Keeping it Simple: A.J. Burnett

In 2010, there was a stark difference in the distribution of pitch locations of A.J.’s curve compared to earlier years. You can see this in the following graph:

In this graph red represents locations that A.J. threw to less in 2010 than in 2008-2009, and blue represents the locations where A.J. threw his curve more to in 2010 than in 2008-2009. The graph is from the catcher’s perspective, so the right side of the graph is close to left-handed batters and the left side of the graph is close to right-handed batters. The dotted box is the strikezone. You can click on the graph (and all following graphs) to enlarge.

For whatever reason, in 2010 Burnett threw his curve significantly lower than before. As a consequence, he also threw the pitch much less often in the strikezone than 2008-2009. It’s not clear whether or not this was an intentional change, but this kind of dramatic shift in the locations of his curves is not likely to occur by random chance. Either way, a proclivity to throw his curve for balls is bad for many compounding reasons. If he is not throwing his curve for strikes, he’s falling behind in the count where batters are more comfortable. He’s especially vulnerable behind in the count because batters can just sit on his fastball, which has been ineffective in recent years. Additionally, throwing his curve for balls makes future curves less effective both in the remainder of the at-bat and the game.

However, there are reasons to believe his curve was unlucky in 2010. This is supported by the pitch’s effectiveness early on in 2011. Consider the following graph:

This graph shows the run value of A.J.’s curveball by vertical location. Red represents 2008-2009 and turquoise represents 2010. The two vertical lines in the background indicate the vertical borders of the strikezone, with the left line indicated the bottom and and the right line indicating the top of the zone. In this graph negative values, or values closer to the bottom of the graph, are good for A.J. Positive values, or values closer to the top of the graph, are bad for A.J. Gray bands indicate confidence. You can click on the graph to enlarge.

Interestingly, it seemed that in 2008-2009 A.J. could still throw his curve in the middle of the zone and it still be effective, while in 2010 when A.J. got his curve up a little it got battered. This is more an indicative of the anger of the BABIP gods than anything else. If I calculate his run value tERA style (by substituting league average run values for groundballs, linedrives, flyballs, and popups instead of using the actual results), his curve comes out as a solidly above average pitch. And while the movement on the pitch was not quite as good in 2010 as it was in 2008-2009, the pitch was and is still nasty.  It seems that we can conclude that the “loss” of A.J.’s curve in 2010 was mainly due to a lack of aggressiveness within the zone and bad luck. There should be no worries about the pitch for the 2011 season.

Early on this year, the results have been promising. The strikeouts are way up and perhaps more importantly, his swinging strike percentage is much improved as well. In 2011 for every 100 pitches he has thrown, he has garnered 10 swinging strikes, which is about his career average and is much better than the figures he posted in the previous years of his Yankee tenure. This tells us that his stuff has been great. He’s also mixing up his pitches more than ever; in 2011 he is throwing less fastballs than before and he’s replacing them with more changeups. The ability to throw his curve for a strike seems to have returned as well. According to Trip Somer’s tool, the pitch has been thrown for a strike 58.7% of the time in 2011, which is both above average and much better than 2010. It seems then that A.J. may not be so complicated after all. Forget all the theories about mental instability, all A.J. needs to do is mix up his pitches and throw his curve for strikes. That doesn’t sound too different from other pitchers now does it?

*pitch f/x data are from MLBAM using Darrel Zimmerman’s pbp2 database.
*run value code is courtesy of Ricky Zanker
*h/t to  for picture of Burnett

7 thoughts on “Keeping it Simple: A.J. Burnett

  1. Josh, another fascinating piece, worthy of some thought. Could there be a relationship between Burnett throwing fewer curve balls for strikes and that these curve ball strikes were hit harder in 2010? The cause-effect thing is confusing here, but the knee-jerk obvious conclusion is that Burnett threw more curves out of the zone because his curves in the zone were being hit hard.

    • That was my first guess too Larry, but I looked at the average heights of his curveball as the season progressed, and the average heights actually increased, not decreased. However, It is possible that this reaction occurred in spring training.

  2. This may be a) happenstance and/or b) heretical, but maybe, just maybe, there was something to the smoke involved in the "Burnett has issues with Posada catching" bits that we've been hearing about for a while? Just going through the first, say, 8 games of 2009 at ESPN Game Logs for Burnett, I think it was like 5 Molina, 2 Posada and 1 Cash (!!!)… but 2010 was like all Posada. I've never heard anything negative (or positive' just neutral) about Jorge's game calling skills, but we HAVE heard that — especially with age — he may not get a stop on a breaking ball in the dirt quite as well as others, which may have caused AJ to have issues tossing the pitch (if he couldn't establish the out-of-zone pitches, batters might sit on the in-zone ones?).

    Is there any way of seeing how many games in 2009 Posada caught vs. 2010 for the whole year? (though I note WP totals were roughly equal from 2009 – 2010)

    • Thanks for adding to the discussion David. About your theory: If Posada was the issue, wouldn't it make more sense that A.J. had to throw his curve higher than before to compensate for the increased possibility of a wild pitch? And it is possible to see how many games Posada caught in 2010, though I'm not sure how helpful that information would be.

      • Ah yes, in traditional late-night style — I read the colors backwards (NOTE TO SELF: less wine before bed). It WOULD make more sense what you are saying! I am genuinely interested though if there was something to the Posada bit, though rationally there's not much data to parse here to get from a + b = c. We have, at best, a + something = fish… >.>

  3. Josh –

    Any thoughts as to whether Burnett's release point on his curve ball changing? That might be a reason behind Burnett's curve missing low.

  4. Here are the averages of his release points in the format (horizontal release point, vertical release point): 2008: (-2.05, 5.82), 2009: (-2.09, 5.9), 2010: (-2.02, 5.87).

    It's hard to say that there's anything important going on here. Also, talking about release points is often kind of iffy, because stadium to stadium calibration errors play a significant role here and the pitcher just may be shifting around on the mound.