The effect of pitch sequence on batting eye and selectivityby Craig Glaser
March 29, 2010
Pitch sequences have always been extremely interesting to me. Throwing pitches in an intelligent sequence is one tool which a pitcher has in order to gain a significant advantage over the batter. When I first started writing about batting eye this was one of the topics I thought would be interesting to look into. In fact one of the first comments I got was from Russell 'Pizza Cutter' Carleton asking when I was going to write about pitch sequences. I think that batting eye and selectivity are new tools which we can use to see what effect pitch sequence has on the batter's ability to perceive the second pitch and their bias on whether to swing or not.
Before we start let's lay out a few things. We're going to look at two-strike counts (for reasons I've explained before) though only the second pitch in the sequence has to be with two strikes. As with the previous articles we're going to stick with the four most common pitch types - change-ups, curveballs, four-seam fastballs and sliders. We're looking at pitch-by-pitch data from the 2009 season (though I have also run the 2008 data and gotten very consistent results). I'm sticking to 2009 because I think that the pitch-fx pitch type classifications are much better for 2009.
Let's start out by looking at change-ups. We saw in a previous article that change-ups are generally the easiest pitch to judge, with batters having the highest batting-eye scores, especially early in the count. I have highlighted the sequence with the highest batting eye score (best for the batter) in red and the lowest batting eye score (best for the pitcher) in blue.
The first thing that stands out here is throwing two change-ups in a row is probably not a great idea. Batters' ability to perceive the change-up seems to improve significantly when the previous pitch was also a change-up. They are still quite biased on this sequence but their selectivity is also quite a bit higher for the second straight change-up. Interestingly the fastball is not the pitch that minimizes batting eye on change-ups. The curveball seems to make the change-up both harder to judge and leads to the biggest bias toward swinging. If you're trying to make a batter swing at a bad pitch you would be well off to throw a curveball followed by a change-up.
Next let's take a look at sliders. Overall the slider was the second easiest pitch for batters to judge (though it became much more difficult on full counts). How does the previous pitch type affect slider perception?
Once again we see that it's probably not a good idea to throw two sliders in a row. Batting eye and selectivity both increase significantly when the pitcher throws two sliders in a row. There are not huge differences here among the other three pitches but it does seem like you get a slight advantage when the pitch before a slider is a fastball.
The curveball is generally the hardest pitch for batters to judge. It's similar to the slider in that batters have more trouble picking it up on full counts (when they don't expect the pitch) but is harder to judge overall. Let's see what pitch sequence shows us here.
Unsurprisingly we see the same effect that we have seen in change-ups and sliders. Repeating a curveball makes it easier to perceive and makes the batter least biased. Interestingly, the curveball seems to be easier to perceive when following a slider as well. It seems like there's actually quite a large benefit to the fastball-curveball combination. The batting eye score of 1.20 for this sequence is the lowest which we'll see for any sequence of pitches.
I've saved the fastball for last and I think you'll see why once we take a look at the chart. The fastball is the most common type of pitch thrown. Count didn't really have that big of an effect on perception of fastballs though they did get a bump on full counts (because they are thought to be the easiest pitch to control and therefore batters expect them).
Here we see the complete opposite of the previous three types of pitches, the fastball actually is most difficult to perceive when it follows another fastball. Once again, change-up to fastball is not a very good combination for fooling batters. There doesn't seem to be much difference in selectivity on fastballs based when it comes to pitch sequence. Slider-Fastball seems to be the combination that enables batters to see the fastball best.
In this article we've looked at the effect of pitch sequence on perception of pitches. We've seen that outside of fastballs it's probably not beneficial to repeat pitches. We've also seen that, at least in terms of perception, fastball-change-up isn't as beneficial of a combination as I expected. Maybe the effect will be seen in outcomes rather than swing behavior. After all, perception is just a part of what determines the effectiveness of pitch sequences. While a low batting-eye score is in the pitcher's favor (and should enable them to get either more strikes looking, more strikes swinging at bad pitches or both) clearly the amount and quality of contact is extremely important as well.
I've got more work to do on the topic of pitch sequences. I want to break the sequences down further (incorporating number of balls as well). I also want to examine outcomes and see how they relate to swing behavior. I think that looking at individual players will be interesting as well (though limited sample sizes will make it a bit harder to draw any conclusions on that level).
Read more of Craig's work on his website Sabometrics