Sunday, July 31, 2005

Defensive Rating Inefficiencies

Baseball was seemingly invented for statistical purposes. One can absolute be buried by batting and pitching stats. There are batting stats to overule batting stats to disprove other batting stats. Yet, when it comes to analyzing defense, statistical analysis is still somewhat scant.

Sure, you have the old reliable - fielding percentage - widely regarded as highly inaccurate. We have range factors and zone ratings, which surely give us a better picture of defensive performance. Then there is Defensive Efficiency, essentially the percentage of balls in play that a team turns into outs.

Does this translate into accurately reflecting a team's defensive prowess? Take a look at the Top 5 A.L. defensive teams according to defensive efficiency the last three years (2005 to date):

2005 - Oakland, Chicago, Cleveland, Seattle, Anaheim
2004 - Tampa, Seattle, Oakland, Chicago, Boston
2003 - Seattle, Oakland, Tampa, Chicago, Anaheim

Interestly, two of the top five in 2005 (Chicago and Anaheim) are also in the top five in fielding percentage. In 2004, Chicago and Oakland managed top fives in both categories, while in 2003 Chicago and Seattle managed the feat (Seattle actually leading the league in both categories).

I was not suprised to see Seattle number one in 2003, when they had essentially three centerfielders in the outfield (Winn, Cameron, Ichiro) and Tampa Bay, too, for all its faults, has speedy and proficient defensive players. Given their successes, Anaheim and Oakland being regular top fivers is not all that eyeraising. However, isn't interesting that the White Sox, a team that supposedly just committed to speed and defense this spring, have been solidly in the Top 5 in both Defensive Efficiency and Fielding Percentage all the way back to 2002?

Once adjusted for diffences in ballparks (courtesy Baseball Prospectus) the range in Defensive Efficiency from first to last is roughly .720 down to .670, year in year out. What does that translate into? Basically, it is five balls in play out of 100 that the best defensive team converts for outs that the worst team lets fall in for hits. Or, more to the point, one ball per game. What does one out mean? Not much in a 7-2 loss, probably a lot more in a 2-1 loss.

Still, is Defensive Efficiency the be all and end all in ascertaining defensive prowess? How do we account for missing the cutoff man, for an outfielder cutting off the gap and turning a probable double into just a long single? Or maybe even more importantly, keeping the runner on first from going first to third on a single. I imagine many, if not most, teams are intricately tracking evey ball in play, every exchange, every misstep, and attempting to compile this mountain of data into some usable measure, but they are not sharing that with the public.

Until such a time as the major leagues do provide this data, we may all be left trying to cobble together a clear picture of defensive worth from a myriad of old and new statisical measuring devices.