Steven Souza has bizarre reverse splits

Steven Souza Jr. hit his 23rd home run of the 2017 season on Thursday, a total that tops his previous career high by six and ranks second on this year’s Rays squad. While the 23 home runs by early August would be notable on their own, what is most notable about those long balls is a rather strange pattern.

They’re all against right-handed pitching.

Just a friendly reminder, Souza is a right-handed batter and in baseball it typically helps to see the ball out of the hand of an opposite-handed pitcher. Around the league this season, right-handed hitters have a wRC+ of 99 against lefties compared to a wRC+ of 92 against righties. For Souza, the numbers are a lot different:

Souza has still managed to be a productive hitter (107 wRC+) with the favorable platoon, but his power has all but disappeared. His isolated power is nearly four times higher against righties than lefties this season—that’s insane.

The first hypothesis would tend to be that this is just a fluke. Sure we’re in August, but once you start dealing with splits, the numbers shrink and can get noisy quickly, so let’s look a bit beyond just the surface.

Here’s Souza’s batted ball profile against RHP and LHP this season:

The 37.7 percent difference between HR/FB rates would certainly suggest some flukyness to this trend (no one runs a 37.7 HR/FB over a legitimate sample size), but there’s definitely something going on. The near-ten percent gap in fly ball rate and near-13 percent gap in hard hit ball rate suggest we should dig a little deeper.

If we take a step back and look at his career as a whole, it also appears as something to see about Souza’s approach versus righties and lefties:

The same exact pattern is there, just not quite to the same extreme.

Souza seems to generate a lot more power against right-handed pitchers than against lefties. So, is there a difference in how each type of pitcher approaches Souza?

Righties appear to stay low and away from Souza, while lefties tend to attack low and in to the 6’ 4” righty. Looking at Souza’s career isolated power zone profile, it doesn’t really explain much at first.

The part of the zone that righties attack most (low and away) is Souza’s worst zone. His ISO is literally .000 in the zone that is by far the most prevalent zone for righties to attack (zone 25 if we start in the top left and assign each a number).

Whiff rate doesn’t explain it either, as he has certain weaknesses (zone 21 against lefties and zone 25 against righties) where pitchers will pepper him, but there is not really much of a difference in terms of whiffs per swing against either righties or lefties. And again, his on base abilities appear to be quite similar with regards to the handedness of the pitcher—it’s the power that is the biggest gap.

How about the pitches he faces against RHP and LHP? Taking a quick glance at the pitch mix for lefties and righties facing Souza, righties tend to throw more sliders and less changeups. However, that once again doesn’t do much for us, as Souza has his highest ISO against changeups among all pitches in his career—the pitch lefties throw him more!

In the end, I think it’s fair to say the original hypothesis—and most boring possible answer—is the correct one. This is probably just a small sample size fluke. Take a look at Souza’s fly balls per ball in play split for his career:

The vast majority of the gap comes from this season, as the two figures were basically even in 2016 and he actually hit more fly balls against lefties than righties in 2015 (he had only 26 plate appearances in 2014, so that number is a tiny, tiny sample). A good chunk of the aforementioned career splits can be explained by 2017, which shouldn’t be too surprising given that Souza is still just 353 games into his MLB career.

If Souza gets to career game number 500 and the RHP/LHP splits are still as loco as they are right now, maybe we’ll be due for a second look at the underlying cause. For now it just seems like some fun trivia, and an excuse to bring up Lee May’s name once again.