What is the Upper Limit of Shifts?

The SABR Analytics Conference took place a couple of weeks ago in Phoenix. I strongly recommend attending this annual conference if you are interested in baseball analytics. During one of the luncheons I was talking to Rob Neyer and Brian Kenny about shifts. The usage of the shift has been increasing over the last few years, but I shared with them how Baseball Info Solutions developed software almost 10 years ago now that suggested a shift on dozens of different batters rather than just the handful that were being shifted then. It took a few years for baseball to catch on, but they did and it’s making a difference.

The software, called BIS-D, currently suggests a full Ted Williams shift (three infielders to the left or right of second base) on well over 100 different batters. Rob asked if it’s possible to use the software to determine an upper limit on the number of shifts. If every batter suggested by the software was shifted, how many shifts would there be? Might that be the upper limit of shifts? A great idea!

The number of shifts in baseball over the last few years have been:

These include both partial shifts and full Ted Williams shifts. If we limit the totals to just the full Ted Williams Shifts, we get:

In 2014, major league teams had a full shift in place 8,354 times when a ball was put in play. How does that compare with what the BIS-D software would have recommended? If you take every player that has put at least 50 groundballs and short line drives in play in their career (and had at least one plate appearance last year), the software recommends a shift on 113 different players. Those 113 players put 24,733 balls in play.

So, the upper limit on full Ted Williams shifts in baseball might be around 25,000, compared to the 8,354 full shifts actually deployed in 2014. Teams are only shifting about a third as often as they could be, according to the BIS-D software.

The Houston Astros, who shifted more than any team last year and saved the most runs defensively (27) with the shift, are close. BIS-D recommended 971 full shifts for them. The Astros deployed 867.

Taking this one final step, teams in aggregate are using a partial shift about half as often as they do a full shift. If that strategy is correct, the upper limit on all shifts, including partial shifts, might be closer to 40,000.