How teams might evaluate Harper's defense

Outfielder's -12 Outs Above Average was fifth-lowest in '18

November 21st, 2018

had a good 2018 season. Not a great one, but a good one. He slugged 34 homers, scored 100 runs and put up a .249/.393/.496 line that was 35 percent better than average, or 17th among qualified hitters. In the second half, it was a sparkling .300/.434/.538, ninth-best in the Majors.
But there's more to baseball than hitting, and in terms of overall value, Harper's season wasn't exactly what he'd have wanted heading into free agency. At FanGraphs, he was worth 3.5 Wins Above Replacement, still above average, but tied for 49th. At Baseball-Reference, it was merely 1.3 Wins Above Replacement, tied for 185th. 
That disconnect between his strong hitting and weak overall value comes down mostly to defense, where every metric had him as a negative. Defensive Runs Saved scored him a -26, worst of any outfielder who doesn't call massive Coors Field home. Ultimate Zone Rating scored him a -14.4, at the bottom of the list. Statcast™'s Outs Above Average, which for the moment includes only range and not arm value, puts him at -12, fifth-worst
Clearly, there's still some variation between several imperfect defensive metrics and how they feed into a player's overall value, but they all agree that Harper's defense in 2018 was poor, and the worst of his career.
But Harper is only 26, and still athletic. So let's try to figure this out in the same way a team might, by trying to answer this question: Was Harper's defensive downturn an indicator of lack of skill, or something else? That is, can a team hope for improvement? When FanGraphs' Jeff Sullivan investigated this, he found that recent examples of defensive decliners rebounded by about a third the next season. 
Let's attempt to answer that question by asking a few other questions.

Did he get slower? (No.) 
This is the first thing you would think of when you're trying to figure out why an outfielder is no longer getting to as many balls, especially when you remember that Harper had a scary-looking knee injury late in 2017. Maybe his peak speed just isn't what it used to be?
But the data doesn't back that up. Statcast™ measures top running speed with a metric called Sprint Speed, measured as "feet per second in a runner's fastest one-second window," with 27 being Major League average. We've begun to use this to try to see what trends we can see in certain players, but Harper's Sprint Speed has stayed relatively constant. It doesn't seem like this is the issue.
2015: 27.8 ft/sec, 72nd percentile
2016: 27.2 ft/sec, 59th percentile
2017: 27.7 ft/sec, 70th percentile
2018: 27.5 ft/sec, 62nd percentile
Harper even stole 13 bases, his third-highest total. In addition, his fastest-tracked run (as a hitter or baserunner) in the four years of Statcast™ came in the final month of the 2018 season, when he hit 30.4 ft/sec going first-to-third on an single. 

Was it because he had to play more center field? (Doesn't seem so.)
Thanks to injuries to and , along with the batting struggles of Michael A. Taylor, Harper played 477 1/3 innings in center, three times more than he had in the previous five seasons combined. As we showed earlier in the year, center field is tough simply because of the many extra opportunities to be exposed while playing there.
But this doesn't seem to be the reason either. Harper's performance in center and right were similar. There's nothing obvious about his position affecting this.
-5 OAA. 91 percent expected catches, 88 percent actual (-3 percent added)
-7 OAA. 89 percent expected catches, 85 percent actual (-4 percent added)
Did he get better or worse as the year progressed? (Not really.) 
As we noted above, Harper's offensive season wasn't exactly consistent. His first half was disappointing, and his second half was spectacular. Did his fielding follow the same track? Not really. A decent May stands out, but otherwise there's no discernible pattern here.
April: -3 OAA
May: +1 
June: -1 
July: -2 
August: -3 
September: -4 
Did he position himself differently? (Yes!)
Interestingly, Harper has moved considerably deeper in right field over the last three years. If we look at all right fielders who played against at least 500 batters in each season, we can see his trend clearly. 
2016: 283 feet deep, tied for shallowest of 71 right fielders
2017: 290 feet deep, tied for 18th-shallowest of 66 right fielders
2018: 295 feet deep, tied for 39th-shallowest of 80 right fielders
There were 36 right fielders who saw at least 500 batters in both 2016 and '18. No one moved back more than the 12 feet that Harper did, from 283 to 295. (The average in the group was +2 feet.)
That's interesting, but it's also not clear that it made a difference. We can use OAA to look at how players have fared in any direction, like we did when we showed how great was at going back. When looking at Harper, he hasn't shown a clear strength or weakness in any one direction. This might have some impact, but nothing we can clearly quantify. 

