Impact of Banning the Shift

For the 2023 season MLB will be banning the infield shift. This ban is supposed to increase action on the field, which in this case more hits is the goal. This post looks into what a potential impact would be and which players will be most affected.

Methodology

The idea here is very simple, but we need to be careful with interpretation as you’ll see. This came from something I saw from Sisu, who coined impact. What we do is take the average of the variables of interest (in this case batting average), take the average after removing a certain category of a variable (in this case when the infield alignment is in a shift), and find the difference.

What I don’t like about this technique is that many people will think “impact” means “causation” when it actually is an association analysis. The reason why is that we are removing rows of data, which eliminates interactions between variables that may be there and are actually causing changes in output. For example, we are not taking something like quality of contact into account. It could be that the quality of contact has a larger effect on success. I think there is still value in the simplicity of this analysis but CORRELATION DOES NOT MEAN CAUSATION. If you want more causal analysis, I would go with a machine learning model or hypothesis testing. Rant over, now let’s get into it!

Data

What is the shift supposed to take away? Baseballs that are pulled and on the ground. With this in mind, the data taken for this analysis are all pulled ground balls for the 2022 season.

Results

Overall

Note that the following batting averages correspond to just pulled baseballs on the ground in 2022.

Batting Average (All)	Batting Average (No Shift)	Impact of Shift
0.191	0.222	-0.031

So what does this mean? Well on average, the impact of the shift is -0.031 in batting average. This is about 3 less hits per 100 baseballs pulled on the ground. This really doesn’t sound like much in my opinion, but again this is an association and not causal analysis so maybe there is a larger impact than shown here.

Player Level

Now let’s look at which players were the most affected, in terms of impact, by the shift. For this I looked at only players that were shifted on 20+ times in the 2022 season and were also not shifted against (at least once) which resulted in 154 players.

Players Most Negatively Affected By the Shift

Name	Batting Average (All)	Batting Average (No Shift)	Impact of Shift
Joey Gallo	0.071	0.500	-0.429
Joey Votto	0.127	0.500	-0.373
Byron Buxton	0.146	0.500	-0.354
Eddie Rosario	0.081	0.429	-0.348
Mike Moustakes	0.069	0.400	-0.331
Jose Ramirez	0.340	0.650	-0.310
Joc Pederson	0.250	0.500	-0.250
Brad Miller	0.160	0.400	-0.240
Salvador Perez	0.208	0.400	-0.192
Ozzie Albies	0.167	0.353	-0.186

Not surprising that Gallo is at the top but hitters like Votto and Ramirez might be a little surprising. Something to note is that the sample size was less than 10 for the top 5 players for pulled ground balls when not shifted on, meaning that it was rare they weren’t shifted on, which is a confounding variable in this case. Still interesting nonetheless and these players should benefit from the shift being banned. Now let’s look into players that did well beating the shift.

Players Most Positively Affected By the Shift

Name	Batting Average (All)	Batting Average (No Shift)	Impact of Shift
JJ Bleday	0.12	0	0.12
Max Kepler	0.1	0	0.1
Mitch Haniger	0.194	0.1	0.094
Michael Perez	0.091	0	0.091
Michael Massey	0.086	0	0.086
Matt Olson	0.135	0.059	0.076
Mitch Garver	0.2	0.125	0.075
Jake Fraley	0.15	0.091	0.059
Rougned Odor	0.055	0	0.055
Brett Phillips	0.042	0	0.042

As you can see, the top positive effects are much smaller than the top negative effects. This means that they managed to beat the shift but really not much difference. Again, sample size of non-shifted ABs is in here as well. Olson is very interesting to me because in Oakland he pulled the ball very often. An extra fun fact is that Shohei Ohtani was 11th, just missing the top 10 cutoff, meaning he was not affected as much.

Conclusion

The shift being banned should have a high association for more pulled ground balls for certain hitters becoming hits but overall we might not see as much impact as we thought. Time will tell though! For what it is worth, I liked the shift because it is just using data to make more informed decisions to take away hits but many would disagree and that is okay too. I have always been interested in a shift based on the pitcher, rather than a hitter, so I think I will look at that next time.

Data

• Baseball Savant