Sabermetrics & Gold Gloves

In the past the Rawlings Gold Glove award has been given to players based on the vote of managers and coaches of other teams.  Over the years, this process was criticized when awards occasionally were handed out to players who didn't seem like the best recipients.  Sometimes the Gold Gloves seemingly were awarded as much on popularity and offensive production as defense.  And, Derek Jeter is a well known example of Gold Glove controversy.  Jeter has won 5 gold gloves, even though all advanced defensive metrics rank him as a very poor defender (Baseball-Reference shows Jeter at -145 to -154 over his career).

To Rawlings' credit, sabermetric oriented fielding metrics will contribute to the gold glove selection process this year.  Rawlings provides a discussion of the criteria here. As mentioned in the link, SABR (the organization) brought together a committee of experts to develop a SABR Defensive Index (SDI).  After describing the SDI components, I will also speculate on how this might affect the chances of any Astros receiving a gold glove.

Rawlings will award 30 votes for each Gold Glove based on SDI results.  Rawlings estimates, based on typical voting patterns, that this gives a 25% contribution by SDI, relative to manager/coach votes.  (Note that managers and coaches cannot vote for players on their own teams.)  But this may not be the end of the sabermetric influence.  Rawlings will include the SDI ratings in the ballots it sends out to managers and coaches.  Some managers and coaches will ignore the SDI when they vote.  But it's reasonable to assume that some of the managers and coaches will take it into account.  Rob Neyer suggests that the bottom line contribution of SDI to the Gold Glove selection may be more like 40%.

The SDI will be comprised of both location-specific zone metrics and play-by-play metrics.  The former are based on location specific information, and frequently data regarding how hard the ball is hit, while the latter are more generalized methods based on inferences from play-by-play data and box score information.   The location specific advanced defensive metrics will be weighted 70% in the SDI, while the play-by-play methods will be weighted 30%. The play-by-play methods are less reliable, which is why they are given a lower weight, but they can be useful because they allow current players' fielding to be compared to players from previous years when the more sophisticated methods didn't exist.

The location specific component of SDI will be based on Defensive Runs Saved (DRS), Ultimate Zone Rate (UZR), and Runs Effectively Defended (RED).  You are probably familiar with DRS and UZR, as the most frequently cited advanced defensive metrics. Both are available at Fangraphs. RED is developed by Chris Dial, a contributor at Baseball Think Factory.  I have seen some of Dial's zone ratings in the past, but I don't have a link for his recent ratings.

The play-by-play component is based on Defensive Regression Analysis (DRA) and Total Zone (TZ).  TZ is shown on the Baseball-Reference web site.  (Note that DRS is the metric used by B-Ref for WAR calculations.)   I'm not familiar with DRA, but my understanding is that it was developed  by Michael Humphreys, as described  here by the Hardball Times.

So, how would this affect any potential Astros' candidates for the gold glove?  The most accurate answer is "I don't know."  But we can get some hints by looking at the SDI components available to us.   Most TCB comments about Astros potentially deserving a Gold Glove center around Brandon Barnes and Matt Dominguez.  So, let's take a peek.

Brandon Barnes, also known as the human highlight film, is a good defender, but I doubt that his SDI will be good enough to get him a Gold Glove.  Note that the Gold Glove is given to 3 outfielders in each league regardless of the position (CF, RF, LF).  I don't know how (or if) the SDI will take into account outfield position.  Barnes is +3.2 on UZR and +4 on DRS based on all OF positions he has played.  This ranks him 15th among AL outfielders.  Barnes ranks somewhere between 8th and 10th among CFers with at least 150 innings.  I don't have the data for Barnes based on RED or DRA, but his TZ is available.  He is -4 on the TZ metric (which may illustrate its lessor reliability), so that will hurt rather than help his candidacy.

Matt Dominguez is even more problematic when we take the saber  advanced metrics into account.  As we have discussed previously,  UZR and DRS give dissimilar results for Dominguez.  He is -1.6 on UZR and +11 on DRS.  My view is that this differential is due to differing treatments of out of zone balls and the way that the trend toward more defensive shifts affect the metrics.   However, even if you accept Dominguez's good DRS result, he ranks second among AL 3d basemen, far behind Machado with a +30 and +32 on DRS and UZR.  TZ has Dominguez rated as +1, and we don't have the rankings for him on the other two methods.  Machado is +27 on TZ.  I would be surprised if Machado isn't the SDI's 3d base gold glove.

Never say "never," but the saber "hints" aren't very encouraging for Dominguez and Barnes.  What do you think?

And while we're on the topic of defense, you may want to read this article on defensive value at Beyond the Boxscore, which is interesting in the context of discussions at TCB about the types of outfielders Luhnow acquires. The artilcle contends that the most undervalued type of player is the corner outfielder who is an average hitter and a good defender-- primarily because GMs are more willing to pay for offense than defense in the outfield.  The article points to the cheaper contracts given to players like Shane Victorino and Angel Pagan compared to the huge amounts committed to Josh Hamilton.  The article suggests that the reason for the undervaluation is the lack of certainty surrounding defensive metrics and the popularity of offensive production in drawing fans to the park.

This is a timely discussion, given the recent TCB comments about the future of players like Hoes, Grossman, Heras, DeShields, and Aplin as Astros corner outfielders.  These players have the potential to be plus defenders in corner OF positions, but they aren't likely to bring the level of offensive power we usually associate with those positions. (Perhaps this is unfair to Aplin who is still in A+ ball and could develop into a power hitter.)  We also know that the 2013 Astros' outfield draft selections appeared to place a premium on defensive ability.  Does this reflect the Astros targetting outfield defense as a market inefficiency?