clock menu more-arrow no yes mobile

Filed under:

Talking Sabermetrics: Defensive Shifts Happen

Do defensive shifts affect the advanced metrics for the Astros?

Justin Edmonds

Topic de jour: Defensive Shifts and the impact on defensive metrics.

Defensive Shifts

Let's start with a candid comment by Pirates' manager Clint Hurdle:

I would think it's been a challenge for the men of my generation. Because this is not the way the game was played. I share this with the players all the time: I played in an era -- this is hard to say but it sounds funny -- I played in an era when you hit a ground ball up the middle it was a base hit (laughter). It was called a diamond-cutter. It's no longer true. Eight ground balls up the middle, five of them the shortstop's sitting right in front of it if there is any type of research done, if you've got any type of major league experience. The game has changed dramatically with the defensive shifts.

Defensive shifts involve the positioning of fielders away from their standard positions based upon the batted ball tendencies of the hitter. Although defensive shifts have been used for 70 or more years, their frequency and scope has increased dramatically in recent years. With a data driven front office, the Astros have leapt to the front of the class for defensive shifts. Compared to 2012, the Astros are on pace to quadruple the team's use of defensive shifts.

The Astros' heavy use of shifts has drawn criticism from some baseball broadcasters and writers, particularly those who are prone to question sabermetrics. Astros' pitcher Lucas Harrell also made public his frustration with the shift, which drew more media questioning of the Astros' use of shifts. The criticisms, though, are not based upon data on the success rate for the shift. In fact, little public data is available to tell us how well the Astros' defensive shifts are working. Who has that kind of data? The Astros' front office. That's why I am fairly comfortable that the shifts have been beneficial. The success of the shifts is purely a data question, and this is an instance where a data driven front office is in the best position--better than broadcasters or even the pitchers on the field-- to evaluate and make adjustments to the defensive shifts.

One of the intriguing questions is how the defensive shift may affect the publicly available advanced defensive metrics. We have only limited knowledge of their effect on UZR and DRS. UZR excludes defensive shift plays from the results--acting as if they didn't exist. One might question how well this "on/off" approach to identifying and excluding shifts addresses the variations and gradations in defensive shifts. DRS historically included shift plays without any adjustments. However, by mid-season 2012, DRS showed a phenomonal level of runs saved for Blue Jays' third baseman Brett Lawrie, who was often used as a "rover" defender in the Blue Jays' shift. In response, DRS began to exclude certain extreme shifts, like those involving Lawrie, from individual DRS results, but included the runs saved in the reported team infield defense. My speculation is that DRS results are more susceptible to the impact of defensive shifts than UZR.

The table below compares the out of zone plays by Astros' infielders in 2013 to the AL average.

Astros 2013
Position Out of Zone Plays Pct of Plays AL Avg. Pct.
1b 9 66 13.6% 15.2%
2b 16 116 13.8% 15.0%
SS 23 98 23.5% 23.1%
3b 24 91 26.4% 20.0%
Total 72 371 19.4% 18.6%

The Astros' first basemen and second basemen are making a smaller percentage of plays outside of their zones than the AL average. The Astros' shortstops and third basemen are making a larger percentage of plays outside of their zones than average. Undoubtedly, the range of the individual players has some effect on these results. But, given that the most common shifts involve pull hitting LHBs, and thereby shift the left side infielders rightward, we would expect the Astros' shortstop and 3d baseman to have a higher proportion of out of zone plays. In this case, the Astros' third basemen are significantly above average in out of zone plays.

An exceptional fielder at third base is a common trait behind an increased use of shifts. Evan Longoria and Brett Lawrie are examples of third basemen who are shifted by the Rays and Blue Jays, allowing those teams to take advantage of their third basemen's fielding talents against LHBs. Ilustrating this point, in a fangraphs' interview, a Tampa Bay sinker ball pitcher realizes that he wanted Longoria in a position to make more than one play per month behind him.

The Astros' defense ranks much better with DRS than UZR. Does this difference reflect the impact of defensive shifts on DRS results? The Astros' 2013 DRS and UZR results, by infield position, are shown below.

2013 Astros Runs Saved
Position UZR DRS
1b -0.06 -3
2b 3.8 5
SS -3.9 1
3b 1.4 8
Total 1.24 11

The DRS results show the Astros' infield with almost 10 more runs saved than UZR. Except for first base, DRS shows more runs saved for every infield position. This is particularly true for shortstop and third base, with 4.9 and 6.6 more runs saved, respectively, for DRS, compared to UZR. I can only engage in guesswork to attribute these differences to the impact of the Astros defensive shifting. But, the positions with the biggest gains from DRS are the positions on the left side of the infield which are shifted closer to the batted ball zones for LHBs. In addition, like most teams with heavy use of shifts, the third base position shows a higher rate of runs saved.

Astros' manager Bo Porter believes that the defensive shifts are responsible for leading the league in turning double plays: "I know for sure that if we're just playing traditional defense where everybody believes the defenders should play, we wouldn't be in the top, as far as the league goes, in double plays," Porter said.

The Houston Press ran a rather bitter and unfair attack on Porter's use of shifts, claiming that his reference to double plays was a "nonsensical stat." It's "nonsense," according to that writer, because the Astros' pitchers put more runners on base, resulting in more double plays. Both DRS and UZR demonstrate that this skepticism is unfounded. Both DRS and UZR contain a GIDP component which is based on double plays per opportunity. As the fangraphs' glossary notes: "There is a separate calculation for GDP’s above or below average, based simply on the number of DP’s turned per DP opportunity, given the speed and location of the ground ball."

According to DRS, the Astros lead the majors in 'rGDP," with 4 runs saved above average by turning the double play. UZR ranks the Astros second in "DPR," with 3.5 runs saved above average due to double plays.

In my view, the Astros' use of defensive shifts will be a continuing adjustment process, with the number crunchers keeping track of individual hitters' response to particular shifts. Statistically, a good case can be made that offensive players should change their approach when faced with a shift. But the creators of the defensive shifts are banking on the fact that many sluggers are not particularly adept at bunting or intentionally hitting to the opposite fielding. Furthermore, some sluggers have a mindset that is averse to "giving in" to shifts. However, some of the hitters will adjust to the shifts by showing that they can bunt for hits or go the other way. For certain exceptionally dangerous hitters, trading singles for extra base hits might be viewed as a defensive success. But in some cases, the defense may have to adjust again.

This is similar to the way that pitchers and hitters are involved in a continuous "cat and mouse" adjustment process. Pitchers pitch to the hitters' cold zones and stay away from hot zones. The hitter may adjust by learning to go the other way on pitches in the cold zone, staying back, etc. This may lead the pitcher to come up with a new game plan to attack the hitter. Sometimes hitters defeat the pitcher even when the pitcher puts the ball in a location which normally produces an out. The same will be true of defensive shifts on occasion.

Rays' manager Joe Madden points out that the defensive shifts are meaningless without pitchers exhibiting good command for locating their pitches. The pitchers have to pitch to the shift in order for it to be effective. This is a learning process for pitchers, and it may take awhile for the pitching-fielding synergy to develop. It behooves the Astros' pitchers to learn how to use the shifts to their advantage.