Turn on a baseball broadcast these days and there's a good chance you'll hear about launch angles, catcher framing, pitch tunneling, bullpenning and the like. Look up at the giant video screen or flip over a baseball card and you might see the hitter's WAR. For the most part, the conventional wisdom in baseball isn't much different from what you're reading about on stathead websites.
But there's one area where the conventional wisdom and the data seem to be holding totally different conversations: the shift. Shifting has gone up more than tenfold in the past seven years, and the conventional wisdom has become that the shift is something so effective, it might need to be banned entirely -- like 16-inch gloves or too-tall pitching mounds. "Many in baseball fear that defensive shifts have simply made it too difficult to get hits, even as home runs increased in recent years," wrote the New York Times after commissioner Rob Manfred signaled ongoing efforts to outlaw it. "[Agent Scott] Boras maintained that shifting essentially has broken the game," wrote The Washington Post last month after Boras said shifts are "discriminatory" against left-handed batters.
"When I'm hitting, there should be no shifts, but when we're on defense, we can shift whenever we want," the Astros' Alex Bregman told Sports Illustrated. A joke, but with a confident premise: The shift works.
The twist is that, eight seasons after the shift era really took off, two years after the first shift data became widely public and now a few months after Statcast-based shift data became available, we have a growing body of research into how the shift affects the game and, indeed, whether it even works. And the data confounds the conventional wisdom.
So what do we know about the infield shift?
The data suggests it does do what it's designed to do: It puts fielders where the ball most often goes and turns more ground balls and line drives into outs. There are different ways to measure this, and different questions one must resolve to do so, but this has been a pretty consistent finding:
• Sports Info Solutions, which until recently was the primary source of data on shift frequency, calculates that 28 of 30 teams this year have saved runs on ground balls and short line drives when the shift was on. At the 2018 SABR Analytics Conference, Mark Simon of SIS estimated that the 10 teams with the most full shifts last year saved, on average, 16 runs on ground balls and soft liners. (SIS has found that partial shifts aren't particularly effective at turning grounders into outs, but full shifts are.)
• Mike Petriello of MLB.com used Statcast data this spring to find that among the 200 most-shifted players, batting averages on balls in play dropped 18 points when the shift was on.
• And Russell Carleton of Baseball Prospectus, and the author of "The Shift: The Next Evolution In Baseball Thinking," looked at batters who had a lot of plate appearances without shifts, as well as some with shifts. For those batters, Carleton found a decreased batting average on balls in play of about eight points when the shift was on.
Of course, we'll still often see a weak grounder trickle past the spot where the shortstop used to play, and the camera will inevitably cut to the irritated pitcher who looks unimpressed with these data. Last season, according to Mark Simon and SIS, Anthony Rizzo got 16 hits he wouldn't have otherwise had because of the shift (including a pair of bunt hits), and those hits all make it look like something failed. But he lost 38 hits to the shift. It's easy for humans to fixate on the mistake they see (the 16) instead of the counterfactual they can't (the 38), but we know that this is fallacious and the shift works. Don't we?
In fact, the data suggest that the shift doesn't do what it's supposed to do. It's designed to stop grounders, but it's supposed to help the defense prevent runs. And, in four fantastic articles (1, 2, 3, 4) at Baseball Prospectus this spring, Carleton found good evidence that it's not doing this.
Even though the shift is good at gobbling up ground balls and line drives, it has the secondary effect of making pitchers throw more pitches out of the strike zone. They don't appear to be pitching to the shift -- by throwing more pitches on the inner part of the plate, for instance -- but merely pitching away from contact, nibbling more and throwing fewer fastballs. This all means more balls. More balls mean more walks, and they also mean more hitter's counts, which means more doubles, more triples, more home runs and fewer strikeouts.
"It's a subtle, marginal effect that produces an extra ball every few dozen pitches or so," Carleton wrote. "By focusing on BABIP, we missed the fact that the battle was already lost. The extra ball means that the count could have been 1-2, but instead it was 2-1."
