Recently, there has been some research on how different ERA estimators (ERA, FIP, xFIP, tERA, and SIERA) perform under different circumstances: in particular, FanGraph’s Bill Petti wrote this fascinating three-part series on how park factors and pitcher’s batted ball profiles affect the performance of ERA estimators. I’ll summarize some of the conclusions, if you don’t feel like reading the entire thing—but you should check it out if you feel like my analysis is incomplete in any way.
Basically, different ERA estimators work better for different types of pitchers and in different types of parks. The performance of a ground ball-heavy pitcher in a pitcher-friendly park can be predicted with slightly more accuracy with one ERA estimator, and the performance of a fly ball pitcher in a hitter friendly park can be slightly more accurately predicted with a different ERA estimator—and so on.
The basic ERA estimators used in the study were ERA (which, unsurprisingly, doesn’t do a great job of predicting future ERA), Fielding Independent Pitching (FIP, which is based on Voros McCracken’s Defense Independent Pitching theory, measures what a pitcher’s ERA should have been if performance on balls in play was league average), xFIP (a park-controlled version of FIP, which substitutes a league average HR/FB rate into the FIP formula—I talked about this in my AJ Burnett article), True ERA (tERA, which takes balls in play into account as measures of a pitcher’s skill), and Skill-Interactive ERA (SIERA, FanGraph’s own ERA estimator, which they tout as the “most accurate,” though some disagree).
The results of the study were a little bit outside of conventional wisdom: in general, the batted ball statistics (tERA and SIERA) tended to have more success in predicting ERA than the purely defense-independent ones (FIP, xFIP), despite the fact that pitchers don’t have much control over batted balls. I looked at hitter-friendly parks in particular because, well, this is a Yankee blog, and found that, in general, tERA has the most predictive power for pitchers in hitter-friendly parks, no matter their style (ground ball or fly ball).
There is one last thing I’d like to mention before I wade through some of the numbers and break down the rotation (as best I can): tERA was never intended to predict ERA—in fact, its’ original intent was to predict runs generally. The theory was designed specifically to control against the semi-arbitrary designation of earned and unearned runs. I mention this because it tends to be the main criticism of tERA (tRA), and because FanGraphs (in particular) emphasizes the E part of tERA. So these numbers might be a bit off in that regard, but bear with me nonetheless.
2011: 3.00 ERA, 2.88 FIP, 3.02 xFIP, 3.14 SIERA, 3.75 tERA, 3.36 tRA
2012 Projections, ERA: 3.33 (Bill James), 3.55 ZiPS
CC is the easiest pitcher on the Yankee’s staff to project. He’s good. Really good. And there’s no reason to think that is going to change—at least not according to any of the numbers. He’s a ground ball pitcher (1.54 GB/FB last year, 1.30 career), who might have actually gotten slightly unlucky last year, with a Batting Average on Balls in Play (BABIP) that was almost thirty points higher (.318) than his career average (.291). Looking at the metrics, they all love him, though tERA less than the others—incidentally, his tRA, the metric the theory was based on, likes him more. He’ll be great. ‘Nuff said.
2011: 3.70 ERA, 4.01 FIP, 4.16 xFIP, 4.29 SIERA, 4.46 tERA, 4.50 tRA
2012 Projections, ERA: 4.28 (Bill James), 4.44 (ZiPS)
Nova probably overperformed in 2011. I say “probably” because he’s a rookie: there is hardly any way for us to know where his average is—and thus, there’s almost no way for us to know what a “regression” will mean, or how to spot one from a simple drop in form. He’s a heavy ground ball pitcher (1.83 GB/FB last year) who pitches to contact, so his numbers are easily affected by luck—both good and bad. We can’t say, for example, that his 2011 BABIP of .283 is low, because it’s more than possible that it is simply a reflection of his style. The fielding independent estimators liked Nova more than the ones that include batted ball profile, but the difference isn’t huge: they’re all indicative of a probable 2012 ERA in the low 4’s.
