- FanGraphs Baseball - http://www.fangraphs.com/blogs -

Theo, Ellsbury, and UZR

Last week, Red Sox GM Theo Epstein did an in-depth interview with WEEI, focusing mainly on the topic of the winter – defense, defensive statistics, and how most of the stat-friendly teams have made moves to improve themselves defensively in recent years. During the interview, he made one specific comment that we feel is worth addressing, since it was aimed our way. In response to a question about whether Jacoby Ellsbury had defensive problems in center field last year, Theo said this:

I think that he is an above-average center fielder now, who is going to be a great center fielder. I know there is a certain number we don’t use that is accessible to people online that had him as one of the worst defensive center fielders in baseball last year. I don’t think it’s worth anything. I don’t think that number is legitimate. We do our own stuff and it showed that he is above average.

Since we publish that “certain number” – more often referred to as UZR – let me weigh in with a few thoughts.

UZR, with its -18.6 rating for Ellsbury in center field in 2009, isn’t exactly out on an island here. John Dewan’s +/- metric had him at 9 runs below average. Sean Smith‘s Total Zone system had him at 10 runs below average. Tom Tango stated that his With Or Without You system had him between 14 and 18 runs below average. This isn’t a case of UZR delivering a strange result that other systems don’t agree with. Pretty much all of the publicly available defensive metrics show Ellsbury had some issues last year.

Now, Theo might lump all of those metrics together as inferior to their proprietary internal metrics, and indeed, they may be. However, we need to keep two things in mind here: first, Epstein making positive public comments about his own players is a classic case of a statement made in self-interest, and second, the Red Sox moved Ellsbury to left field. Theo’s not going to come out and trash any of his own players, and it’s in the Red Sox best interest to fight any perception of Ellsbury as a defensive liability. If they engage in trade negotiations with another team, it would not be helpful if the league comes to a consensus that Ellsbury really does have some defensive issues, considering that is the biggest selling point for his particular skillset.

The Red Sox decision to move Ellsbury to left reinforces that idea. If Theo had left it unchallenged, it would essentially amount to a tacit acceptance of the rating, which would not be good for Jacoby’s trade value. Even if the Red Sox internal metrics had not shown Ellsbury as above average, it would still behoove Theo to publicly defend his player against the perception that his defense in center field may be questionable. Once the Red Sox decided to shift Ellsbury over, it became necessary for Epstein to make a statement to this effect, whether he believes it or not (and I’d bet that he does – this is not intended to question his integrity).

Finally, this is a big point – Jacoby Ellsbury played 1,302 innings in center field last year, basically one full season’s worth. As has been noted many times, one season’s worth of any defensive metric is not a very large sample size. Due to the amount of marginal plays that a player is judged on over the course of a single season, a few bad breaks here or there can make a pretty significant impact on a player’s overall rating. We have always suggested that you want more than one year of data before you start making judgments about a player’s true worth defensively. No one should look at Ellsbury’s 2009 UZR and state definitively that he is a poor defensive center fielder.

In fact, UZR doesn’t even support that assessment. In 2007 and 2008, Ellsbury racked up a +14.4 UZR in 777 innings between left and right field. That equates out to about 20 runs above average, if you extrapolate out over a full season. UZR loved Ellsbury in the corners, and historically, the defensive gap between a CF and a corner OF is about 10 runs. Given how well UZR rated him as a corner outfielder (again, in a very small sample), we can use that data as information about how well he should be able to handle center field. An overall view of Ellsbury through UZR, including all of the data from 2007 to 2009, would have him as a barely below average CF, not anything close to a disaster, and not that far from what Epstein is claiming.

There is a school of thought that these swings suggest an underlying flaw with UZR, but I’d suggest that it may be evidence that the perception of perfectly consistent defensive value is a myth. We know that hitters and pitchers often see wild swings in their performance, but no one thinks its proof that home runs are bunk when David Wright gets out-homered by Ichiro Suzuki. Wright obviously has more power, but over one season, he didn’t show it. It is certainly reasonable to believe that a player that Epstein believes to be “a good defensive center fielder” could simply have a bad year.

In the end, there’s no huge disagreement between FanGraphs and the Red Sox on how to evaluate defense, even if they prefer their internal metrics to UZR. We love the defense that Mike Cameron and Adrian Beltre provide and, obviously, so do the decision makers in Boston. Additionally, that Ellsbury was shifted to left field to make room for a 37-year-old suggests that the Red Sox may agree that he’s not yet an elite defender, even if they think he may become one. In this instance, I think that actions may speak louder than words, and I don’t think that Theo sees Ellsbury all that differently than we do.