“Statcasting” the Giants Outfield Defense

Statcast is probably the greatest advancement in MLB during my lifetime, and quite possibly ever. The fact that literally everything that moves between the lines of a baseball field can now be simultaneously measured, tracked, and timed, is pretty unbelievable. It’s also difficult for me to wrap my head around.

While the public has fallen in love with certain aspects of Statcast data more than others (I’m thinking exit velocity for hitters and spin-rate for pitchers are the current darlings), there’s a couple of overlooked areas I really enjoy scouring on the ever-growing Baseball Savant website.

Expected stats for hitters is a newer phenomenon, but one that really intrigues me. Basically, the system can now predict a player’s average, slugging, and weighted on base average (wOBA) just by tracking his exit velocity and launch angle. That’s pretty incredible, and it can help tell a lot more about a player’s season than by just looking at his BABIP.

According to xStats, Joe Panik was one of the most unlucky hitters in MLB last year, with an expected avg of .285 & an xwOBA of .317. Both of those numbers were over 30 points higher than his actual production, indicating he was actually much closer to a league-average hitter than we all realized. On the flip-side, Alen Hanson’s xwOBA of .236 was 60 points lower than his actual numbers, and among the very worst in the game. What does that mean? Well, if Hanson doesn’t make better quality of contact this season, we might expect his numbers to plummet.

But I didn’t come here to talk about offense. No, the topic of this post is defense, namely of the outfield variety. Statcast has added quite a few new metrics and graphics to its fielding and positioning leaderboard sections, but it’s the trio of outfield defense categories that I’ve been heavily scouring this winter. I’m talking about Outs Above Average (OAA), Catch Probability, and Directional OAA. This is the meat and potatoes of how Statcast tracks outfield play, assigning a catch probability for every ball in the air. If the player makes the catch, he gets a certain percentage of points. If he doesn’t he loses a percentage.

Here’s a quick example. Say Steven Duggar plays a flyball that had a 30% catch probability (the system would call this a 4-star opportunity). If he makes the play, he gets credit for the remaining 70% (or 0.7 of an out). If he doesn’t make the play, he’s docked the 30% (or -0.3 outs). Simple enough, right? Generally, the elite players usually post a season total north of 20 Outs Above Average. In the first three seasons of data, there have been 8 such 20+ OAA performances. On the flip-side, the worst defenders usually get credited with 20 outs below average (-20 OAA).

According to Mitchell Lichtman, who works with Statcast and goes by the handle Tom Tango, or @Tangotiger on Twitter, a defensive “out” is worth roughly 0.8 runs. It takes about 10 wins (9.7 in 2018) for a player to earn one WAR. So, 12 outs (OAA) x 0.8 = 9.6 runs, or almost exactly 1 WAR. In 2018, 8 OF’s posted OAA totals of 12 or more, topped by Lorenzo Cain’s +22. So, according to Statcast, we can say that Cain gave the Brewers 17.6 runs of defensive value, good for 1.8 WAR. Baseball-Reference has him at 20 fielding runs, which is pretty darn close. Compare that to Nick Castellanos, who posted a dead-last -24 OAA, and Cain was 36.8 runs, or nearly 4 WAR more valuable. That’s a pretty decent gap!

The problem with well-known advanced defensive metrics like Defensive Runs Saved & Ultimate Zone Rating are their sample reliability. According to most analysts, a season’s worth of DRS or UZR data is hardly enough to tell you anything about a player’s defense. I’ve also noticed that defensive ratings can be quite volatile from year to year. It’s easy to understand this when we consider that all defensive measurements pre-Statcast were accumulated mostly through play-by-play data. It’s just not very reliable… yet these are the metrics used in calculating player WAR. I have a feeling that won’t be the case for much longer, as we see more and more advancements in Statcast.

So I wanted to know just how much the Giants could stand to improve in their outfield defense, which hasn’t been up to par for a few years now. My eyes and memory definitely knew things could be better, but the numbers are pretty revealing. Since Statcast started posting OAA totals in 2016, the Giants have been worth -37 outs. That’s a WAR score of -3.0, essentially a win subtracted per season. That’s not good!

Here’s the per-season breakdown in spreadsheet format:


Screen Shot 2019-02-20 at 10.08.40 AM


Screen Shot 2019-02-20 at 10.08.48 AM


Screen Shot 2019-02-20 at 10.08.56 AM

It’s not a pretty picture, especially for a team that relies on pitching and defense to compete. It also makes me wonder a bit more why Zaidi decided to non-tender Gorkys, who is literally one of the only bright spots defensively over the past 3 years. Metrics like UZR & DRS don’t like him as much, but that still seems like a move that could come back to haunt the Giants a bit.

There’s definitely an argument to be made here about playing Duggar and Mac as often as possible. Duggar played exactly 1/4 of a season in the majors last year and earned 4 OAA. Simple math says that’s 16 OAA over a whole season, which would have tied him with Billy Hamilton and Adam Engel for 4th in MLB. That’s big time.

