Gold Glove Award Voting Is Changing for the Better but Still Needs Fixing

Zachary D. Rymer@zachrymerMLB Lead WriterOctober 28, 2013

April 13, 2012; Los Angeles, CA, USA; Los Angeles Dodgers right fielder Andre Ethier (16), starting pitcher Clayton Kershaw (22) and center fielder Matt Kemp (27) receive their 2011 Gold Glove awards before playing against the San Diego Padres at Dodger Stadium. Mandatory Credit: Gary A. Vasquez-USA TODAY Sports

The 2013 Rawlings Gold Glove Award winners will be announced on Tuesday night, so it's time for your annual warning that there will be griping when the hardware is given out.

There usually is with the Gold Gloves. Other baseball awards foster arguments, but no other award has a history of missing the mark quite like the Gold Gloves. For an award that's supposed to honor the best defensive players in the game, it sure has failed to do so quite a bit over the years.

Call it the inevitable upshot of an imperfect process. But if there's a bright side, it's that the process of selecting the Gold Glove Award winners is trending in the right direction.

This year's selections will be the first for a new selection process that is several years overdue and very much welcome. Exactly how the selection process is different this year is information that's been available for a few months—this article is another thing that's overdue to a certain extent. But if you need a rundown on the changes, you've come to the right place.

If you're also looking for a take on how an already improved process can get even better, you've come to the right place for that as well.

How It Used to Be and How It Is Now

/Getty Images

In the past, Gold Glove ballots were distributed to managers and coaches in each league, and they were tasked with picking out the best defensive players at each position. The only thing they couldn't do was vote for their own players. Aside from that, there was nothing in place to influence for whom they voted.

That's what's different now.

In March, Rawlings announced that it would be collaborating with the Society for American Baseball Research (SABR) to add a sabermetric component to the Gold Glove selection process. The goal of the collaboration was summed up like so:

...SABR will develop an expanded statistical resource guide that will accompany the Rawlings Gold Glove Award ballots sent to major league-level managers and coaches each year. In addition, SABR will immediately establish a new Fielding Research Committee tasked to develop a proprietary new defensive analytic called the SABR Defensive Index™, or SDI™. The SDI will serve as an “apples-to-apples” metric to help determine the best defensive players in baseball exclusively for the Rawlings Gold Glove Award and Rawlings Platinum Glove Award selection processes.

In August, Rawlings and SABR formally announced the list of people who would form the research committee, as well as the solution they had come up with for incorporating statistics.

Here are the names, complete with their ranks.

  • Committee chair Vince Gennaro, SABR President and author of Diamond Dollars: The Economics of Winning in Baseball
  • Sean Forman, founder of Baseball-Reference.com
  • John Dewan, owner of Baseball Info Solutions
  • Bo Moon, executive vice president and co-founder of Bloomberg Sports
  • Chris Dial, author and recognized expert on defensive metrics
  • Michael Humphreys, author of Wizardry: Baseball’s All-Time Greatest Fielders Revealed
  • F.X. Flinn, SABR board of directors

John Dewan, he of Fielding Bible fame, is probably the biggest star of the seven. That said, anyone who's familiar with the sabermetric Illuminati should recognize him and maybe a couple of the others.

As for the SABR Defensive Index, what these seven fellows came up with is something that "draws on and aggregates" five different statistics.

Three are derived from batted-ball and location-based data, and they are:

  • Defensive Runs Saved (DRS)
  • Ultimate Zone Rating (UZR)
  • Runs Effectively Defended (RED)

DRS and UZR are probably the two most widely known and widely used advanced metrics for evaluating defense, and both basically attempt to do the same thing with essentially the same methodology.

UZR and DRS attempt to quantify how many "runs" a player is worth on defense above or below the average player. Both metrics use play-by-play data from Baseball Info Solutions to make their calculations, taking into account the "zones" where players play and what happens within those zones in light of what could/should happen within those zones.

The differences between the two stats are in the subtle differences in their calculations and adjustments. Anybody who's interested in exactly how UZR comes to be can go check out FanGraphs' lengthy breakdown. Likewise, anybody who's interested in exactly how DRS comes to be can go check out the FAQ at FieldingBible.com

As for the RED metric, I've actually been unable to find a breakdown of how it works. But if it's based on batted-ball and location-based data, then the concept is the same as that of UZR and DRS: assign players "zones" and make determinations based on what happens within those zones.

The other two metrics that make up the SDI are based on play-by-play data. And they are:

  • Defensive Regression Analysis (DRA)
  • Total Zone Rating (TZ)

A primer on DRA can be read at Hardball Times. It includes the following summary:

DRA...systematically works through and determines the statistically significant relationships between traditional, publicly available pitching and fielding statistics and the actual number of runs allowed by a team. DRA yields formulas...that enable us to estimate the number of runs saved or allowed by pitchers and fielders (a) relative to the league average and (b) independently of each other.

