The Embattled Shift
There have always been those who have doubted the merits of the shift in Pittsburgh.
Back in the summer of 2013, after a Pirates’ shift was beat by a well-placed ground ball that resulted in a Texas run, A.J. Burnett exchanged words with shortstop Clint Barmes on the field and later in the dugout. Later approached by a television reporter in the clubhouse, Burnett explained, with emphasis, his frustration.
“Listen, I did not have a problem with Clint! I do not have a problem with Clint! I had a problem with the fucking shift! We play people in the wrong spot!”
Of course, that was also the season the Pirates increased their shift usage by 400%, their staff produced an MLB-record ground-ball rate, opponents hit an anemic .207 against the shift, and the club dramatically improved its defensive efficiency en route to its first winning season and playoff berth since 1992.
Despite the general success of the shift for the Pirates, there were those once again bemoaning it in Pittsburgh (where I still keep a primary residence) — in the media and the public forums — after Sandy Leon beat the shift with a well-placed bunt on Monday.
As one can see, Leon’s bunt is followed by what appears to be a rather unpleasant gaze from Gerrit Cole into the visiting dugout, ostensibly in response to the club’s positioning.
And one batter later, Andrew Benintendi was able to cover an elevated-and-in 98 mph fastball.
Said Cole to reporters afterward: “They just hit them where we weren’t.”
Indeed.
After the game, the results-over-process narrative was the easy one: if the Pirates hadn’t shifted, the day might have turned out quite differently for the road grays. Of course, that’s the nature with the shift: people take it for granted when it works and rage against it when it fails.
The Pirates have invested in the ground-ball-and-shift philosophy as much as any team, increasing their use of shifts in each year since 2012, coupled with the game’s highest ground-ball rate. Though the four-year results against the ground ball are perhaps not as overwhelming as one might expect, the shift has generally worked against grounders. The Pirates’ 2013-15 playoff berths were, in part, credited to ramped-up shift usage.
| Season | At Bats vs. shift | Opp. BA | At Bats vs. no shift | Opp. Ba |
|---|---|---|---|---|
| 2013 | 242 | .207 | 2065 | .219 |
| 2014 | 376 | .234 | 1869 | .213 |
| 2015 | 702 | .204 | 1530 | .218 |
| 2016 | 895 | .226 | 1261 | .236 |
Overall, the industry had dramatically increased shift usage every year since 2012. Shifts were up 58% last season. Major-league teams seem to think they work — or just want to fit in with the crowd.
| Season | Shifts |
|---|---|
| 2012 | 4,576 |
| 2013 | 6,881 |
| 2014 | 13,298 |
| 2015 | 17737 |
| 2016 | 28,072 |
In the 2017 edition of the Bill James Handbook, John Dewan calculated that the game’s 28,704 shifts in 2016 saved 359 runs, and that 1.28 runs were saved for every 100 shifts (down from 1.5 runs in 100).
But shifting, the fate of most trends, isn’t as cool as it used to be, and some are questioning its merits.
Eno Sarris recently pondered whether we are soon to reach peak shift. Wrote Sarris: “If you poke around a little more, there’s evidence that the shift might be providing less value in the past.”
Russell Carleton did a lot of poking around Tuesday at Baseball Prospectus in making a case teams should shift less often. You should read the piece but, in short, it’s complicated. There are unintended consequences.
Wrote Carleton in closing:
“The Shift was sold as a miracle drug. Like most miracle drugs, it does what it says on the label and everyone is happy with it, until they realize what the side effects are. Sometimes the treatment is worse than the disease.”
Typically, when we’ve considered how batters might fight back against shifts, there have been two prevailing thoughts: we’d see more bunt hit attempts down the third-base line and we’d see more ground balls and low liners hit to the opposite field.
Bunt attempts, and bunt hits, to the left side of the infield against the shift are up, as you can see thanks to our handy splits leaderboards:
| Season | Bunt Attempts | Bunt Hits | AVG |
|---|---|---|---|
| 2012 | 80 | 29 | .363 |
| 2013 | 60 | 30 | .500 |
| 2014 | 100 | 50 | .500 |
| 2015 | 595 | 154 | .259 |
| 2016 | 561 | 174 | .310 |
But batted-ball tendencies of ground balls, the types of balls in play against which shifting is most effective, are also holding steady. Most ground balls are hit to the pull side. It’s the nature of the ground ball, of the swing plane, of rolling over on a pitch. It’s difficult, it seems, to purposefully hit a ground ball to the opposite field.
