13 November 2015

Did Manny Machado Deserve His Gold Glove?

Click here for Ryan Romano's archives.

On Tuesday, the Orioles saw something good come out of 2015. Manny Machado, who had made a triumphant return from knee surgery, won his second Gold Glove for his work at third base. Anyone who has seen him make plays such as these over his time in Baltimore knows that he fields his position incredibly. But did he truly lead the entire American League? The answer to that is kind of tricky.

By the standard (for sabermetricians) fielding metrics, Machado certainly had an outstanding campaign. He racked up 14 Defensive Runs Saved, to go along with 8.4 runs by Ultimate Zone Rating. While those marks couldn't match up to the ones his sensational 2013 season, they still drew the envy of almost everyone in the Junior Circuit. Almost — but not all.

Adrian Beltre, who finished in the top three behind Machado in the final voting, has always excelled at the hot corner. For his career, his play there has netted him four Gold Gloves. He hasn't acquired one since 2012, however, and despite struggling in 2013 and 2014, he regained his superiority this year — at Machado's expense. With 18 DRS and 11.4 UZR, Beltre led all qualifying third baseman in the AL; Machado had to settle for second and third, respectively (Josh Donaldson narrowly beat him in UZR).

These statistics have error bars, so a difference of decimal points — such as the one between Donaldson and Machado — shouldn't lead to any bold proclamations. That doesn't apply to Beltre and Machado, who have a four-run gap by DRS and a three-run gap by UZR. When both of the widely used defensive systems come to an agreement like this, it's hard to argue otherwise.

With that said, Machado's case doesn't end there. In addition to the aforementioned statistics, FanGraphs puts out Inside Edge Fielding data. It relies on scouts, rather than stringers, to take a more subjective look at the plays on the field. They grade all plays into six categories, each with a rough percentage range of the player making it: "impossible" (0%*), "remote" (1-10%), "unlikely" (10-40%), "even" (40-60%), "likely" (60-90%), and "routine" (90-100%).

*In the four years for which these numbers exist, no player has made an impossible play. 

The Inside Edge data flips the script — here, Machado came out ahead of Beltre:

Name 1-10% 10-40% 40-60% 60-90% 90-100% AL 1-10% AL 10-40% AL 40-60% AL 60-90% AL 90-100%
Manny Machado 9.1% 41.2% 57.1% 74.8% 96.9% 3.4% 25.3% 49.1% 72.9% 96.5%
Adrian Beltre 5.6% 12.9% 46.7% 74.3% 96.0% 3.4% 25.3% 49.1% 72.9% 96.5%

In the most trying times, Machado clearly reigned supreme; Beltre closed the gap a bit for the easier plays, but the disparity remains. Although both cleared the league averages in almost every regard, Machado did so to a greater extent than Beltre did.

Plus, Machado didn't receive as much help from his pitching staff, who gave him more hard-hit balls than Beltre's did:

Player # 0% # 1-10% # 10-40% # 40-60% # 60-90% # 90-100% AL # 0% AL # 1-10% AL # 10-40% AL # 40-60% AL # 60-90% AL # 90-100%
Manny Machado 2.4% 6.6% 3.4% 2.8% 9.6% 75.1% 1.7% 5.7% 3.4% 3.8% 8.4% 76.9%
Adrian Beltre 1.2% 4.4% 2.5% 3.7% 8.5% 79.7% 1.7% 5.7% 3.4% 3.8% 8.4% 76.9%

Overall, 8.1% of Machado's plays fell on the difficult side of the spectrum, compared to 12.5% for Beltre. So not only did Machado cruise through the biggest challenges, he encountered far more of them. In terms of both quantity and quality, the Inside Edge Fielding data comes down pretty firmly on Machado's side.

It would seem that this question boils down to the defining conflict of baseball's modern age: statistics versus scouts. Do we trust the systems that theoretically track every play objectively, but which don't take into account confounding variables like starting position? Or do we trust the scouts who can better gauge the nuance of a play, but who carry their own human biases? While I don't have an easy answer, I know what might mediate the discord.

Most baseball analysts (like me) use solely DRS and UZR to appraise defensive value. Those two don't act alone, though: Fielding Runs Above Average, from Baseball Prospectus, also measures players' glovework. FRAA uses a different formula, looking at plays made rather than basing its analysis on zones. Many other potentially influential factors — the ballpark and the base/out state, among others — also make their way into FRAA, whereas DRS and UZR (to my knowledge) ignore many of these.

FRAA agreed with DRS and UZR, insofar as it thinks very highly of both Machado and Beltre. However, it diverged from the narrative those two have established, in a major way. According to FRAA, Machado gave the Orioles 20.3 runs, and Beltre cost the Rangers 2.6 runs. By the (theoretically) most advanced publicly available defensive metric, Machado demolished Beltre by a 24-run margin.

Maybe you want to split the difference: Since Beltre wins in DRS and UZR and Machado wins in FRAA, they should tie in the end. But the duo didn't have a wide gap in the former two, and they certainly had one for the latter. As August Fagerstrom observed at FanGraphs, a weighted average here still puts Machado ahead.

Even if Beltre had taken home the Gold Glove, few people outside of Maryland would disagree with it. At age 36, he succeeded as much as ever; even if he decided to retire right now, his resume would surely transport him to the Hall of Fame. Machado nevertheless warranted the award, providing the best third base defense in the AL for the second time in three seasons. Keeping this up for several more years could earn him a ticket to Cooperstown as well.

No comments: