keyboard_arrow_uptop

Now pitching for the Milwaukee Brewers. Number 21. Jeremy Jeffress.

Check that. Number 13. Will Smith.

Wait… Number 34. Rollie Fingers?

The Brewers may not bring out their closer of yesteryear (or his Hall of Fame mustache) any time soon, but Jeffress and Smith make quite the pair, don’t they? Smith is left-armed and listed at 6-foot-5. Jeffress is right-armed and listed at 6 feet. Jeffress throws his fastball at 95 mph. Smith throws in the low 90s but offsets it with a couple of high 70s breaking pitches. Jeffress is a proverbial “groundball machine” while Smith tends to have a more average profile. But if you venture across the Cheddar Curtain for a three-gamer with the Brew Crew, you might have a game where you get to face both men in two consecutive plate appearances. That might make your head spin so fast that you might think Rollie Fingers was back out on the mound.

Well… does it? Sure, the hitters will notice that Jeffress and Smith pose two completely different challenges, but does the fact that they are so radically different from each other mean that they will have extra difficulties? Does Jeffress’s fastball “look” faster when the last thing that the batter saw was a 79 mph curve out of Smith’s hand? Does the change from right to left (or left to right) confuse the hitters? If it does, does it really have any effect on what happens in the plate appearance?

Warning! Gory Mathematical Details Ahead!

I used data from 2011-2015. I looked at plate appearances in which a batter facing a different pitcher than the one that he saw in his previous at-bat. This means that all of these plate appearances took place against relievers. To make sure that we control properly for the fact that some relievers (and some hitters) are simply better than others, I created a control variable using the log-odds method that takes into account a pitcher’s and batter’s abilities and normalizes our expectations for what to expect in this plate appearance. I also had a variable indicating whether the pitcher and batter were of different or the same handedness.

Then, I looked to see whether big changes from pitcher to pitcher caused changes in the outcome of a plate appearance beyond what those factors might account for. For my first trick, I looked at whether (or not) the batter was facing a pitcher of a different handedness in this plate appearance than he was in the last one. We already know that it is a different pitcher, but maybe the pitcher changing sides that throws him off. That change of sides might be to his platoon advantage or disadvantage, but we’ve already controlled for that.

So were batters freaked out by the fact that there’s a group of people out there who throw with their other arm? No. At no point did that significantly predict anything. Hitters performed as well (no more, no less) as we would have expected based on their own talent, the talent of the pitcher, and any platoon advantage. (I also ran models which included pitch count, which ended up not making a difference for our findings.)

Well, what about different types of pitchers. For example, what if you go from a guy who gets a lot of strikeouts… then again who doesn’t get a lot of strikeouts anymore… to a guy who’s not a K machine? To test this, I took the raw difference in K rate from one pitcher to another. (Smith strikes out 15 percent of the hitters he faced, but Jones has relieved him now and he strikes out 20 percent, so it’s a 5 percent increase.) We generally expect more K’s from Jones, but we’ve already controlled for that. Does the fact that he’s coming after Smith (as opposed to Davis who strikes out 25 percent and would represent a 5 percent decrease)?

Nope.

What about grounders? Is it possible that facing off against a guy who’s a GB machine (at least compared to the last guy) confuses guys? I used the same basic method as above and found that there was an effect, but a very strange one. Facing a pitcher who is more likely than his predecessor to get grounders actually makes it less likely that the batter will hit a grounder than we would otherwise expect out of him. (Rates of singles, which usually parallel grounder rates also go down, as do balls in play in general.) Oddly enough, strikeouts–not usually associated with GB pitchers–go up. It’s as if when a batter is facing a new pitcher who is much more grounder heavy than the last one he saw, he says “I know what they’re trying to get me to do… they want me to hit a grounder… well, I’m not going to give them the satisfaction of giving into The Man!”

Priming the Pink Elephant

For the most part, we don’t see many advantages (or disadvantages) to giving a hitter a “different look” from at-bat to at-bat within a game. Then again, using the same sort of method, I’ve found that there’s no effect for having a similar (or different) starter two days in a row. Players, for the most part, adjust. And I suppose we really shouldn’t be surprised. When this sort of idea is floated, it generally takes the form of “Well, the fastball looks faster, because you aren’t used to seeing it.” But the reality is that even if it does “look faster” it doesn’t make a difference in what we might expect from the batter.

I think we shouldn’t be too surprised by this finding. When watching a game on TV, it’s easy to get caught in the frame of reference of the information that’s being presented to you. You know that Smith came up in the sixth inning against “I’m only in here because my team’s starter did a five-and-fly” guy, but after he does his duty, you don’t think about him. You think about the next hitter. So, when Smith comes back up in the eighth against “set-up guy” you haven’t really thought about him in half an hour, although when he comes up, you are reminded of what he did in the sixth. And since you haven’t been meditating on him, it’s easy to fall into the trap where you assume that he has also had your experience that you are now having in your head and that there’s only been about 10 seconds between those appearances. And maybe he’s gotten stuck on the idea that the pitcher on the mound is a righty.

In reality, he’s sat around waiting for his teammates to do whatever they’re going to do, played the field twice, and maybe consulted with the hitting coach. He’s also seen a lefty before. He knows that they exist and that sometimes they pitch to him. There was a meeting before the game where they went over the relievers who might be throwing today and what to expect from them. And this sort of switcharoo happens a lot, so he’s figured out how to handle it.

The one finding that we did see suggests that when groundball pitchers come into the game to relieve non-worm-burners, they tend not to have an advantage in making their groundball magic work, but in fact actually get fewer GBs than we might expect. The effect size wasn’t very big, but I think it might speak to the idea that the batter might be even a little too aware of the differences between this time and last time and might be over-compensating a little. It’s like when someone tells you not to think about a pink elephant. The first thing that you will think about is a pink elephant.

In all honesty, I’m less interested in the exact finding of this particular research than the topic that it brings up, which is the idea of priming effects. The idea that what came before has a direct effect on what comes next. There probably are priming effects that are real in baseball. Just got buzzed by a 97 mph fastball near your chin? You probably are going to be a little more hesitant to lean out over the plate and a little more susceptible to a ball just off the plate outside. But not all priming effects are created equal. The thing about pitch-to-pitch effects is that they take place within 20-30 second of each other. Pitch sequencing effects might very well be real (and we can study that) but just because one effect is real, doesn’t mean that all sequencing effects are.

Sequencing is one of those fun fields that hasn’t been mined very much. I personally love it because it’s all about tinkering with the behavioral quirks inherent in human beings, but from a functional standpoint, exploiting those weaknesses (or perhaps inoculating your own hitters against those weaknesses) can be worth real value.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
bornyank1
7/19
I'd be interested in seeing some tracking-based stats (release point, maybe spin or perceived velocity) incorporated into this to get closer to the "different looks" effect Jerry Dipoto said he was going for (which Sam described here).
pizzacutter
7/19
Me too, Ben...
jfranco77
7/20
What about the opposite? Of course a lefty isn't throwing you off... dude is throwing with a different hand so you know he's not the same guy. Unless you're Manny Ramirez and literally don't do anything except see ball hit ball. What about 2 different righties? Maybe Thornburg followed by Jeffress is more confusing?
pizzacutter
7/20
That's in there. In the first set of analyses, everyone is facing a new pitcher whom they did not see the last time they were up. The key variable is whether that pitcher is of the same or a different handedness. Doesn't make much difference.