Carleton first suggested this possibility in 2016, but the data was too limited to draw conclusions from. Shift data at the time included only balls in play, so if the defense was shifting when a batter struck out, walked or homered, the shift was treated as if it never happened. But Petriello, using Statcast's data -- which includes all pitches -- found the same effect this spring:
"Now that we have pitch-level data on positioning, we can look at what happens when the ball is not put into play. Here, it's pretty clear. Our group of 201 players walked 9.8 percent of the time against the shift and 9 percent of the time without the shift. Eight-tenths of a percent doesn't sound like much, does it? As it happens, that's almost exactly the jump we've seen overall in the Majors from 2015 to 2017. On that scale, that was almost 1,800 more walks. It's hard to say why, for certain, but it's happening."
We're talking about fairly small effects, but, as Carleton puts it, "they are literally all pointing in the wrong direction other than the singles." And even if it were a complete wash, it would still counter the premise that shifts are stifling offense. It's not the league that should be eliminating shifts, Carleton suggests, but teams themselves:
"The threshold for when a shift should be used should be very high. It should be practiced only on those hitters with the most extreme pull tendencies. There should probably be a few thousand shifts per year, league wide, and they should all be against the same 20 or so guys. And those guys should be bunting against The Shift anyway."
But if the shift doesn't work, then are teams being irrational? Assuming they don't have different data showing different outcomes, why do they do it?
Well, for one thing, not all do, at least not all equally. The Cubs, famously, rank near the bottom of the league in infield shifts, despite being managed by the former shift trailblazer Joe Maddon. The Angels have cut their shifting by more than 30 percent this year, and the Padres and Rangers have also decreased their shifts substantially this season. The wisdom of the shift is not universally conceded. For a few extreme hitters, the shift probably does work. But right now the league's teams are engaged in their own debates about how much it works, how often it works, whether it still works and whether it works specifically for their pitchers and their infielders.
Further, it'd be easy for a team to convince itself that the shift perhaps hasn't worked, but that it will. The data strongly supports the infielders' positioning in theory; the problem is one of management, of making pitchers feel secure in throwing strikes. Teams might be convinced that the central insight of the shift is correct and worth pursuing and that the new problem is one that can be solved through better management or greater familiarity. Maybe they're right. Or, perhaps, a pitcher will always feel vulnerable when one part of the field is entirely unguarded and seems easily exploitable.
Which raises the question of whether industry opponents of the shift are being irrational. If the concern is that the shift has cost the league offense, and offense sells tickets, then the shift is probably the wrong villain. But if the concern is that shift costs the league the right kind of offense -- the kind that involves batters swinging, balls being put in play, fielders chasing baseballs and runners sprinting for bases -- they have a point. The shift has essentially, uh, shifted offense from the ground ball-through-the-hole kind to the waiting-out-a-walk kind. It has given batters less incentive to put the ball in play and more incentive to take pitches. It's plausibly a small part of baseball's pace problem. Further, if we can't say that the shift has stifled offense, we also can't say that it won't in the future as pitchers get comfortable with it or teams get better at managing it.
But for now -- and beyond the basic question of whether it's at all appropriate for the league to legislate basic strategy like this -- baseball's most fallacious assumption isn't that a defensive strategy that apparently boosts offense is actually crushing it. Rather, it's that the shift is somehow making baseball boring, more monotonous.
In fact, it has added variety and mystery. Teams are finding their own strategies, gambling on these strategies, tinkering with these strategies, differentiating themselves from their peers with these strategies, and then challenging hitters to adapt with their own individual strategies. They're doing it all visibly, on the screen, so we have something to talk about -- and, even with mountains of data, something to disagree over.
Rather than vilifying the shift, Manfred should be broadcasting the ambiguities of it. He should be emphasizing each team's unique approach to it. The shift might or might not be working. But it isn't boring.