Buxton tops in new Statcast metric Outs Above Average

Byron Buxton checks every box defensively. He glides in the outfield, befitting his status as one of baseball's two fastest players. Buxton makes spectacular plays that look great, and he makes spectacular plays that look easy, too.
It was clear back in March that the Twins' outfield defense would be a strength, and that Buxton would be the primary man responsible for it. He passes both the eye test and the data test, and as we introduce our latest Statcast™ metric, Outs Above Average (OAA), Buxton is a wonderful name to have at the top of our leaderboard. Having saved +23 outs over an average outfielder -- and nearly 40 above those at the bottom of the list -- he isn't just baseball's smoothest-looking outfielder, he's also the best at tracking down flies, so says the data.
Given that Outs Above Average is a brand-new number, let's explain how it works. Prior to the season, we introduced Catch Probability, which took the distance an outfielder had to go and the time he had to get there (later including the direction as well) to put a difficulty number on each outfield ball. A soft fly with a 99 percent Catch Probability is virtually always caught, while a batted ball with a one percent Catch Probability is virtually never caught.
All year long, we've been accumulating Catch Probability numbers on individual outfield plays, and sharing leaders in Five-Star catches (1-25 percent), Four-Star (26-50 percent) and so on on our leaderboards. What we do with Outs Above Average is to compile the outcome of all of those individual plays -- over 300 in Buxton's case -- into one season-long number.
Let's show a specific example. On Opening Day, Buxton robbed Alex Gordon by making a diving catch on a ball with a mere 24 percent Catch Probability, meaning that under the same conditions of time, distance and direction, this is a ball that an average outfielder will fail to come up with 76 percent of the time.

By making the play, Buxton was awarded a +.76 on his season-long tally; if he'd missed it, he'd have been hit with a -.24 penalty, reflecting the plays made on that ball by other outfielders. By comparison, when Baltimore's Adam Jones was given an easy 99 percent catch opportunity against Cleveland's Edwin Encarnación on Sunday, it would have only given him a +.01 if he'd caught it, because it was a play any outfielder is expected to make -- and it hit him with a -.99 because he did not.
So we take the +.76, and add it to the +.82 Buxton got for an 18 percent Catch Probability play he tracked down off the bat of Salvador Perez on Sunday, and a +.01 for an easy 99 percent fly ball hit by Khris Davis on May 3, and we dock him -.40 for failing to come up with a 60 percent Mookie Betts play that fell for a single on June 26, and so on and so on across every ball hit his way this year. This accounts not only for the number of plays he makes (or doesn't), but the difficulty of them. It also allows us to free metrics from the tyranny of "hits" and "errors," because it's simply about "was the play made or not and how hard was it to make," regardless of scoring.
"Us outfielders have this thing where nothing falls but raindrops," Buxton said after the Gordon catch on Opening Day. "We take that to heart and want to be the best outfield out there."
Nearly a full season later, that's exactly what's happened. Thanks largely to Buxton, as well as Max Kepler's +7 and Zack Granite's +3, the Twins do indeed rank at the top of our team OAA leaderboards.
Team Outs Above Average (OAA) leaders in 2017
+29 -- Twins
+18 -- Red Sox
+17 -- Rays
+13 -- Mariners / Reds
Behind Buxton, you'll find some pretty familiar names on the individual OAA leaderboards, like Ender Inciarte, Betts and Lorenzo Cain.

Now, what about those other three columns? Let's focus on Kevin Kiermaier to show how this accounts for varying levels of playing time -- important because Kiermaier missed more than two months with a broken hip this summer. OAA is a counting stat, in that the more you play, the more opportunities you have to accumulate value (or not), and that Kiermaier is back up in the top five anyway says a lot about his skill. But for most players, limited playing time will prevent compiling Outs Above Average. We can still credit them with the three Catch Percentage columns on the leaderboards.

Expected Catch Percentage shows how many plays an average outfielder would be expected to come up with based on the difficulty of the batted balls hit to the outfielder in question. This is a good proxy to see if the fielder is getting easy or difficult opportunities. Of the 118 outfielders with 100 chances, you will see that Mike Trout's Expected Catch Percentage of 90 is the highest, meaning an average outfielder catches 90 percent of the balls Trout has seen. This shows he's getting the least challenging opportunities sent his way, which helps explain his lack of great catches this year, while Robbie Grossman, at 75 percent, is getting the toughest.
Actual Catch Percentage shows the production of the actual fielder on the balls hit his way. On Trout, for example, given the difficulty of batted balls, we'd expect him to make the catch 90 percent of the time, but in actuality, he's making the catch 88 percent of the time. Grossman is expected to make the catch on 75 percent of his plays, and that's exactly what he's doing -- 75 percent.
Catch Percentage Added is the most important, because it's the difference between the Expected and Actual numbers. For example, Jason Heyward and Denard Span have each been given the same difficulty of opportunities, being expected to catch 85 percent. But since Heyward's Actual Catch Percentage is 89, he's added four points of value; since Span's Actual is 81, he's subtracted four points of value. Heyward is eight points more valuable on a rate basis than Span, and +19 (comparing his +9 to Span's -10) OAA on a cumulative basis.

You can use either way to judge season improvements, too. For example, Tommy Pham has gone from -6 OAA to +3, in part because a year after underperforming his Expected Catch Percentage by three points, he's now overperforming by two. But Joc Pederson has dropped from +2 to -7, as he's now underperforming by four points after being essentially even last year, and Kevin Pillar has gone from elite (+15) to average (-1), thanks to a combination of fewer tough chances (from 84 percent Expected Catch Percentage to 87 percent) and a drop from overperforming by four points to being basically scratch.
What doesn't it include?
Right now, Outs Above Average is a range-based metric to show outfielder skill at collecting fly balls. It doesn't currently include other defensive plays like throwing errors or non-fly miscues, like when Kiermaier overran a Whit Merrifield grounder that turned into a home run. It also currently is not split by outfield position.
What about "wall balls?"
As explained by MLB.com's Tom Tango, we've defined a wall ball as any ball caught within eight feet of the wall. For plays made, the fielder retains the credit; for plays not made, the play is removed from the sample, a change which is now reflected in our original Catch Probability leaderboards. This helps a player like Boston's Andrew Benintendi, who was being unfairly hurt by the Green Monster, by removing those uncatchable plays. Red Sox left fielders are collectively -6 at home and -2 on the road, while visiting left fielders are +1 at Fenway, which gives us confidence they're being treated fairly. This offseason, there's a plan to revisit wall balls to attempt to more thoroughly account for their unique properties.
Outs Above Average (OAA) is now available on BaseballSavant both in team and invididual leaderboards, and will update once daily.