Why MLB teams are shifting on defense more than ever
The count was 2-and-0 to Mike Trout on April 4 in Houston, and when he looked up, he couldn't believe what he saw: The Astros had shifted their infield to put three infielders on the pull side. "Me? The way I hit?" Trout said. "I was surprised."
Trout's approach to hitting is based on letting the ball travel and hitting the ball up the middle or over the second baseman's head. What did the Astros know about the centerfielder of the Los Angeles Angels to cause such a shift that was so odd it shocked Trout himself?
The answer is, not much. Trout is only 22 years old. Until then, he had put the ball in play on 2-and-0 counts only 34 times in his young career. In those counts, he did not profile as a pull hitter: He hit the ball up the middle 17 times, to the pull field 13 times and to the opposite field four times.
The Astros again overshifted their infielders against Trout in another at-bat when the count went to 3-and-1. The career sample size for that count was a little bigger -- 51 balls in play -- but that spray chart showed even less of an indication of a pull-happy hitter: 30 pitches hit up the middle on 3-and-1 counts, 11 pitches pulled and 10 pitches hit the other way. Still, the Astros, armed with more sophisticated data about Trout, moved their infielders.
"It's working," said Houston general manager Jeff Luhnow, whose team deploys more shifts than any club in baseball. "We feel confident the evidence is very convincing when we analyze how we use our innovative defensive alignments. We believe our data shows we're on the right path."
The Houston pitcher facing Trout was Jerome Williams. Trout flied out to centerfield in both of the at-bats with Houston deploying the count-specific shift. But the results are not so important. What is important is what those seemingly odd shifts represent: Defensive positioning has become one of the biggest changes to how baseball is played since the designated hitter was adopted in 1973. What may seem strange quickly is becoming common.
Baseball likes to hold on to this quaint if incorrect notion that the sport is so timeless that someone from a century ago could sit down to watch a game today and enjoy fairly the same game he saw when night baseball or games west of St. Louis never existed. The adoption -- no, the flat-out overnight embracing -- of defensive shifts blows up such a romantic idea. Anybody from just 10 years ago has to look at the way baseball is played today and find defensive positioning to be wholly different. The takeaway now is this: My goodness, we were playing defenders in the wrong place for 150 years!
It's easy for even a casual fan to notice this revolution, but Baseball Info Solutions has provided the raw data for confirmation. In just three seasons, from 2011 to 2013, teams more than tripled the number of times they used a shift on balls put into play (from 2,358 shifts to 8,134). And the numbers from this April indicate another large leap is happening this season. Many clubs are hiring the equivalent of a "defensive coordinator" to plan and arrange the many versions of the shifts.
The defensive overshift has become standard operating procedure. Why? It's the data, of course. The more data that is collected and analyzed, the more teams are shifting. And the more influence data collectors have in front offices, the more the manager and coaches are simply middle managers making use of the information.
But it's more than just data collecting that is growing. It's also the faith in the data. Trout may have pulled only 13 3-and-1 pitches in his entire major league career, but the Astros trusted that Trout would pull the pitch -- trust that also involved confidence that Williams can execute an inside pitch with precision and that Trout is more likely to pull the ball when he hits it on the ground.
Is such faith in numbers justified? Are defensive shifts really (if you want to play along with a hypothetical parlor game) "saving runs?" That's the key question, of course. The answer is less obvious than you think. Think of defensive shifting as an immature industry: It still is evolving, as is the offensive approach against it.
The shift does seem to work especially well against dead-pull hitters, but not as well overall as you might think. You would think that all this data and all this shifting would make life increasingly difficult for hitters -- that we would be seeing a marked improvement in the rate of how defenses turn batted balls into outs, and a sharp downturn in batting average on balls on play. But we're not.
Teams turn 68.9 percent of batted balls into outs this year. That's down not just from last year (69.2%), but also down from 1984 (69.9%), 1974 (70.2%) and 1964 (70.5%). Likewise, batting average on balls in play this year (.296) is higher than what it was 30 years ago (.286), 40 years ago (.282) and 50 years ago (.279).
It sounds counterintuitive. But shifts are not a fad; teams are convinced they work, and overall, they do. But how can it be that the Yankees, who along with the Astros shift more than any other team, were a more efficient defensive team back in 1990, with a 95-loss outfit under Bucky Dent and Stump Merrill, than they were last year or have been this year?
David Ortiz is a pull hitter. You knew that even without the data. Just watch the Red Sox play a few games and you can see that Ortiz prefers to pull pitches, though he also is apt to hit the ball the other way against some lefthanders and in many RBI situations. But years ago, no one was certain exactly how often and when he pulled the ball. Life was good for Ortiz back when opponents relied more on advance scouting reports and institutional memory than sophisticated data to position their defense.