However, line-drive rates to the opposite field are up slightly, though batters’ average on line drives when facing shifts was actually down last season.
What’s also true, and perhaps in part consequence of the shift, is that batters are trying more often to launch the ball over shifts. More batters are trying to produce fly balls. Of course, fly balls hit with authority do not concern themselves with shifts. But are shifts compelling more batters to hit fly balls?
An examination of the data reveals that fly-ball rates have remained steady since 2013.
So it’s a bit of a mixed bag. While we’re seeing more batters bunt in shift situations over the last two seasons, while line drives to to the opposite field are up six percentage points — from 19% in 2010 to 25% the last two seasons — ground balls are still most often hit pull side and hitters are not hitting more fly balls when shifted upon.
Batters have adjusted, but have they adjusted enough to reduce shift usage?
That brings me back to the outrage, locally, over the shift against Leon.
Should the Pirates have been guarding against a two-out bunt with Leon batting?
In 1,276 minor-league plate appearances, he was credited with 16 sacrifice bunts. In 521 major-league plate appearances, Leon has bunted 14 times and has three bunt hits. Though he’s not an extreme pull hitter, more of an argument against vacating third base, he’s hit 155 ground balls and the plurality have been to his pull side.
Said Pirates manager Clint Hurdle after the game:
“It’s a well-placed bunt on his end. We’re well aware of the guys that bunt, and when they bunt, and it was a situation that he hadn’t shown, with two outs and a runner on first. They are an opportunistic team. They got the bunt down, which extended the inning.”
While different analyses suggest different degrees of effectiveness from shifts, most hitters have not shown and obvious ability or willingness to counter-punch the shifts regularly. At least not yet. We’ll perhaps have to see what Leon did Monday more often to say, for certain, shifting should be reduced.
It was something of a surprise attack. The odds suggested Leon was much more likely to hit a ground ball into the shift than to bunt. Teams should play probabilities. And we should reward good process, not whatever results follow.
As for Benintendi’s following swing, there’s no defense for that.
A Cleveland native, FanGraphs writer Travis Sawchik is the author of the New York Times bestselling book, Big Data Baseball. He also contributes to The Athletic Cleveland, and has written for the Pittsburgh Tribune-Review, among other outlets. Follow him on Twitter @Travis_Sawchik.
We can’t put defenders in the right field bleachers, Gerrit.
Not with that attitude….
Tell ’em, Wash.
You can! The rulebook does not prohibit an outfielder playing on the other side of the fence. Only that fielders are in fair territory.
Just a couple of years ago the dominant narrative amongst fans regarding the shift seemed to be the other side of it: the fucking shift was robbing their pull-happy batters of hits they would’ve otherwise gotten. Everyone was calling for more bunting to foil the shift, keep teams honest, and reduce its use. Seems like that’s happened? Baseball: Game of Adjustments, continues to adjust.
Whining doesn’t have to make sense to the whiner. And whining is always a hindsight reaction.
If “my” player hits into the shift, it’s a strategy so good it’s tantamount to cheating, because I want my player to get a hit.
If “my” team gives up a bunt hit to a catcher, it’s a terrible strategy, because I want my team to get outs.
Most (all?) complaints about the shift are hindsight bias, imho.
One can hold both opinions. I think defenses should shift like crazy, like Houston and Tampa do. I also think batters should try to get around it via bunting and uppercut swings.
“They just hit them where we weren’t.”
Very true. Benintendi noticed the lack of defenders in the right-field bullpen. Hurdle needs to fix that.
You’re a good writer. Good addition FG.
This is the exact comment I made fifteen minutes ago in response to Craig Edwards article. Cole was clearly fuming and lost his composure: The bunt Sandy Leon laid down was NOT a 13% hit. It was a perfect bunt that was closer to 90%. Leon has done this before and is an excellent bunter. All bunts may be equal but some bunterss are more equal than others and the real culprit in Cole’s collapse is Hurdle,or whoever decided to apply the shift in this situation. Cole looked into the dugout after that play then came apart at the seems.
