Over the past year and a half, I’ve written about the Twins’ use of defensive shifts on the infield a couple of times.
The summary version of those old posts is that Minnesota had increased its use of defensive shifts on the infield each season since 2016. They had done so to the point where they ranked 7th out of the 30 teams in the percentage of plate appearances with a shifted infield deployed in both 2019 (35.5%) and 2020 (41.3%). Within that, they were one of the leading teams in deploying shifts against right handed batters, ranking 2nd in that split in 2019 (34.9% of PAs) and 6th in 2020 (31.5%).
This past offseason some new public research pieces by Tom Tango and Russell Carleton looked at offensive production against infield shifts and concluded that right handed batters had actually produced at higher levels against infield shifts than they would have against standard defensive alignments. Tango’s study showed that left handed batters suffered a 24 point drop in wOBA against infield shifts while right handed batters gained 38 points of wOBA against infield shifts.
Those conclusions called into question the efficacy of deploying shifts on right handed hitters and led some to estimate that shifting a right handed hitter only makes mathematical sense if that batter pulls upwards of 80% of their ground balls. To put that in context, about 65% of all ground balls have been hit to the batters’ pull sides over the past five years and only about 10% of hitters who played significantly in 2019 and 2020 pulled 80% or more of their ground balls.
Given all that, I wrote this past winter that it would be interesting to watch for a change in the Twins approach in 2021. Now that we have a couple months of data from this season, it’s time to revisit this subject with a few updated charts.
First, the numbers about the overall use of infield shifts:
In the chart above, we can see some evidence that the use of infield shifts — at least as defined by having 3 defenders on the same side of second base — has declined slightly this season. The data shows this is true MLB-wide and for the Twins, although it is interesting that the Twins use has declined about three times more than the MLB average declined.
Now the splits by batter handedness:
The chart above is showing the deployment against left handed batters. We can see pretty clearly in it that the overall decline in shifting we saw from the first chart is not being driven by shifts against lefties. Both the MLB-wide and Twins numbers against left-handers have increased again in 2021. The Twins increase again outpaces the league change by about three times.
So, that must mean the overall decline in shift deployments must have come against right handed batters:
In this one we see the decline in both the MLB-wide and the Twins numbers. At the league level, this is the first decline in infield shifts against right handed batters since 2017. For the Twins, it’s a continuation of a trend that first showed up last year. You can see there on the right side of the chart that the club has roughly halved its use of infield shifts against right handed batters from last year’s figures.
Does this imply the Twins changed strategies? This data would suggest so.
But, it might not be quite that cut and dried. Infield shifts are somewhat messy from a data perspective because the choices about where to position the fielders are much more plentiful than the simple binary choice of shift or standard alignment. There’s a whole range of possibilities between those two ends of the spectrum.
The data shown so far in this article has been limited to the situations when there are three infielders on one side of the second base bag. Statcast captures the various other kinds of defensive alignments as “strategic” shifts.
To know more definitively if the Twins have moved away from shifts against right handed batters, it is instructive to consider their use of strategic shifts, too:
The blue columns are the data against right handed batters and the orange columns are the data against left handed batters. There is a very small uptick in strategic shifts against batters from both sides of the plate in 2021 relative to 2020, but it’s not nearly large enough to make up for the decline in shifts against right handed batters we looked at in the charts above. In fact, with just 7 fewer strategic shifts against right handed batters, the Twins usage in 2021 would be the same as last year’s 7.6%. So, there’s really not much change in this category.
Let’s put all the data together in one last chart that shows Minnesota’s total rates of deploying all types of infield shifts:
Overall, there looks to be a pretty clear shift in the Twins strategy in 2021. They are leaning more than ever into shifting left handed batters while at the same time decreasing their deployments against right handed batters.
It’s hard to know if these changes are an indication the team agrees with the public research findings or if it is just a byproduct of having different personnel at the middle infield positions (or something else entirely). Certainly, having Andrelton Simmons (and his +10 Outs Above Average, 3rd best among all infielders in 2021) at shortstop, instead of the well below average Jorge Polanco, could play a factor in the team determining how often it needed to add a third defender on the right hand pull side of second base.
Whatever the reasons might be, based on what we know to be the general public consensus on the usefulness of right handed infield shifts, the Twins new strategy should probably be viewed as a good thing.
But, is it working?
Since the logic behind deploying infield shifts is mostly about converting more ground balls into outs, one way we can assess how well their approach is working is by looking at opposing batting averages on balls hit on the ground.
We’re past the 60 game mark of this season’s schedule so it’s reasonable to compare stats to last season without much adjustment. The table above also includes the prior seasons for context and shows the Twins right handed opponents batting averages on ground balls by defensive alignment. The 2021 figures in the table above are lower than every other data point in each defensive alignment subset.
Here is the same table, but for left handed batters:
Here again, the 2021 numbers are the best on the table (save for the .192 against shifts number in 2018).
You’ve probably read many times in the past few years that batting averages are declining across the game, due in part to infield shifting but also to the increase in strikeouts. Given those environmental changes, the batting average numbers in the tables above need to be put in context.
For this, I completed one more analysis to index the Twins results with the league average batting averages on ground balls. In the table below, 100 is set to league average and every point above or below 100 is worth 1 percentage point. Since we are talking about pitching and defense, lower is better.
For the first time since 2016, the Twins infield defense is better than average on ground balls against both right handed and left handed batters. Their 92 index value against right handed batters this season is their best performance relative to league average in the time span and their 95 index value against left handed batters is 2nd-best (trailing 2018 by one point).
We’ll see over the rest of the season if these good results hold, but at least in terms of the batted ball type that infield shifts are most designed to stop (ground balls), the Twins overall approach has worked out in 2021.
In a season where so little has gone according to plan, it’s worth noting something that seems like it has.
John is a contributor to Twinkie Town with an emphasis on analytics. He is a lifelong Twins fan and former college pitcher. You can follow him on Twitter @JohnFoley_21.