Did he just not receive difficult opportunities? (Yes!)
Not even the greatest defender can simply will difficult chances into existence. If he's positioned too well, the chances are easy. If he's too far away, the chances are impossible. Throw in the vagaries of batted-ball luck, and this can have a real impact. We looked at this in 2017 while trying to figure out why 's defensive numbers weren't strong, and we noted he'd received the easiest chances in the game that year.
We can do the same for Harper. There were 87 outfielders who received at least one chance per team game. No one had easier overall chances than he did, as 90 percent of the balls hit to Harper would likely have been converted by an average outfielder, tied with , and for the highest in the game. 
It's difficult to set yourself apart when the toughest chances aren't there. So that explains the lack of positives, perhaps. What about the negatives? 
Was the problem the easy plays, or the difficult ones? 
We're trying to get to the heart of the problem here: Did Harper have a harder time making difficult plays, or was it more of an issue making easier plays? Let's find the easiest 10 percent of plays, defined as those with a Catch Probability of 90 percent or more, and make that one bucket. Everything else, we'll put into a second, more difficult group.
Let's set a minimum of 30 "difficult" opportunities in each season, and see where Harper ranked. 
2016 (117 qualifiers)
Harper converted 98 percent of chances, where the average was 98 percent.
That was 61st of 117, in the 48th percentile.
Harper converted 51 percent of chances, where the average was 50 percent.
That was 58th of 117, in the 50th percentile.
2017 (109 qualifiers)
Harper converted 99 percent of chances, where the average was 98 percent.
That was 13th of 109, in the 88th percentile.
Harper converted 37 percent of chances, where the average was 49 percent.
That was 88th of 109, in the 19th percentile.
2018 (114 qualifiers)
Harper converted 97 percent of chances, where the average was 98 percent.
That was 95th of 114, in the 17th percentile.
Harper converted 33 percent of chances, where the average was 50 percent.
That was 105th of 114, in the 8th percentile.
So it's true that he had some issues with the 'difficult' plays, though when he robbed playing center field in June, it was on a ball that is caught only 39 percent of the time, his best-rated catch since 2016.
But what about that drop in the "easy" plays, from slightly above average down to poor? What happened there?
OK, so what did the catchable balls he failed to catch look like? 
Look, it's a long season. Over hundreds of chances, silly things happen now and then. It hurts your metrics and it looks bad, but it's not a reflection of skill, or lack thereof. For Harper, that play came last September, when Matt Carpenter "doubled." It happened, but it probably doesn't mean anything going forward. 

We watched 18 other Harper chances where he didn't make a play when the Catch Probability was 50 percent or higher, and a few trends stood out.
There were a few plays where he just seemed passive, not going for balls that with slightly more aggression likely could have been caught, like this single from August.

It's somewhat similar on this Zack Wheeler single, where Harper appears to decide it wasn't a ball he could get to, letting a fly ball with a 60 percent Catch Probability drop for a hit.
On some, Harper seemed to have poor reads or slow reactions. You can see that on this Danny Valencia single.

Something similar appeared to happen on this hit, too.
On this double in September, Harper wasn't able to get there in time and pulled up short of the wall, though the ball dropped on the warning track. It's easy to imagine another outfielder getting there and making it look easy.

On others, he got there with a glove on the ball, but just wasn't able to convert. Against Miami's , Harper was charged with his only non-throwing error of the year. 

Early in the year, he could have taken a double away from , but couldn't close the deal despite the ball hitting him in the glove. Late in the year, something similar happened against -- Harper got there, just couldn't hang onto the ball.
This could be considered good news, in that his ability to get to the ball hasn't disappeared. A little more aggressiveness, and slightly better execution when the ball hits his glove, and this looks like a different story entirely. That's what interested teams are hoping, anyway.
* * *
Now: Is any of that predictive of the future? We can't say for sure yet. Coming in the spring, we'll introduce some new Statcast™ metrics that will allow us to dig a little more deeply into things like reaction time, burst, and route running. A very preliminary look at those numbers shows that Harper didn't actually decline in those areas in 2018 from 2017, though it's too soon to say. 
On those 'easy' plays, it seems Harper still has the skill to get to many of those balls. For whatever reason, the execution just wasn't there in 2018. That was bad news for his defensive metrics. It might be good news if you're a team hoping he'll improve in the years to come. It doesn't mean he'll be great. It doesn't mean he won't be a first baseman at some point in the future. It does mean that he's still just 26 and enormously talented. Teams aren't just going to assume he can't play defense anymore.