But that was then, and this is now. So how does the current crop of outfielders competing for spots in Arizona grade out?

Screen Shot 2019-02-20 at 10.35.33 AM.png

There’s a good chance this group is going to struggle with the bats, but on paper it looks like a solid improvement defensively to what we’ve seen since 2016. Just having Duggar in CF for a full season is a huge upgrade itself, but there’s some sneaky good defenders mixed in here. Parra & Maybin are at the age where their defense could crater. Parra may be on that track already, but Maybe was in the 85th percentile for sprint speed last year. That makes me believe he’s still got decent OF range, and the numbers seem to back that up. Still, I think we’d all feel better if the Giants added one more proven player to this group.

So, who’s available? Funny you should ask.

Remaining Free Agent Outfielders

Screen Shot 2019-02-20 at 10.41.53 AM

Wow. Ok, don’t start in with the “Giants don’t need Bryce Harper…” comments. They absolutely need him. For the life of me, I can’t figure out why the Nationals played him in CF for 477 innings last season. Despite that, it sure looks like you wouldn’t want him playing RF at Oracle Park. His -26 Defensive Runs Saved mark backs that up. Actually, what worries me more than anything about Harper’s OAA score is that his “expected catch percentage” of 91 was tied for the highest in MLB. That means, on average, the plays he had to make were statistically considered among the easiest in baseball. But he only made 87% of those plays.

On the others side of the coin, Carlos Gonzalez’ expected catch % of 84 was the lowest in the league (tied with teammate Gerardo Parra), and he was able to convert 85% of his opportunities. What does that mean? Well, I think it means CarGo could probably hover around league average as a RF in San Francisco.

So, what about trade options?

Here’s a few familiar names I cherry-picked, with OAA totals (and MLB rankings) included.

Josh Reddick, Astros | +7 OAA (T-16th)

Adam Duvall, Braves | +6 OAA (T-21st)

Aaron Altherr, Phillies | +3 OAA (T-37th)

Kole Calhoun, Angels | +2 OAA (T-41st)

Kevin Pillar, Blue Jays | +1 OAA (T-50th)

If I’m Farhan Zaidi, and I end up losing out on Harper, I think I’m going to be making a lot of phone calls to Houston, in hopes of reacquiring my old pal Reddick.

If the future of Giants baseball is about making “good baseball moves,” then continuing to improve the outfield defense is one very accessible avenue. And thanks to Statcast, we don’t really have to throw darts anymore.

2 thoughts on ““Statcasting” the Giants Outfield Defense”

  1. It’s called rate. Hernandez earned 1.0 WAR in 532 attempts on 12 OOA. Mookie Betts has 12 OOA in 278 attempts. Betts is a very good defensive fielder. Hernandez, in his best year, was in the 63rd percentile based on his RATE. That’s above average, but that’s hardly ‘defensive wiz’ and does not offset his abysmal 14th percentile exit velocity giving him an xwOBA in the 22nd percentile. The guy can’t hit, no matter how many effusive articles and tweets Schulmann dropped on the guy last year.

    Hernandez is not the only one who can’t hit. Duggar hasn’t shown he’s got a serious MLB bat and I expect a .245 season out of him unless he can put a little more oomph in the bat. But Duggar is an elite (Top-10) fielder already. And he may grow with the lumber whereas Hernandez has failed to seize anything but a limited bench role in his career.

    As for Harper, you’re wrong and will continue to be wrong. You just keep not accounting for Oracle. I’ve gone over his batting and normalized it to the weather (20 degrees cooler on average) and park factors.

    The bottom line that his biggest HR spots are the same as Belt’s, which is Triples Alley and center-right and he will LOSE TEN FEET ON EVERY FLY BALL. There’s a reason that Belt leads all first basemen over the past four years in doubles & triples despite having hundreds of fewer ABs because of injuries. Harper will be the same instead of playing in a park that’s very favorable to him. We’re talking going from a 1.217 HR factor to a .500 HR factor for LHs!

    You’re a smart man, but I do not understand why you must keep on believing he’s not going to be affected by the park. It’s going to be severe, like Belt who is, at a neutral MLB park a 30-35 hitter based on his road performances the past few years. But when he’s at Oracle, he’s a 10-12 HR hitter. And the rest are doubles and triples. And that will be Bryce Harper at home — a poor-fielding, under-performing power hitter.

    You want power, it’s got to be RH. Like Pence used to be, who did suffer at Oracle. But not so bad as the LHers.

    1. I love your analysis. I am very numbers oriented and your work makes so much sense. I use statistics to try to predict where HRs will land and when. While my statistics are crude, with them I grab a HR roughly every 10 games each year. Mainly in McCovey Cove but also inside parks with my mitt. McCovey Cove Dave

I'd love your feedback!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s