“In 521 major-league plate appearances, Leon has bunted 14 times and has three bunt hits”
I think you and I have slightly different definitions for the word “excellent.”
“…the real culprit in Cole’s collapse is Hurdle … Cole looked into the dugout after that play then came apart at the seems.[sic]”
Even if the shift was a mistake, that doesn’t give Cole a pass for losing focus, any more than if it was a booted ground ball or a dropped fly ball. Sometimes your teammates (and managers) make mistakes. Forgetting them and moving on to the next batter is a mental skill pitchers are expected to learn in High School. Cole is 26.
I read the Carleton article and my takeaways were (a) there’s probably a correlation between hitters who hit fly balls and those who pull ground balls (in other words, they’re often the same hitters); (b) pitchers are most likely pitching lower in the zone to try to induce a ground ball when these sorts of hitters are at the plate, whether consciously or subconsciously.
The cause and effect is messy but if pitchers are, in fact, pitching differently when they know a shift is on, then the overall spray charts on a guy don’t really reflect reality. You’d need to isolate shift on / shift off data for each hitter and combine that with pitch locations on batted balls to get a true sense of whether the hitter is performing differently against the shift and, if so, to what extent the pitcher has played a part in that change in performance.
But, what’s the differneces of WRC+, ISO, or slg between the two? Guessing the difference is greater there. Bunts don’t clear basses, and trying to beat the shift by goi g the other way makes it less likely to go yard.
Weird that Cole gets upset by the beaten shift, but not the homer he allowed.
2013-2016 the Pirates ranked 25th in Def. Not sure what that means, if anything, but I thought it was interesting.
Don’t the DEF ratings only take into account non shifted plays? If so that would explain it, they shifted a ton because they didn’t have good defenders
Two thoughts:
1)The data that leads you to shift a player in the first place is based on standard infield alignment, but when you present a batter with a shift, does he approach the at-bat differently? In the past, I don’t think players were willing to take “cheap hits”, like Leon’s bunt or Headley slap single the day before, but as players get more saber-savvy, they are realizing they need to fight fire with fire, and change their approach to an at-bat when shifted. Maybe that’s ok, maybe teams are willing to have an increased likelihood of bunt singles for a lower likelihood of extra base hits, but a change in approach will narrow the statistical gap between shifted and non-shifted BABIP.
2) Considering #1 above, I’d argue teams need to be more thoughtful about when they shift. Leon may not be a great example, since he’s got a limited MLB track record, but for Headley, we’re talking about a career .260 hitter, with power a tic below league average, and not really extreme pull or groundball tendencies, so how much value is there to shifting him? Essentially, what the difference in expected value of an at-bat shifted vs. not-shifted, then weigh that reward versus the risk of a bunt or shift-aided hit. I’d guess for a below league average hitter like Headley, a team wouldn’t gain much by opting to shift him.
I’m not sure how accurate the Shift vs. Non-Shift split stats are on the FG stats page, but they show Leon as one of the Top 3 players in the league vs. the shift on a wRC+ basis last season…..so, maybe not the guy to shift?
Also agree. As shifts become more analyzed and reactive, one constant will be an ability for some players to exploit the shift over a range of AB, accruing baserunners over the normal expectation as well as adding traits for some players who can adjust against the shift better.
There should also be emphasis on the zone map for pitching with runners on base, as some pitchers may be prone to exploit holes or weak contact to induce out vs others who are trying to get the hitter out. The battery scheme changes with the shift, especially when you could have a situation with a pull-heavy player that may not be exploited correctly with their weaknesses if the pitcher is prone to inducing GB% from outside corner of the plate, which lead to 3B plays at SS which also can lead to inefficiencies, errors etc.
We are starting to literally see a shift from pull-heavy players having a one-on-one advantage to a non-shifted field, and some of these players have become mortal due to these shifts, which demonstrates the trend is extremely important from a player perspective. Especially if these players are seeking big contracts, they need to realize these adjustments are of the utmost importance.
