clock menu more-arrow no yes mobile

Filed under:

Carlos Lee: elite defender?

Fukudome, Denorfia, Carlos Lee.

That sounds like one of those SAT test questions where they ask, "which one of these is unlike the others?"

If you were to look at their body shapes, Lee is different than Denorfia and Fukudome.  If you follow their reputations as outfielders, you are aware that Denorfia and Fukudome have reputations as good defenders and Carlos Lee does not.

AstrosB sent me an email pointing out that has Lee, Denorifa, and Fukudome tied with the top defensive WAR so far in 2011.  I'm sure that some of you would find that to be a surprise.  Baseball-Reference indicates that, prior to 2011, Lee hasn't been a positive defensive player since 2005.  Fangraphs WAR indicates that Lee hasn't been a positive defender since 2008.  In 2010, Fangraphs WAR ranked Lee as the fifth worst defender in baseball with a -17 UZR.  What a turnaround, right?

Although all of the advanced fielding metrics show Lee with a good year, the metric used by Baseball-Reference (Total Zone) is much more favorable to to Lee.  UZR and DRS are the most sophisticated advanced defensive metrics and they both place Lee in good, but not elite, defensive company in 2011.

For example, Fangraphs uses UZR for its fielding WAR component, and Carlos Lee is ranked 18th among ML players, which is good (+7) but not top of the heap.  I like to use DRS, and it shows a similar result for Lee.  In my view, UZR and DRS are more accurate than Total Zone (TZ).  The underlying data for DRS and UZR is derived from Baseball Information Solutions, while TZ relies upon retrosheet data, which can vary in quality. TZ has some advantages if you want to compare fielders for different eras of baseball.  But UZR and DRS are more detailed, accounting more accurately for location of fielded balls and utilizing information on how hard the ball is hit.

Also, we should be careful about ranking all major leaguers' fielding WAR.   The advanced metrics' plus or minus ratings are based upon comparisons to the other players at the same position.  A player can be above average relative to other first basemen, but that doesn't mean he is a better fielder than an average shortstop.  In an absolute sense, shortstops as a group tend to be better fielders than first baseman.  In the overall WAR calculation, this difference is taken into account with the positional adjustment.  Also, a players' fielding ranking can change if the fielders at his position are better or worse in a particular season.  Carlos Lee plays at positions (LF and 1b) which are the usual spots for teams to hide their worst fielders. 

Furthermore, as we have said here on occasion, a larger sample size is required for accurate fielding metrics than normal hitting/pitching statistics.  One season of data probably is more accurate for hitting stats than fielding metrics.  Some would argue that you should use 2 or 3 years of advanced fielding data to get a better idea of a defenders' skill level.  Advanced defensive metrics are subject to regression to mean, just like other stats.  But there is more "noise" in the fielding data, which means that a larger sample size is required.  For example, the coding of batted balls as fly balls and line drives, and  how hard they are hit, is subject to some degree of human error.  In a large enough sample, these errors will tend to wash out.  Sometimes all we have to review is one season of fielding data, and I'm not saying you have to ignore it---but keep in mind that it may have a larger error band around the sample average.

So, what should we take from this discussion of Lee's 2011 fielding metrics?

First, look at UZR and DRS before you make a conclusion based on TZ.  Also, for position players, my personal preference is to trust Fangraphs WAR more than Baseball-Reference's WAR.  You may or may not agree with my preference, but if you see a difference between WAR values that are referenced in the media, keep in mind that the defensive metrics for B-Ref vs. Fangraphs can produce significantly different results.


Second, it is possible that Lee has made a concerted effort to improve his defense this year.  Based on what I have seen in games, Lee seems to be fielding the ball better and with more hustle.  Therefore, I wouldn't rule out the possibility that some of the improvement in fielding metric results is due to an actual improvement in skill and effort.  Perhaps Mills' visit to Lee's home in Panama has something to do with this.

Third, maybe Lee wasn't quite as bad a fielder as we thought.  Based on Lee's 2010 UZR results--the worst of his career--we thought Lee had become one of the worst fielders in baseball.  Given his advancing age, it's easy to see how we could jump to that conclusion.  But maybe we are seeing a regression process over a sample of multiple years.  From 2008 - 2011, Lee averages about -4 per year.  This would make him somewhat below average, but nowhere near the worst LFer over that period.