Sunday, November 07, 2010

Evaluating Third basemen in 2010

In recent posts, I have been looking at fielding measures for present Tigers and some free agents.  Because there is a good deal of disagreement between measures, I have been computing averages across measures instead of relying on just one measure.  Now, I'm going to start a series of posts looking at four measures and computing averages for all regulars in the majors in 2010.

The four measures are:

For all of the above measures, the result is the estimated number of runs that a player saved/cost his team with his fielding compared to an average player at his position.  All of the metrics can be found at  The exception is that TZ is not yet available for 2010.  I got that number at  TZ is called Rtot over there. 

I have a couple of goals in mind as I do this series.  First, I want to compute averages for each player in order to arrive at a conservative estimate of a fielder's runs saved/cost.  What we are doing in effect is regressing to the mean.  This will dampen the effect of a really large number on one measure which disagrees with other measures.

For example, if a player has a +16 on UZR and 0, 1, 3 on the other three measures, he'll have an average of +5.  So, instead of just going with UZR and taking +16, we have a less extreme estimate.  What if this particular player really is close to +16 though?  Don't we come up with a worse estimate by using the average?  Yes, but I think if we are going to plug these numbers into WAR, I'd rather err in the middle than err at the extremes.    

The other purpose of this exercise is to look at which players disagree most across measures.  This is important to know because I think we can be less certain about a player with a lot of disagreement across measures.

The third basemen with 1,000 or more innings (INN) in 2010 are listed in the table below.  I'll use Brandon Inge as an example:

DRS = +2
UZR = +3
TZ = +8
FSR = +16

The first thing you might notice is that the fans rated him better than any of the statistics did.  His average (AVG) is +7, so we estimate that he saved 7 runs compared to the average third baseman in 2010.  The second to the last column of the table gives the standard deviation (SD).  This number tells us how close the four estimates were to each other.  A low number indicates a lot of agreement between the four measures and a high number indicates a good deal of disagreement.  Inge's SD is 6.4 which indicates a moderate amount of disagreement.   

Table 1: Aggregating Fielding Measures for Third basemen in 2010

Here are some highlights:
  • The top third basemen according to the average is Evan Longoria (+15), Chase Headly (+14) and Ryan Zimmerman (+13).
  • The bottom three are Aramis Ramirez (-9), Michael Young (-8) and David Wright (-7).
  • The player with the most disagreement across measures is Jose Lopez: 15, 8, 17, -7.  The fans apparently did not think he was as good as his numbers suggested.  

No comments:

Post a Comment

Blog Archive


My Sabermetrics Book

My Sabermetrics Book
One of Baseball America's top ten books of 2010

Other Sabermetrics Books