Watch out for shifts: Where defensive alignments could hurt Atlanta Braves hitters

Aside from “pace of play,” frosted tips, take-out slides, and replay review, infield shifts have somehow become another relatively hot-button topic in baseball. In some ways, it’s not hard to see why: baseball keeps having to endure the calumny of “not enough offense,” which is often spliced with “not enough excitement” to produce an executive-level conniption fit. But, hey, don’t take my word for it:

Arguably at the beginning some teams had an advantage because they shifted and others didn't. What I'd argue to you about shifts, everybody's doing it now, it's just changing the game with no competitive advantage, so let's just get rid of them, you know?

—Rob Manfred, from an interview with Scott Miller of Bleacher Report, March 27, 2017

Yes, that was something baseball’s commissioner said very recently, and no, I don’t really get it either.

(I get that Manfred thinks the shift is doing something untoward to the game, but I don’t understand his last sentence, where he suggests that anything capable of being applied or implemented by all major league teams is something to get rid of. Curveballs? Relief pitchers? Sunglasses? Pants? Can you imagine if the MLB Commissioner said, “Look, what I’d argue to you about pants, everybody’s wearing them now, it’s just changing the game with no competitive advantage, and advertisers are mad about the lack of legs on America’s TV screens, so let’s just get rid of them, you know?”)

Anyway, even a casual observer of the modern MLB game can notice that the usage of infield shifts has become widespread. This type of shifting isn’t a newfangled doohickey, of course: usage of the shift goes back to the 1920s, when it was used against Williams. But no, not Ted Williams (and if you think Ted Williams played ball in the 20s...), who rightly terrorized opposing teams enough to warrant special circumstances, but Cy Williams. Of course, it was popularized further when Lou Boudreau used it against Ted Williams in the 1946 World Series, though that was done in part to psych out The Kid (and it kind of worked).

But, whether the shift has origins in the early 20th century or not, it’s hard to dispute that infield shifting exploded in popularity:

Rob Arthur and Neil Pane, FiveThirtyEight, using data from Fangraphs.com

The ball is now in the court of the hitters, to see whether they can adjust to the changing reality of defensive alignments, or whether this type of technocrat-ization of the game is difficult to overcome. For what it’s worth, in a big-picture sense, hitters have not really been adjusting: the total percentage of pulled balls has not changed much since 2010, and if anything, it’s actually increased slightly; the real changes have been improvements in hitter launch angle, leading to fewer weak flies and more liners, as well as more no-doubters leaving the park.

---

If you’re reading this, though, you’re probably aware of much of the above, and you’re also possibly thinking, “Okay, but, what about the Braves?” That’s what the remainder of this post is about — let’s look at the 2017 Braves, their batted ball tendencies, and how often teams shifted on them last year.

The first thing to think about is that infield shifting is really meant to stop groundballs (and some low-angled line drives) from turning into hits, and that hitters tend to pull their groundballs, such that the shift puts additional fielders on the right-hand side of the infield for lefties, or the left-hand side of the infield for righties. There are, of course, player-specific deviations, but the data I am using here are simply those recording whether or not a “traditional shift” was applied to a given plate appearance and ball in play, as recorded on Fangraphs.

So, one rudimentary way to think about the batted balls for the 2017 Braves hitters is simply how often they hit grounders, and how often they pull the ball.

The red dotted lines represent league average, and each quadrant is labeled relative to league average. You can see that, actually, the Braves don’t have any players that both pull the ball and hit grounders at rates above league average. However, they have players that hit lots of grounders without also pulling the ball a lot (like Jace Peterson), and the interesting duo of Matt Kemp and Kurt Suzuki, who both pull the ball at an above-average rate, but hit grounders at a below-average rate.

Of course, there are also the remaining players, like Freddie Freeman, who both don’t pull the ball much, and don’t hit many grounders. Those players, almost definitionally, may be difficult to shift on.

However, despite the fact that I generated the above chart, it’s actually not really the correct way to think about batted ball profiles and shifts. The reason is that major league hitters tend to be special snowflakes that, for a lack of a better turn of phrase, do weird stuff. 2016 Jace Peterson is perhaps a decent example of this.

As you can see, Jace Peterson hits a lot of ground balls. (This is actually not a terrible strategy for him, given that he lacks the power to do damage with fly balls very often.) However, the chart also indicates he pulls just 35 percent of his balls in play, whereas league average for pull percentage is closer to 40 percent. The thing is, though, that delving into his stats shows that Peterson actually pulled about 47 percent of his grounders in 2016, and the reason for his low overall pull percentage is that he pulled a laughably low 13 percent of his fly balls, and a quite low 23 percent of his line drives. (This actually goes back to my first parenthetical in this paragraph. Maybe if Jace Peterson was able to wrap the bat head around balls he elevated and pull them, he’d have more success on fly balls. As is, alas.)

So, the correct thing to do is actually to examine the proportion of pulled groundballs, rather than pulled balls or ground balls separately. That’s exactly what the next chart does.

