in this specific case, the stat gets it wrong. There's no shame in that - advanced stats got Tom Glavine and Ichiro Suzuki wrong for a long time, and most recognized that it was due to specific qualities in their games that some statistics simply can't measure - for Glavine, it was his measurable ability for "situational pitching" and with Ichiro it was a heretofore unmatched combination of bat control and speed. In 13 seasons, Burnett's ERA has equaled or exceeded his xFIP exactly twice. There's a specific reason for this - more of his fly balls go out of the park than other pitchers. Approaching 2000 career innings, this isn't a fluke, or even a trend - it's the pitcher that Burnett is. Expecting that to "correct itself" at this point is silly.

We wake up this morning, and A.J. Burnett is 9-10 with a 4.96 ERA. We don't have any evidence that he's better than that. We do have evidence that the other four guys are better than he is. Burnett has gotten his chances, and needs to be pushed to the side. For the first time in my life, I think most Yankee fans will agree with me.