2011: 3.74 ERA 3.42 FIP, 3.53 xFIP, 3.36 SIERA, 3.42 tERA, 3.48 tRA
2012 Projections, ERA: 3.07 (Bill James), 4.11 (ZiPS)
Now we get to the interesting part: Pineda, coming from one of the best pitcher-friendly parks in the country, is a dead fly ball pitcher. He benefitted last season from a pretty low 9.0% HR/FB rate, which could be partially due to the spacey confines in Seattle. But the batted ball metrics like him a lot! His SIERA and tERA/RA were lower than his FIP/xFIP, and they’re both park-controlled. Is it possible that Michael Pineda might actually be better than we thought? Yes and no. It is possible that his profile might fit at Yankee Stadium, despite the fly ball tendencies, but I wouldn’t be so quick to rejoice—his stellar 2011 performance is complicated by a super-low .258 BABIP, lower than average. Though again, like Nova, we have so little date on Pineda that we can’t say for sure what mean he has to regress to, and it’s possible that many of his peripheral stats are simply indications that he’s just really good—at least that’s what most of the ERA estimators suggest. And because I’m a Pineda believer, I’m inclined to think they’re right.
2011: 3.07 ERA, 3.78 FIP, 3.56 xFIP, 3.66 SIERA, 4.22 tERA, 3.98 tRA
2012 Projections, ERA: 3.57 (Bill James), 4.33 (ZiPS)
Kuroda is a very talented pitcher who has put up great numbers in the National League for his whole career. But that’s the problem. How can we project numbers for a pitcher moving from league to league, and especially from the terrible NL West to the hyper-competitive AL East? Well, we can’t really. Kuroda is a ground ball pitcher (1.52 career GB/FB), but his rate dropped off last year (1.24 GB/FB); he should fare well in the AL East if he can maintain some of his peripherals (he had a slightly above-average HR/FB rate last year, and a very reasonable K and BB rate). At least that’s what the ERA estimators tell us: he’s a high-3’s, low-4’s skill level, who overpeformed ERA-wise last season. At the same time, I don’t see a huge reason top panic about Kuroda, as the batted ball ERA stats don’t hate him, and they tend to predict ERA in hitter-friendly parks better than the totally fielding independent metrics. I’m cautiously optimistic about him.
2011: 3.62 ERA, 4.12 FIP, 4.36 xFIP, 4.43 SIERA, 4.86 tERA, 4.83 tRA
2012 Projections, ERA: 4.25 (Bill James), 4.85 (ZiPS)
Ah, Freddy. By all accounts, Garcia hugely overperformed in 2011: his 3.62 ERA is half a point lower than any of the other metrics, and more than a point lower than his tRA. He’s a pretty neutral GB/FB pitcher over his career, with a career 1.11 average, but he skewed towards fly balls last season, with a .88 GB/FB. Weirdly, that’s not consistent with his below-average 8.2% HR/FB rate (which is below his career 11.1%), suggesting that he got pretty lucky. Nothing in this profile suggests that the Yankees should expect another similar performance from Freddy, and his 2011 tERA/tRA suggests that he might really drop off this year. Writing this article has only reinforced my belief that the Yankees need to do their absolute utmost to make sure Phil Hughes has a good season.
I know this is going to sound like a cop out, but I don’t feel comfortable writing a projection for Hughes, or even discussing possible numbers. His 2011 could have been a complete anomaly, especially when we consider his mysterious disappearing fastball and half-year injury. There’s very, very little data on which to make any sort of prediction (I know, that didn’t stop me from talking about Nova and Pineda), and because of the interruptions in his ball playing life, I don’t think there’s that much to be gleaned out of his stats to date. He’s working on new pitches, trying to find velocity, and master his control—if he gets to a place similar to 2010 (with his velocity, command, etc) then we can talk. Until then, I don’t know if it’s even worth discussing Hughes’ stats: we should treat 2012 almost like a new beginning for him, I think.
But then again, I might just be stupid and totally wrong. And that goes for all of this.