I took the liberty of also plotting how often teams actually shifted against each player. Once again, the league-average lines are also displayed, separating the hitters into four quadrants.

For Matt Kemp, all alone in the upper-right quadrant, teams seem to be taking a pretty sensible approach: he hits lots of pulled grounders, and teams employ a shift on him a lot. Whether or not teams are infield shifting on Kemp enough is a topic for another day, but in broad strokes, their approach to defending Kemp in the infield makes sense.

Similarly, the bottom-left quadrant features a bunch of hitters that don’t hit many pulled groundballs. Not much point of putting on a traditional infield shift there, as the hitters tend to spray it around if and when they do put the ball on the ground.

The upper-left quadrant is one of the strange ones. Teams love shifting on Freddie Freeman. Only five other players (David Ortiz, Anthony Rizzo, Curtis Granderson, Kyle Seager, and Curtis Granderson) had more infield shifts applied to their PAs. The thing is...

...that’s pretty hard to defend.

At this point, you might be wondering, “But wait! Teams shouldn’t base their infield shift decisions on how likely a player is to put a pulled groundball in play. Instead, they should base infield shift decisions on how likely a groundball from that player is to be pulled.” And, you’d be right — but the goal of this exercise is to examine where teams could really hurt hitters by applying the shift. Teams may be minimizing their exposure on groundballs by shifting against Freeman (though that’s somewhat up for debate, given his spray chart), but they’re probably not harming him that much: despite the high shift burden against him, he still had a monster year, and his 2016 groundball wRC+ (28) was pretty similar to and slightly better than his overall career groundball wRC+ (20).

The bottom-right quadrant is the most interesting. Kurt Suzuki, Adonis Garcia, and Jace Peterson all pulled a lot of ground balls last year. But, opposing teams were not very interested in deploying the infield shift against them. All three of these hitters, and especially Jace Peterson, look fairly vulnerable to increased shifting. Given that these three players averaged a 90 wRC+ last year, and are projected for an average wRC+ of around 85 in 2017, things may go further south for them in a hurry if teams try to further lower the effectiveness of the many pulled groundballs they hit. In particular, both Peterson and Garcia tend to hit a bunch of grounders while lifting balls in the air the other way, making them ripe for shifts plucking hits away; Suzuki tends to hit the ball on the ground less often but is very pull-oriented, which makes him easier to defend overall, though not necessarily as far as grounders go.

Of course, teams may also not really be motivated to shift on Peterson, Garcia, and Suzuki at the plate, given how that they do not necessarily constitute an intimidating lineup presence. Still, if trying to garner every advantage, I would not be surprised to see teams shift against them more often in 2017.

The main takeaways are, essentially:

• Matt Kemp will continue to suffer some detriment to his batting average and on-base percentage due to the presence of infield shifts, unless he can adjust to hit fewer grounders and/or hit more balls up the middle or to right field.
• Teams may be interested in consistently applying the infield shift against Freddie Freeman, but his overall avoidance of both hitting the ball on the ground and consistently hitting it to the right side makes the shift somewhat of a minor factor in impinging on his success.
• Jace Peterson, Adonis Garcia, and Kurt Suzuki should think about adjustments to avoid hitting so many pulled grounders; if teams begin to shift them more heavily, their averages and OBPs may crumble and make them (even more?) problematic plays at the major league level.

---

One quick item - to go back to the earlier point about shift decisions being made in terms of pulled groundballs, I also pulled together data on that.

Not surprisingly, the overlap and conclusions are fairly similar. It makes sense to shift on Matt Kemp — it might make sense to do so on Kurt Suzuki as well, as noted above, though teams haven’t been doing it so far. Anthony Recker is a bit of a strange case: he pulled a lot of his groundballs in part-time action in 2016, but he doesn’t hit very many groundballs to begin with, and he’s not particularly pull-heavy, so that may be a small sample size artifact.

Adonis Garcia and Jace Peterson are below-average in terms of what proportion of their groundballs they pull. That’s a bit of a confounding outcome given the other charts above, but it’s really just that they hit so many groundballs that they still hit more pulled groundballs than other hitters, even if they don’t pull their groundballs at a prodigious rate. Still, the more your batted ball profile is rote and predictable, the easier you are to defend, so I wouldn’t take the fact that they don’t pull all that many of their groundballs as a saving grace for Adonis Garcia and Jace Peterson that they can avoid further damage if teams start to shift on them as well.

---

Last minor note: Freddie Freeman’s production is actually harmed more by those shifts that place an infielder in a “rover” position in very shallow right field. Those don’t tend to be groundballs, but rather line drives, and that sort of shift may not be effectively captured with these data. Still, to the extent that the data yield to teams incrementally better ways of understanding the tendencies hitters have when putting balls in play, hitters will need to keep adjusting to prevent tailor-made defensive alignments from depressing their offensive production. (Or Commissioner Manfred could take a sledgehammer to a non-problem and ban shifts...)