What the data says about defensive shifts in Major League Baseball
In sports, any proactive team seeks to find advantages within (and sometimes outside of) the rules of the game. Teams like the Oakland Athletics and Boston Red Sox of the early 2000s found success by exploiting league-wide inefficiencies in evaluating players. It wasn’t long before the rest of the league caught up and the era of advanced data analytics was born in baseball.
Recently, we’ve seen a resurgence of defensive shifts in Major League Baseball and increasing numbers of calls to limit or even eliminate shifts.
The modern-day baseball game now has some sort of shift for nearly one out of every three batters, more than double the number of just five years ago. For some teams, the shift has become the majority of what they do as both the Dodgers and the Mets utilized a shift on more than half of their opponents’ plate appearances in 2021.
Baseball has seen many rule changes over the years. Baseball Almanac has a list of such changes. The most celebrated change of recent baseball history was the lowering of the mound following the 1968 season from 15 inches to 10 inches. This was a result of the historically low offensive output of previous seasons. Interestingly enough, that same offseason, the strike zone was returned to the smaller zone that had been in use through 1961. It would seem to me that more balls and fewer strikes would have much more to do with the increase in offense but we’ll never truly know. George Resor did a more in-depth look at this back in 2014 for The Hardball Times.
And so, in 2022, there is a call from some for a change in the rules to limit shifts. A quick Google search will find numerous articles calling for the end of the shift, often citing low offensive output as their reasoning. In recent years, several players have called for some limitations. Joey Gallo was quoted in a recent article by Jayson Stark at the Athletic as saying, “I think at some point, you have to fix the game a little bit. I mean, I don’t understand how I’m supposed to hit a double or triple when I have six guys standing in the outfield.”
Looking for statistical data on the shift is not as easy as I had hoped. Of the publicly available shift data, the two best sources of information I found were at FanGraphs and MLB’s site Baseball Savant. The latter only provides statistical information for the number and percentage of plate appearances (PA) with the shift and the weighted statistic wOBA. FanGraphs has more detailed splits for the shift (both traditional and non-traditional) and will be the primary basis for the information in this article.
The reason a team would utilize a shift against a player would be to turn more balls in play into outs than would happen with a traditional defensive setup. Most commonly, the shift results in having an additional infielder on one side of second base to defend against ground balls hit by extreme pull hitters. At times, this has often resulted in one infielder turning into more of a roving outfielder in shallow left or right field. Other shifts like having four outfielders or a fifth infielder are much less common but still happen often enough in today’s game.
According to Baseball Savant, which has shift data going back to 2016, nearly 31% of PA had shifts in 2021. This number is a slight decrease from 2020 when it was 34% but is significantly increased from 2017 when it was 12.5%. Over the last two seasons, 52% of all PA by a left-handed hitter had some sort of shift while only about 17% of PA by a right-handed hitter did. Every single team in MLB employed a shift for at least 1 out of every 6 ABs over the past two seasons.
So how effective is the shift in Major League Baseball?
The common perception is that the shift works better against left-handed hitters than right-handed hitters and that is likely the reason why it is used three times as often against lefties. Baseball Savant confirms this as, according to that website, righties had a .340 wOBA (weighted on-base percentage) against the shift while lefties have a .320 wOBA in both 2020 and 2021. The league average wOBA in 2020 was .320 and in 2021 was .314. If those numbers are accurate, over the past two seasons lefties have had a slightly higher wOBA against the shift than with no shift and righties have had a significant advantage.
I am a fan of the wOBA statistic as it weights every significant event that goes into a batting line based on the run expectancy of that event. Still though, as good as wOBA is, no one statistic in baseball tells the full story. FanGraphs provides more data across the league and for each individual and has shift data dating back to 2010. At that site, they focus solely on balls that are in play and eliminate home runs, walks, hit by pitches, and strikeouts from a players’ record. The result of this is a record that includes batting average on balls in play (BABIP), hits, doubles, triples, and ground into double plays.
