Why do advanced metrics dislike Michael Brantley's defense?

On Thursday, the finalists for the 2014 Rawlings Gold Glove Awards were announced, and Michael Brantely was recognized as a finalist in left field. Congrats, Michael! This is Brantley's first selection as a finalist, and a lot of Tribe fans think that's long overdue.

The next day, I published the first of two "Gold Glove Awards, Strictly by the Numbers" pieces over at FanGraphs, where Brantley also showed up on a list of three players worth noting in left field! Problem is, he was on the other side of the spectrum.

The advanced defensive metrics have long disagreed with the commonly accepted notion that Brantley is a plus defender in left field. Fans don't like this. Fans like Brantley, and they like Brantley's defense. Which is fine, because defensive metrics aren't perfect, and you're allowed to disagree with them and form your own opinion. This is probably a better than time than ever to talk about Michael Brantley's defense, so let's do that.

First, I should point out that the numbers don't exactly hate Brantley as a defender. This year, his defensive value in left was estimated to be worth about -3 runs per 1,000 innings (about a full season). For his career, it's -2. So, the numbers say he's slightly below average, and a couple runs over the course of a season is pretty small.

We know that with defensive numbers, a one-year sample is not meant to be taken as an indication of true talent level. This may seem odd, because a year is a long time, but the number of defensive chances each fielder receives in a single season is still pretty small in terms of how many repetitions it takes for data to be significant, so you want to look at several years before you start drawing any meaningful conclusions. Luckily, we have that with Brantley.

To gain a better understanding of where Brantley grades out, I created a custom FanGraphs leaderboard of left fielders and set a minimum of 1,400 innings played since 2012. This gives us a pool of 28 players, or just about one left fielder for each team in the league.There's still going to be some amount of noise in the data for the guys at the lower end of this spectrum, but Brantley has played 2,200 innings in left in that span, and so his results should be fairly indicative of what the metrics believe to be his true talent level.

We've got two widely accepted metrics - Defensive Runs Saved and Ultimate Zone Rating. Neither are perfect, but they're the best we've got and most can agree that, together, they do a pretty good job. From our pool of 28, I used a 50/50 split of DRS and UZR, prorating each figure to 1,000 innings so everyone was on the same scale. Brantley came out 16th, right beneath Alejandro De Aza and Nate McLouth and right ahead of Justin Upton and Carlos Gonzalez.

There are three main compenents that factor into these metrics: Arm, error prevention, and range. Here's Brantley's career profile, represented by runs above or below average and relative league rank, out of 28:

• Arm: +3.1 (5th)
• Errors: +1.0 (5th)
• Range: -6.2 (21st)

Brantley's got a great arm. We know this, because of the outfield assists. There’s really no arguing with the errors, because there’s really no arguing with the official scorer. These are the positive traits for Brantley, the defender. But you see the negative. It's the biggest number, and it's the reason Brantley grades out as a negative defender. Range is the most important skill for an outfielder to have, and it happens to be Brantley's worst skill.

The numbers say the guys with the best range are Starling Marte, Christian Yelich, and Eric Young. That sounds about right, because those guys are all really fast. Their range is estimated to be worth about +10 runs over the course of a season, equal to about one win. The numbers say the guys with the worst range are Jason Kubel, Jonny Gomes, and Domonic Brown. I don’t think many people would disagree with those names, because they're all very slow. Their range is estimated to be about -10 runs over the course of a season, costing their team about one win.

The metrics pass the eye test on either end of the spectrum. The same metrics which conclude Starling Marte has great range and Jason Kubel has poor range say Brantley’s range grades out to about -6 runs over the course of a season. It hasn’t been the worst, but it hasn't been good.

The thing some people don’t like about the range number is that it's just a number. With offensive statistics, we can see the singles, doubles and homers, so we know exactly what the inputs are, giving us transparency. Even with errors and the arm, we can see the isolated plays. With range, there's very little transparency without some sort of play log, because no two plays are created equal. But they can be close. And we do have spray charts, thanks to the wonderful folks over at Inside Edge, which show us A) the location and B) the difficulty of fielding each batted ball. Using these, we can overlay Brantley’s "missed plays" spray chart from the last three years onto the "made plays" chart of a player with better range, find the plays most similar and watch some footage, to see for ourselves. This can help give us some notion of transparency with regards to what's going into these numbers.

Let’s compare him to someone at +6 runs per season, like Alex Gordon. I know, I know - Gordon is a four-time Gold Glove Award winner! But that's more due to his arm and his lack of errors than it is his elite range. To put it another way: Gordon's range is as close to being league-average as Brantley's, they're just on opposite ends of the spectrum.

Here's a play Gordon made a couple months ago:

Here's a similar play Brantley didn't make in 2013:

That was a tough play for Brantley, and Gordon simply made an amazing catch. You can't really fault Brantley here, and you've just got to tip your cap to Gordon. The good defenders make these plays more often than the rest of the league, but you don't really expect it. Inside Edge gave both these plays a 1-10% chance of being made. Let's look at some easier plays.

Here's a ball from the 40-60% bucket that Brantley didn't catch, and a ball that is expected to be caught more often than not by those with plus range:

A similar ball, caught by Gordon:

These weren't the exact same play - Brantley's was hit a little harder and Gordon had to range a little further - but they're similar in terms of difficulty. See if you can kind of watch them at the same time. It seems clear, to me, that Gordon covers ground at a much quicker rate than Brantley - and remember, Gordon doesn't even have elite range. Gordon got to his ball, saving his team what would have likely been two runs. Brantley didn't get to his, and so his team gave up two runs. These don't go down as errors by Brantley, and they don't show up in the box score. But they show up in the advanced metrics and, over time, these things add up.

A ball that gets caught 60-90% of the time, caught by Alex Gordon:

A similar play Brantley didn't make:

Brantley's ball was pretty close to an outfield wall with which he isn't too familiar, but so was Gordon's. Gordon wasn't quite as close to the wall, but he had to range further. Again - slightly different plays, similar difficulty, different outcome. Indians broadcaster Rick Manning said, "nine times out of 10, Brantley catches that ball. Probably more." Manning might argue it should be in the next bucket up.

Another ball from the 60-90% bucket that Brantley couldn't get to:

And here's a much harder play, in nearly the exact same location on the same field, made by Gordon. Because this one had far less hang time, Inside Edge gave it just a 1-10% chance of being caught:

Anyway, you get the picture. Could you find some plays the other way around, where Brantley catches similar balls Gordon doesn't get to? Probably. But these clips alone aren't meant to be the basis of the argument that Michael Brantley isn't a particularly good defensive left fielder. The numbers are the argument. The clips just serve as support. You don't have to believe the numbers, but there are smart, unbiased, mathematical methods behind the numbers and when you watch some film, the numbers seem to make some sense.

Of all the skills which make up an outfielder's defensive ability, range is the most important. Alex Gordon looks like a left fielder with good, but not great, range. Michael Brantley looks like a left fielder with subpar, but not terrible, range. Gordon's good-but-not-great range, paired with his plus arm and error prevention, make him the best defensive left fielder in baseball. Brantley's subpar-but-not-terrible range, paired with his plus arm and error prevention, make him an average to slightly-below-average left fielder.

At least that's what the numbers say.