Put simply: While batted-ball and location-based data effectively try to balance what happened against what could/should have happened, DRA draws conclusions only from what actually happened.

Primers on Total Zone Rating can be found at FanGraphs and Baseball-Reference. It takes play-by-play data from Retrosheet files and, in keeping with the general theme of all advance statistics, uses the data to quantify how many runs above or below average a given fielder is/was.

Individually, the five stats might paint completely different pictures of a given player's defensive value. But put them all together, and what you get is a group of statistics that leaves no stone unturned.

Per an FAQ on SABR's website, lumping all five statistics together allows the SDI to cover the following:

...a fielder's range, his throwing arm, his sure-handedness, his ability to turn double-plays (for SS and 2B), his ability to convert bunts into outs (primarily P, C, 3B, and 1B), scoops of throws in the dirt (1B), as well as the number of "excellent" and "poor" fielding plays. In addition, for outfielders, his ability to prevent runners from taking an extra base is also rated. For catchers, blocking balls in the dirt and stolen bases/caught stealing are also included in their ratings. Finally, for pitchers, we include his ability to control the running game by holding runners on base, in addition to many of the above criteria. (Pitch framing by catchers is not currently included in the defensive metrics that comprise the SDI.)

For the record, the five metrics that go into the SDI aren't given the same amount of weight. SABR says that the batted-ball and location-based statistics account for 70 percent of the SDI, whereas the play-by-play-based stats account for only 30 percent. The reason is that the latter take "more generalized approach" to evaluating defense.

Lastly, there's how much the numbers matter in the grand scheme of the selection process. Here's SABR's explanation:

The committee has created a simple way to convert the SDI—which is expressed in the number of runs a fielder "saves" his team—into 30 "votes" for each Rawlings Gold Glove Award awarded.

These 30 SDI-based votes account for 25 percent of the total process. The ballots of the managers and coaches account for the other 75 percent. And when managers and coaches get their ballots, Rawlings and SABR give them a "revamped statistical resource guide" that includes the SDI.

That's the short version—as short as I can make it, anyway—of how things have changed.

As for how all this is for the better, that part's much simpler.

Why This Is for the Better

Rafael Palmeiro won his second Gold Glove in 1998. When he won his third in 1999, he had only played at first base in 28 games all season long.
Rafael Palmeiro won his second Gold Glove in 1998. When he won his third in 1999, he had only played at first base in 28 games all season long./Getty Images

The initial decision to have managers and coaches in both leagues cast their votes for the best defensive players in the business wasn't a bad one. The mistake was never putting any real guidelines in place, as it left the door open for managers and coaches to make it up as they went along.

It's no wonder the Gold Gloves became less about honoring the best defensive players and more about, well, other things. The Gold Gloves came to honor highlight-reel-only defenders, popular players, veteran players and popular veteran players. And once the voters got fixated on a certain player, they tended to keep the Gold Gloves coming.

Matthew Leach of MLB.com hit the nail on the head back in 2010:

One joke goes that Gold Gloves are like Supreme Court appointments: They're hard to get, but once you have them, they're for life. Another often-heard wisecrack is that a given player didn't hit well enough to win a Gold Glove. Lines like that wouldn't have caught on if they didn't have an element of truth.

The part about Gold Gloves being given to good hitters is also true. Veteran middle infielder John McDonald recently told FanGraphs:

When you put managers and coaches in front of a ballot, offensive numbers factor into their Gold Glove awards. I know that, because I’ve heard coaches talk about it. ’He didn’t hit well enough for me to vote for him.’ That blows me away.

Obviously, having such major awards being determined by such silly motivations was a problem that needed correcting.

Especially, mind you, in a day and age where advanced defensive metrics have come out of the shadows. It made no sense for the Gold Gloves to continue to treat defensive evaluation as an inexact science when defensive evaluation is more of an exact science than ever before.

But now here we are in 2013. For the first time in the history of the Gold Gloves, numbers have found their way into the process. Having the numbers influence the voting is good enough in itself. Having the numbers inform the managers and coaches doing the voting is even better.

And yet, there are still some cracks.

Why It's Still Not Good Enough Yet

April 2, 2012; St. Petersburg, FL, USA; Baltimore Orioles designated hitter Chris Davis (left) smiles with center fielder Adam Jones (10) after hitting a 3-run home run against the Tampa Bay Rays in the seventh inning during opening day at Tropicana Field

How do we know there are still cracks in the system?

The first sign of trouble was the problems with the list of finalists.

The finalists for the Gold Gloves were revealed on Friday. Exactly how the finalists did in 2013 according to the SDI is unknown. The SDI calculations will be made public eventually, but for now we're all in the dark.

In taking the time to cross-reference the list of finalists with the testimonies of Defensive Runs Saved and Ultimate Zone Rating, however, I can vouch that there aren't many complaints to be made about the list of finalists.

It's just that some of the complaints that can be made are loud complaints.

For example, there's Boston Red Sox first baseman Mike Napoli. Per FanGraphs, he led all qualified American League first basemen in both UZR and DRS. Yet somehow, he's not a finalist for a Gold Glove. That he didn't make the cut over Baltimore Orioles slugger Chris Davis, who was below average in the eyes of UZR and DRS, is outrageous.

Davis isn't the only questionable Gold Glove finalist to come from the Orioles. Center fielder Adam Jones and right fielder Nick Markakis are also finalists, despite the fact they were two of the worst outfielders in the league by UZR and below average by DRS, according to FanGraphs.

Then there's New York Mets center fielder Juan Lagares. Only the Milwaukee Brewers' Carlos Gomez had Lagares beat in UZR and DRS among National League center fielders, according to FanGraphs. Yet Lagares didn't get nominated over Denard Span and Andrew McCutchen, who the metrics say were very good but not quite brilliant on defense in 2013.

Elsewhere, Josh Donaldson could have been chosen over Adrian Beltre at third base in the American League. D.J. LeMahieu could have been chosen over Mark Ellis at second base in the National League.

You're free to make your own gripes, but these are the big ones because of what they tell us about the voting process.

That Davis got chosen as a finalist tells us that offensive production is still a factor. That Jones, Markakis, Span, McCutchen, Beltre and Ellis—all six of whom are past Gold Glove winners or generally acknowledged as great defensive players—are finalists tells us that reputation is still a factor.

It's as simple as this: If these things are still factors, then the process is still flawed. An award that rewards defense should care about defense and nothing else.

So allow me to propose a few notions.

Since statistics are the only thing that can view defense objectively, the easiest way to settle things would be to just give the SDI the only say in the matter. The players with the best SDI at each position get Gold Gloves, and that's that.

But since this would be about as silly as giving the MVP award to the player with the highest WAR, going so far as to make the SDI count for 100 percent of the selection process is a bit much. To use a contemporary baseball phrase, there should be some sort of "human element" at play.

So how about half SDI and half votes from managers and coaches? Rather than 25-75, it would be 50-50. If it comes to this, the Gold Gloves would evaluate defense as defense is best evaluated: half stats, half eye test.

Granted, making the SDI's role even bigger will require the SDI to be legitimately foolproof. It's up for debate as to whether that's even possible. As Matt Hunter of Beyond the Box Score put it: "You can't just combine a bunch of metrics, each of which are calculated differently, together and call it progress. That's not how math, or rational thinking, works."

A fair point, but it's perfectly OK to be optimistic about the SDI. The right people were chosen to create it, and they did zero in on the right stats. If they conclude that the formula needs tweaking, they'll surely tweak it. If they determine that the SDI needs to evolve to stay current with the times, it so happens they've already vowed to do so.

In the meantime, however, SABR and Rawlings need to seriously reconsider whether the SDI in itself is going to have the desired effect on managers and coaches.

“There will be some that would look at it,” Tampa Bay Rays manager Joe Maddon told the Los Angeles Times when he was asked about whether managers and coaches would consider the SDI when it came time to vote. “I think it would be a minority that would actually look at it.”

Maddon added: “I think staffs are still trained to go by the eye, what they see and who plays well against them."

There's more than just the SDI in the statistical resource guides that are given out to managers and coaches, but not much more. Per SABR, the guide only includes "advanced statistical metrics and player eligibility requirements" along with the SDI. 

Since managers and coaches across the sport lean so heavily on what their eyes tell them, simply presenting them with a bunch of numbers is only going to get them to turn their heads so much. It would be better if SABR and Rawlings presented with numbers and words.

The ideal system involves managers and coaches being presented with the five SDI statistics broken up individually with written summaries for how the scores came to be. The three batted-ball and location-based metrics could get particularly in-depth, as the sheer complexity of determining the numbers allows for them to pinpoint specific strengths and specific weaknesses. 

For example, John Dewan—who, once again, runs The Fielding Bible, which is responsible for the Defensive Runs Saved metricwas able to get pretty specific about Adam Jones in Volume III of The Fielding Bible. Via SI.com:

According to Dewan, no centerfielder over the past three years has taken more bad routes or broken in the wrong direction more often. Dewan also reports that Jones often gives runners extra bases (compared to the average centerfielder) on low-trajectory fly balls hit into the gaps and by mishandling balls after they fall for hits. 

If managers and coaches had summaries like these to go along with the numbers, they would be able to cross-reference the statistically informed opinions with their own opinions. Their final decisions would be products of the best of both worlds.

Perhaps it will happen someday. And if it does, maybe then the Gold Gloves will finally start going only to the players who deserve them the most, no questions asked.

For now, the process will have to do as is. It's not perfect, but at least the Gold Gloves have caught up with modern times.

If you want to talk baseball, hit me up on Twitter.

Follow zachrymer on Twitter


The latest in the sports world, emailed daily.