BABIP is a statistic that focuses on the batting average of only balls in the field of play, eliminating HR, BB, K, and HBP from the equation while also accounting for sacrifice flies. Focusing on the data since 2016, the BABIP for PA against a shift has been .296 whereas the BABIP with no shift has been .297. This is over no small sample size either, in those six seasons there have been over 600,000 balls put in play with or without a shift.
Looking even deeper at the numbers, however, there appears to have been a shift, so to speak, in the last two years. From 2016 to 2019 the league had a higher BABIP against the shift (.301) than with no shift (.297), actually providing an advantage to the hitter being shifted against. Over the last two seasons that advantage has flipped and the league BABIP against the shift has lowered to .288 while the BABIP with no shift has remained the same (.297). The league BABIP for this last period was around the midpoint of these two scenarios, .292. For the last two seasons, the shift has taken away a hit about 1 out of every 100 times a player puts a ball in play across the league, or a 1% reduction.
Are extra-base hits harder against the shift in Major League Baseball?
The Gallo quote earlier in the article states that it’s harder to get extra-base hits against the shift than with a traditional defensive alignment. Over the last six seasons, against a shift, 26.3% of all hits (excluding home runs) have been for doubles or triples. With no shift, 25.3% of hits have been for extra bases. Even over the last two seasons, when the tides have turned ever so slightly for BABIP, 26.5% of hits against the shift have been for extra bases while only 23.9% with no shift have had that result. The data states that extra-base hits happened about 10% more often against the shift since 2020 and about 4% since 2016.
There is one baseball event that happens at a significantly reduced rate when playing against the shift, double plays. Over the last two seasons, 3.9% of all balls in play with no shift have resulted in a double play and only 2.3% have had the same result against the shift. When looking at the last six seasons those numbers are slightly closer, 3.4% and 2.5% respectively. That difference may be small but ground balls are about 25% less likely to result in a double play against a shift than with no shift over the past six years and 40% less likely over the past two seasons.
Put another way, based on the BABIP and double play information, over the course of 100 plate appearances against a shift in 2020 and 2021, a team will give up one less hit than they would without the shift. They will also convert one less ground ball into a double play.
How does the shift impact individual Major League Baseball players?
There has been a fair amount of analysis out there about the effect of the shift on individual players. It wouldn’t be surprising to me that some players are impacted more than others but it’s also apparent that because of extremely small sample sizes, the true impact of the shift on individuals is very difficult to measure, at least with publicly available data.
For a player like Gallo, FanGraphs has 0 PA where a ball was put in play with no shift in 2021 and one PA in the last three seasons combined. There simply isn’t enough data on how he fairs with no shift. What we do know about Gallo from 2021 is that 59.7% of his PA resulted in a HR, BB, K, or HBP. Of his 498 AB, 251 were either a HR or K and his average was .151 in those AB. Even Ted Williams in 1941 would hit below .300 if he hit that poorly in over 50% of his AB.
Now when you look at the league data, there are tens of thousands of data points that paint a picture of a slight advantage in reducing hits over the past two seasons by shifting but also a disadvantage in the ability to turn double plays. Across MLB in 2021, 36.3% of PA resulted in a HR, BB, K, or HBP. Compare that to the 29.7% of 20 years ago and it makes you wonder what the real reason is for the decline in offense.
Next. Why banning shifts won't fix anything. dark
Baseball teams are smart and have an unbelievable amount of data at their disposal. There must be significant, data-driven, reasoning put into their increased efforts to utilize shifting defensively. In the future, if some of that data becomes more readily available to the public, I look forward to going through it and seeing what it says. Until then, we have the wOBA data from Baseball Savant that indicates an advantage to the hitter when shifts are in place (at least to righties and no disadvantage to lefties) and the FanGraphs data that essentially says it’s a wash.