The correlation to hit the other way against the shift or bunting hasn’t gone up as much in relation to the number of shifts themselves, showing that development of these adjustments by current MLB veterans are slower and giving many pitchers great advantages. Teams that will be adamant in exploiting these shifts, gaining more practice drills with bunting and coercing their veterans to do the same will counteract these changes quickly, and even just an extra 2 baserunners a week beating these shifts vs the norm can give a team an extra 2-4 wins. Nothing to sniff at, especially when it would also lead to less shifting, allowing their hitters pull without punishment.
As we see few top tier players (that didn’t grow up in the shifting age) doing this, it will most likely occur from the development / scout squads, and sometime soon will see a team where they’ve built most of their lineup to beat the shift. Flyballs are important and I cannot see any statistical change from it, but mid-tier players and players with speed will be the ones most likely to gain from this, in stats and in $$$$ as teams will see this as a clear advantage, and not some controversial topic where players are given freedom to hit into easy outs.
You seem to be under the impression that MLB managers are complete morons.
Of course they’re aware that hitters can adjust; it’s been discussed since the shift was invented decades ago. When a team puts on a shift, they have *already* factored in the batter’s anticipated reaction into that decision, and decided that the odds of this given hitter beating the shift (at least with the pitches they plan to throw) are lower than the odds of him getting a hit on an unshifted infield. They may be wrong in that assessment, but it’s absurd to say that they just took a standard spray chart and shifted based on that, with no thought for what hitters would do when they saw a big gaping hole.
Historically — talking pre-1980s or so here — most batters had enough bat control and bunting skill to take advantage of those holes, which is why the shift was reserved for the most serious power hitters where teams were pretty much willing to happily concede a single. But in the last few decades, as the game has become more power-focused (and not illogically), you’ve had a whole generation of hitters and hitting coaches who haven’t emphasized the “hit ’em where they ain’t” skills … and the shift is the obvious countermove.
Yes, obviously there is a countermove to the countermove, but for some of these guys trying to do that goes against the basic approach they’ve developed for their whole careers. It’s a little bit like a scrambling college QB who gets to the NFL and all of a sudden he has to make more plays from the pocket. It goes against their instincts, and while some can do it, some can’t; and good NFL D coordinators will try (and usually succeed) in taking away his scramble game until he shows he can.
And, to bring this full circle, that’s what MLB defenses are doing to hitters: taking away the thing the hitter likes to do and has spent years doing, unless and until he can force them out of it. Right now, most guys can’t.
Really awesome write-up, Travis!
Chase Headley had three hits on Opening Day beating the Tampa shift: a bunt and two weakly hit ground balls down the line. Tampa’s shift is not even as extreme as most teams. Beating the shift for a left-handed batter does not require anything more than halfway decent contact in the right location.
This is probably analogous to why Maury Wills was able to steal 104 in 1962. Teams were not stealing bases at any significant rate for a long time so teams stopped caring about catcher throwing arms and pitcher pick-off moves. This became very easy to exploit. Once it became obvious that base stealing was back to stay, the teams adjusted their defense accordingly and things calmed down. The shift is the same deal. You have lots of players who simply cannot go the other way or bunt. If they are not sufficiently talented to overcome that, they will be replaced by players who can. The shift will become less common after turning 250 hitters into 300 shift hitters.
If you tried this shifting in the Dead Ball Era, I suspect the team would lose almost every game. It would be nothing but bunts and slap singles until the sun went down.
Is there any data for attempted bunts during shifts that lead to strikes, foul balls included? I am sure if you incorporated this it would deflate the *effectiveness* we see in the stats for bunting against the shift. For example, if any slugger attempts to bunt but instead ends up in a one strike count, the odds of them doing damage goes down. I think this is why a lot of guys refuse to bunt because they simply aren’t good at it.
If I remember correctly, the Cubs shifted very few times last yearin the sense of 3 infielders to one side of second. I seem to remember them shifting slightly based on a multitude of things rather than just employing a full on shift. I would guess that an even more granular approach such as this is probably the next evolution and the full on shift will probably start being used less. Especially since it was so effective, and we all know that the more successful something is, the more it is copied.
I think the shift will remain effective. Slapping the other way would be a win for the defense, in these days most players want to lift the ball and hit over the shift, you can’t defend a Homer.
Pulled fly balls are the most effective hits and anyone wants to become Brian dozier and hit a ton of them. I think the key is to keep the grounders down because those are hurt most, elevate the ball.