Then the Tampa Bay Rays hired Joe Maddon.
In 2006, in what was his first trip to Fenway Park in his first month on the job, Maddon broke out a defensive alignment he had kicked around in his head in his years as a coach with the Angels. When Ortiz came up, Maddon stationed six players in the outfield -- four of them, including the third baseman, spread among normal outfield depth, and the shortstop and second baseman positioned on the grass in short rightfield, behind the infield cutout. Maddon would admit privately there was some gimmickry to the alignment; Ortiz was wearing out the American League, and the manager figured the trickery might play with his head. But it also was based on Ortiz's spray charts, which showed that he rarely hit a groundball to the left side.
Life at the plate would never be the same for Ortiz. Maddon's defensive strategies first drew chuckles, but when Tampa Bay won the 2008 pennant, Maddon and the Rays suddenly looked more like geniuses than tricksters. From 2007 to 2008, the Rays improved by 31 wins, and the third year of their defensive positioning system -- they were using data-based shifts routinely -- deserved much of the credit.
How much did Tampa Bay improve on defense? The answer could be found in Defensive Efficiency, which measures the percentage of balls in play that a team turns into outs. From 2007 to 2008, the Rays improved from last in the league in Defensive Efficiency (.652) to first (.708), an extraordinary leap.
The rest of baseball noticed. The shift, especially against obvious pull hitters such as Ortiz, became more common and more extreme. Look at how deeply the emergence of shifts against Ortiz changed his career:
|David Ortiz Batting Average When Pulling the Ball|
"The first time I really noticed it was with Joe Maddon," said Ortiz, who entered this week with 2,045 career hits. "I'd probably have like 2,600 or 2,700 hits without the shift. It's just something that I learned not to worry about. It's just common now. Some teams, they play me so deep with the second baseman, there was a game against Baltimore where I hit a flyball and the second baseman caught it two steps from the warning track."
Ortiz is far from alone. Shifts are deployed much more often against lefthanded hitters than righthanded hitters, especially the prototypical sluggers like Ortiz. And in those cases, the shifts appear to be working very well.
Check out this chart, which shows the year-by-year batting average for all MLB lefthanded hitters when they hit the ball to the opposite field and when they pull it:
|Batting Average by MLB Lefthanded Hitters|
While the batting average to the opposite field for lefties has gone down slightly (overall batting averages have been trending down), it absolutely has cratered when lefties pull the ball. Shifts appear to be working very well against lefthanded pull hitters.
"That makes sense," Luhnow said. "That's the group with the highest concentration of shifts used against them."
The Houston Astros and New York Yankees have something in common, and it's not their payrolls, the age of their rosters or their chances of playing postseason baseball. The two otherwise very different clubs share some of the strongest beliefs in the shift. Unlike the data against lefthanded pull hitters, the overall data here provides mixed results.
The Astros seem to be improving incrementally as they shift more; the Yankees, not so much. Check out their four-year trends:
|Astros/Yankees Defensive Efficiency/BABIP|
The Astros still aren't as efficient as they were a decade ago, but you can see improvement in recent numbers. The Yankees still don't measure up to the awful Yankees of 1990, a team that featured the likes of Steve Sax, Alvaro Espinoza, Jim Leyritz and Oscar Azocar. That team's Defensive Efficiency was .699, with a .288 BABIP against.
How could that be? It must be a statistical quirk, right? I'm sure if we looked at overall MLB Defensive Efficiency and BABIP over the years they will tell us that today's sophisticated defenses are turning more batted balls into outs and making it harder for hits to fall -- right?
Uh, no. Take a look at snapshots in 10-year intervals:
|MLB BABIP/Defensive Efficiency|
I've asked a few baseball people to explain this counterintuitive trend and gotten everything from blank stares ("Run that by me again," said Orioles manager Buck Showalter, a major believer in shifts) to blaming the root of all evil in the modern game, strikeouts (Cubs president Theo Epstein said fewer hitters use the old-school "just make contact" two-strike approach, turning what used to be easy groundball outs into strikeouts). Luhnow said it may be best to ignore such comparisons because "Defensive Efficiency includes outfield defense," the data contains so much "noise" and can be "dirty," and besides, he said, he is not interested in "the global trends," only in how his team is improving according to its proprietary metrics.
"It's like macroeconomics and microeconomics," he said. "You might expect the data to match up and show the same trendline, but that's not always the case. One of the things you have to remember is that we're only talking about 10 percent of the plays. For 10 percent of the plays to impact the 100 percent is not something that's going to show up clearly. The average team shifts four times per game."
