Six years ago, Matt Kemp hit .324/.399/.586, good for a .413 wOBA* that was the fourth best in the Major Leagues that year. It earned him a second-place finish in the National League Most Valuable Player Award voting, largely because Ryan Braun's Brewers won 96 games while Kemp's Dodgers won just 82. But over the next five years, Kemp was besieged by injury and got traded twice, never coming close to repeating that production. As his 33rd birthday nears this fall, it was fair to assume his days as a star were over.
But after blasting his 10th homer in a 2-1 loss to the Angels on Wednesday, Kemp is hitting .345/.381/.608, which is good for a .416 wOBA. It would be the best mark of his career if he could keep it up all season; it's the 11th-best mark in baseball this year, behind a group of elite sluggers like Freddie Freeman, Mike Trout and Bryce Harper. (And, somehow, Zack Cozart.)
Where in the world is this coming from? And can it possibly sustain?
(*wOBA is Weighted On-Base Average, which is very similar to On-Base Percentage except it gives increasingly more credit to extra base hits as opposed to treating every time on base equally, as OBP does. The 2017 Major League average wOBA is .318.)

The easy answer is to go with the narrative that seems to be cropping up, that Atlanta's new SunTrust Park is the biggest launching pad this side of Coors Field. The problem with that is that it's just not true, as we explored in April. The rate of home runs per batted ball in Atlanta is only 20th in the Majors; the average distance on hard-hit fly balls is tied with several other parks for eighth most. There have been 2.5 homers per game in Atlanta, which is 14th most. It may be a better place to hit than Turner Field was, but it's not really any sort of offense-fueling outlier.
Even if it were, Kemp has actually been far more dangerous away from Atlanta, anyway. While he's hit well at home (.303/.352/.455, .344 wOBA), he's been far more effective on the road (.371/.400/.705, .462 wOBA) where eight of his 10 homers have come. Kemp hit homers on back-to-back days against Washington in mid-May, and otherwise he hasn't homered in Atlanta all year. So no, it's not that.

So we've talked about what it isn't, but what is it? Digging into the data, it seems to boil down to two things:
1. Kemp has become a more disciplined and legitimately better hitter, and
2. He's been the benefit of some amount of good batted ball luck.
Both those things can coexist at the same time, and we'll show you what we mean. First, let's show you that Kemp really is earning a lot of this improved production, and a big part of that is simply that he's hitting the ball harder. Last year, his hard-hit percentage (which we define as a ball hit 95 mph or harder) was just over 19 percent; and this year, it's nearly 25 percent. Over the past three years, Kemp hits .581 with a 1.300 slugging percentage when he hits the ball 95 mph or harder, so it's beneficial to do it more. He's definitely doing it more.
But unlike so many others around the game this year, it's not about elevating or joining the "fly-ball revolution," as Kemp's launch angle is slightly down this season. Kemp was among the dozens of sluggers we talked to this spring about why they thought they'd hit more home runs in 2016, and he made it clear that he had no intention of changing his swing path like Daniel Murphy or Yonder Alonso.
"I know if I hit the ball hard, it's going to go out," said Kemp. "I'm not going to change my approach. Everybody might be different, but for the power hitters I know, I don't think they're going to change their approach. If they square the ball up, most of the time, it's going to go out."

Kemp is right. Not only is he hitting it harder more often, he's hitting it more productively more often. What we mean by that is that while it's great to hit the ball hard, it's most important to do so at a productive angle, because a rocket straight up is still going to be an out. To that end, we identified six combinations of batted ball contact, three good for the pitcher and three good for the hitter. Last year, among 337 hitters with 50 balls in play, Kemp's 43 percent productive contact rate was 41st. This year, among 113 hitters with 50 balls in play, he's fourth, behind only Aaron Judge, Miguel Cabrera and Alonso.
Kemp has done that without increasing his strikeouts, which would be at their lowest rate since 2009, and that's key. In addition to measuring a hitter's actual wOBA, we can measure their "Expected" wOBA (xWOBA), too, which is a lot simpler than it sounds. Take the ball Kemp squared up a few days ago in San Francisco at 107.8 mph, the kind of ball that based on exit velocity and launch angle falls for a hit 73 percent of the time. In this case, it didn't, because Denard Span made a great play on it, but the quality of the defense has nothing to do with Kemp -- so, we credit him for the 73 percent Hit Probability.

If you do that for each batted ball over the course of a year, and include strikeouts and walks, you can come up with a seasonal xwOBA. Again, the real-world wOBA average this year is .318.
For Kemp, in 2015, his xwOBA was .336, and his actual wOBA was .325. So, what he "earned" was slightly above league average, and what he actually got was just a bit under that.
In 2016, Kemp's xwOBA was .350, and his actual wOBA was .333. So he hit a bit better than he did in 2015, and he slightly underperformed that by the same amount, perhaps in part due to declining speed.
In 2017, Kemp's xwOBA is .399, and his actual wOBA is .416, and now you can see what we're talking about. At .399, his combination of batted ball contact and strikeouts and walks has very truly been a big step up over last year, one of the 25 best in baseball, ahead of stars like Buster Posey or Francisco Lindor. But instead of underperforming as he had in the [ast two years, Kemp has overperformed this year by 17 points. You can see that in his .398 BABIP, which is 100 points higher than last year. He's hitting better, and he's also getting fortunate.
As an example, this weak grounder to third base in Miami on May 13 probably should have been an out (.070 expected average), almost certainly should have an error, but was scored a hit:

It's not really about luck, though. Kemp will never get back to that 2011 MVP-caliber level, simply because he's no longer got the speed he once had and he's become one of the larger liabilities on defense of any regular outfielder in the game. But the bat, at least so far, looks as dangerous as it did in Kemp's prime. It's not about Freeman or SunTrust or luck. It's about laying off the bad stuff, and crushing the good ones as hard as he can.