Of course there is a slight downside to this too. Guys with fringy power need to pull the ball for power and statistically pulled balls are grounders more often because you hit them more out front which increases the chance for a rolled over grounder. You want to pull it in the air but despite the uppercut swing you will simply have a few more roll overs.
There are a few guys with great bat feel who can selectively pull fly balls but most pull hitters will just take their 15 points babip dip that is caused by the shift and also the fly outs and often also increased pop ups in trade for 25 bombs and rightfully so because it is likely more productive to have a 320 opb with 25 HR than 330 with 12.
The ideal hitter against the shift is a flyball hitter with power to all fields but for that you need the right swing and also true plus power because most guys can pull 360 foot homers if they adjust the swing but for oppo bombs you need real power.
So if dozier decides to hit fly balls to the big part of the park, his pulled fly balls become warning track outs, so he rather accepts that he has a 270 babip but hits 30 bombs which makes him a productive hitter.
Many things come at a cost and now players think about ops and not average anymore. Hitting it low the other way means better babip but no power. And pulling balls in the air means good power but low babip and often higher pop up rates and a fair share of rolled over grounders.
They only approach that has no cost is hitting high line drives to all fields but not anyone can do this while most guys can adjust to either Slapping low or pulling high to Maximize their talents. For guys with 0 to 5 hr power Slapping low the other way is probably most productive while the traditional 10 home run guys can get to 20 with more of an uppercut and more pulling (those are the guys of the “swing revolution”). However that is limited too. Many guys increased their power with a swing adjustment but those guys don’t suddenly become miguel Sano and hit 450 foot oppo bombs,they basically increase their power by hitting wall scrapers to the pull side.
So when recommending going the other way many people don’t realize that comes at a cost, for most this will cause a big drop in power. Babip is affecting batting a lot but it is not the only variable, you must know how much you want to sacrifice for it, especially if you strike out 150 times a year and even a better babip won’t make you a 300 hitter.
For most the choice is not 320 with 8 homers or 240 with 35 homers but 250 with 22 homers or 265 with 13 homers. Is that worth it? Any hitter has to decide that for himself.
An excellent article, inspiring excellent comments.
MLB’s skill requirements are always evolving. When a particular skill wanes, its prior prominence tends to be exaggerated, especially by those who witnessed the decline.
But if place-hitting had really been practiced as widely and well in the dead-ball era as some would believe, why did MLB bat a combined .254 from 1901-19, despite a K rate less than half of today’s?
And if adapting to new conditions is easy, why did so few batting stars of the late dead-ball age become sluggers? Of the 9 guys who totaled .310 BA or better from 1916-19, only Hornsby ever had a 20-HR year. The other 8 guys averaged 7 HRs per 650 PAs from 1920 on, with 61 years as regulars. Meanwhile, 30 different guys had a 20-HR year in the ’20s.
Adapting to and selecting for new conditions is a slow process. The shift as now employed is still young, in evolutionary terms, and the cost-benefit analysis from a batter’s standpoint is complicated. It’s debatable whether the shift’s overall impact is worth radically changing a batter’s approach.
No matter what’s shown by advanced analysis, the rising BAbip of the last 5 years (and the scoring rise of the last two) makes it hard to argue that batters should behave quite differently. To be sure, there are situations where the “free” base offered by a shift is quite valuable, and we naturally bemoan failing to take them. But are there enough of those to drive the engine of evolution towards a skillset better able to exploit those chances, given the trade-offs that may be involved? I’m not so sure.
The game situation Travis discussed here is brutally hard to analyze. Boston already led, but just by one run. We don’t know how big a threat Leon “really” posed, in that moment; his 2016 was way better than his past suggested, but he also has never hit RHPs well. Bunting forfeited a chance to drive in the runner from first base, but it passed the baton to a high-average hitter whose mere single would score him, and it upped the chance of a big inning. Good luck cooking all that into an informed decision by either side.
Babip was not going up the last 5 years. It did go up the last 40 years constantly probably due to harder and better contact but it reached the current level already in the mid 90s and stayed there ever since.
But you are correct that the shift at least did not have a big impact on babip. What er don’t know is whether the shift is not that effective/not affecting enough hitters or whether contact quality continued to get better but is offset by better positioning and defense so that the net result is neutral.