There is no going back. Shifts are here to stay. They have become so widely accepted as "state-of-the-art" baseball and "smart" baseball that non-believers are dismissed as Luddites of the game. As Indians manager Terry Francona told MLB.com, capturing a kind of peer pressure, "There's so much good information now that, if we don't take advantage of it, then I think you're looking at an older manager stuck in their ways. That's not good."
Older managers don't get hired any more, and nobody wants to be considered uncool. Even Mike Scioscia, in his 15th season managing the Angels, deploys more shifts, though he prefers to call it "shading" and is less autocratic about it than other clubs.
"We're doing a better job shading than we ever have," he said. "But I don't think you will ever replace the instincts an infielder has to have to move on a pitch and cover the ground he needs to cover."
Everybody is on board to some degree, even if defensive shifting is not as simple as it appears. It changes relays, cutoffs, double plays and the angles of throws. It threatens to leave third base uncovered. It sometimes places fielders in awkward, unfamiliar positions (i.e., Adrian Beltre turning a double play at second base, Derek Jeter playing third base). It takes leaps of faith, such as when the Yankees overshifted on rookie Jackie Bradley Jr. (career balls in play by direction: 36 pulled, 53 up the middle, 24 to the opposite field) and when they got burned by one against Mike Carp against lefty Matt Thornton (Carp has faced lefties in only 19 percent of his career plate appearances), or when the Red Sox overshifted on Ryan Flaherty, the Baltimore infielder with a .359 career slugging percentage. Flaherty poked a single to leftfield.
"Now, see, we learned something there," said Red Sox coach Brian Butterfield. "He saw the defense and tried to go the other way. We take note of that. Some guys will try to counter the shift. He's one of those guys. The next time we won't overshift as much. You learn who those guys are."
Brian McCann of the Yankees is another "one of those guys" who has tailored his swing to occasionally take advantage of what the defense gives him.
"It's the next phase," Yankees hitting coach Kevin Long said. "Hitters are adjusting. It was one thing when you just had Tampa Bay and maybe a few other teams shifting on you. But now it's 30 teams. Everybody is doing it. You see it all the time. So when you see it all the time it's time to start thinking about making adjustments. Brian had two hits going the other way against Tampa. He's a good hitter. Not all hitters can do that, but the good hitters can do it."
McCann has four hits to the opposite field; he had 16 all of last year. Opponents might tell you they are fine with that occasional choice by McCann. Why? He slugs .786 when he pulls the ball but only .392 when he goes the other way.
Part of the "next phase" undoubtedly also will include more bunting against the shift. On Saturday, for instance, Ian Stewart of the Angels dropped a bunt against a Yankees shift for a hit in the eighth inning of a one-run game, giving Los Angeles the potential tying run at second base and the potential go-ahead run at first base. More hitters will practice bunting against the shift -- the ball should be pushed, not just "dropped" as in a sacrifice situation -- to at least make it a counter-option.
"Any time there's a change in strategy, there is a counter to it," Luhnow said. "We know that. It's not like the spots our infielders are standing in now are going to be the same spots over the next 20 years. It's a constant battle of adjustments."
People may argue how much the shift works, but that it does work and that it is here to stay should no longer be an argument.
These are bad days for hitters. It is harder to get a hit in the major leagues today than at any time since 1972 -- the last of the seasons without the DH. It would be easy to ascribe much of that difficulty to the wonders of number-crunching and the sophisticated defensive positioning it inspires.
Actually, the downturn in hitting has much more to do with pitchers than it does fielders. In 1972, for instance, hitters batted .244 overall, including .272 on balls in play. This year, hitters are batting .248, but that average now includes a much more robust .296 average on balls in play. Here's the difference between 1972 and 2014: The percentage of plate appearances that are strikeouts has soared from 14.8 to a record 20.8.
A key part of the "problem," if you prefer more offense, is another post-DH change in the game that is far more impactful than the shifts: The modern specialized bullpen. In 1972, hitters actually hit for a higher average (.250) and struck out less often (14.3%) in innings seven through nine than they did in the first six innings. This year, hitting gets even more difficult in the late innings (.238) with even more strikeouts (21.4 percent) as managers use a parade of hard-throwing relief specialists to quash offense.
Put another way, just when a baseball game should be building to an exciting climax -- the last three innings -- the modern game devolves into 1968, the infamous Year of the Pitcher, when hitters batted .237. It is the only full season since 1871 in which hitting was harder than it is in the last three innings of today's game.
So if you crave more action, don't place too much of the blame on the proliferation of the shifts. Sure, they are designed to take away hits, but the modern run-starved environment is less about what happens when the ball is put into play and more about how often it is not.