Which Hurlers Have the Best Pitch Sequences in MLB?
Analyzing How Much Pitchers Change Their Approach in Subsequent Times Through the Order
Last year, I wrote about the most unpredictable pitchers in Major League Baseball. Using an approach called Shannon entropy, I looked at what every qualified pitcher threw during each count and added up how varied pitch types were in every situation. While those with more pitch types in their arsenal obviously tended to be more unpredictable, Arizona’s Merrill Kelly proved even a six-pitch arsenal could still keep hitters guessing.
Perhaps a more interesting question revolves less around individual pitches and more involving sequences a pitcher will throw. Is there an identifiable pattern a hurler will resort to, and more impactfully, does a starter change this pattern when he faces the same hitter a second time thru the order? Or a third time?
I was intrigued by this idea after reading a blog post by data scientists Sean Sullivan and Ajay Patel. There, they explained how to mathematically identify the same elements in subsequences, compare them from one sequence to the next and come up with a Sequence Similarity Ratio (SSR) for how similar and different they are. In our study, we’re analyzing how different sequences of pitch types are from one plate appearance to the next. Here’s an example of how World Series MVP Yoshinobu Yamamoto was able to record a complete game in Game 2 of the 2025 Fall Classic by having different plans for every different plate appearance, keeping one of the more feared hitters on the planet in Vladimir Guerrero Jr. off balance (courtesy: Baseball Savant):
Granted, Guerrero singled in the second plate appearance, but other than consecutive splitters, Yamamoto kept going to different stuff and went with a new pitch ultimately to force a strikeout or weak contact.
Without getting too deep into the math, this approach for measuring the difference between two sequences uses a Sequence Similarity Ratio from the fuzzywuzzyR package in R1. Here, the approach compares two sequences of pitch types and finds how much of the order they have in common versus how much is different. In this study, we treat each plate appearance as a sequence of pitch types and compare how similar those sequences are when a pitcher faces the lineup the first time versus later passes through the order in the same game. Then, those similarity scores are flipped into an Adaptation Index. The higher the index, the greater the change in sequencing from the first time through the order to later plate appearances.
Looking at the 2025 season and only analyzing pitchers who faced hitters a second time through the order at least 100 times, here are the top five hurlers in most unpredictable sequences from one at-bat to another:
Leading the way is former Kansas City turned Colorado starter Michael Lorenzen. Of his seven pitch types recorded on Statcast, Lorenzen does not use any one of them more than 22% of the time (his four-seam fastball). No matter the hitter’s handedness or skillset, Lorenzen seems to have a unique plan for every situation.
Though Lorenzen has left the City of Fountains, they still boast two of the top five hurlers in this metric: Seth Lugo with his ten (TEN!) pitch types and Stephen Kolek whose fastball and offspeed run value both rank in at least the 85th percentile. Look for the Royals’ starting pitching rotation to go deep into ball games and likely exceed expectations.
As for the five pitchers who are the most predictable:
Leading the way are a pair of Cubs’ hurlers: Ben Brown who only throws a four seamer, knuckle curve and occasional changeup, and Shota Imanaga who’s elite not walking hitters but largely goes to his fastball and split finger delivery. As I mentioned on a recent radio interview, there are other ballclubs worth backing to win the N.L. Central, and pitch sequencing is one reason for it.
The importance of pitch sequencing can be shown with analytics and conventional wisdom. Sullivan and Patel’s piece found weak but positive correlations between their version of an Adaptation Index and things like swinging strike percentage and batting average against. They also found pitchers who excel in this area tend to continue being unpredictable in future seasons. Lastly, most any pitching coach will stress the importance of using one pitch (or at-bat) to set up something different to keep hitters guessing. It’s the cerebral part of a pitching matchup, that could very well determine who gains a major advantage.
This dogged reporter can confirm Fuzzy Wuzzy had no hair.