Painting the Strike Zone: A Blurry Beginning

A while back, I investigated how a varied pitch mix helps produce weak contact. At the time, I wondered if varying pitch location would be a benefit for the pitcher. After a first stab at the data, the answer is somewhere between no and just not known.

The theory goes that a hitter would have a tough time squaring up a ball as it gets located in different parts of the strike zone. The results could even be more swing-and-miss. With this focus, I just dove in to see what stuck.

The first hurdle was finding a way to measure pitch location variation. I ended up using nine zones with nearly the same number of pitches in each zone. Deciding on just nine zone drives the rest of the results. Should there be more? Fewer? Some removed? If/when I reexamine the data, I’ll start here with some adjustments.

With the nine zone zones, I used the inverse of the Herfindahl-Hirschman Index (full explanation at this article’s end) to show diversification in the nine zones. A value of 1 means all the pitcher’s pitches were in one zone while a value of 9 means they were evenly distributed between all nine zones. Most pitchers had a value around ‘7’.

I determined I needed to limit my player pool to just starters. Relievers throw fewer pitchers therefore their pitch location variation is lower because they have fewer opportunities to mix things up. Also, relievers have a higher strikeout rate (9.7 K/9 vs 9.1 K/9 in 2021), so all the low variance pitchers had high strikeout rates. To try to even out the talent pool, I settled on starters with a minimum of 80 innings from the 2017 to 2019 seasons.

And after trying every possible input, I found nothing showing any relationship between any plate discipline or power allowed metric and the diversity value. Zip.

I had a backup plan. Instead of looking at all pitches, I limited it to just fastballs. Using the same group of starters, I got no correlation. When I say no correlation, nothing over an r-squared of 0.02. Here is that 0.02 correlation for strikeout rate.

And the correlation isn’t even in the expected direction with a lower diversity leading to more strikeouts. This should not be a surprise with high fastballs leading to whiffs and pitchers attacking that section of the strike zone. The starters with a diversity factor under 7 threw 49% of their fastballs in the top half while the rest were at 34%. Possibly diversity isn’t as important as the location’s overall effectiveness.

It’s back to the drawing board. I’ll have to think about the lack of results to see if I can find any useful information. One possible variation is to go with more or fewer zones. I like the idea of fewer since each fastball will have its ideal plane (four-seam: high, sinker: low, cutter: middle/low), and maybe the answer is variation in each plane. Maybe.

I’d also love some help from the crowd. Maybe some will notice a trend. Here are the diversity leaders and laggards from this and last season for all pitches and just fastballs.

Pitch Location Diversity Leaders & Lo
2020 All Pitches 2020 Fastball 2021 All Pitches 2021 Fastball
Name Diversity Name Diversity Name Diversity Name Diversity
Luke Weaver 8.9 Adam Wainwright 8.5 Chase Anderson 8.6 Jose Urena 8.6
Tyler Anderson 8.8 Anibal Sanchez 8.5 Jordan Lyles 8.6 Trevor Cahill 8.4
Kyle Hendricks 8.8 Jon Lester 8.4 Brandon Woodruff 8.6 Jack Flaherty 8.3
Chris Paddack 8.8 Kyle Gibson 8.3 Hyun-Jin Ryu 8.5 J.A. Happ 8.1
Chase Anderson 8.7 Patrick Corbin 8.3 Marco Gonzales 8.5 Drew Smyly 8.1
Tanner Roark 8.7 Dinelson Lamet 8.3 Sean Manaea 8.5 Zack Wheeler 8.0
Hyun-Jin Ryu 8.7 Gio Gonzalez 8.3 Tyler Anderson 8.5 Adam Wainwright 8.0
J.A. Happ 8.6 Clayton Kershaw 8.3 Danny Duffy 8.5 Zach Eflin 8.0
Austin Voth 8.6 Austin Voth 8.2 Tyler Mahle 8.4 Brandon Woodruff 8.0
Yusei Kikuchi 8.5 Kenta Maeda 8.2 Zach Eflin 8.4 Charlie Morton 8.0
Gerrit Cole 6.9 Dylan Cease 5.9 Shane Bieber 7.0 Chase Anderson 6.0
Jack Flaherty 6.9 Sandy Alcantara 5.9 Clayton Kershaw 6.9 Johnny Cueto 5.9
Derek Holland 6.9 Gerrit Cole 5.9 Rich Hill 6.9 Gerrit Cole 5.8
Garrett Richards 6.9 Dylan Bundy 5.9 Justus Sheffield 6.7 Trevor Bauer 5.7
Daniel Hudson 6.8 Tyler Mahle 5.8 Joe Musgrove 6.5 Dylan Bundy 5.7
Shane Bieber 6.8 Taijuan Walker 5.8 Jacob deGrom 6.5 Carlos Rodon 5.6
Zach Davies 6.8 Nathan Eovaldi 5.7 Matt Moore 6.5 Matt Moore 5.3
Michael Pineda 6.6 Rich Hill 5.6 Garrett Richards 6.4 Jacob deGrom 5.2
Justin Verlander 6.1 Blake Snell 5.1 Josh Tomlin 6.3 Andrew Heaney 5.2
Drew Pomeranz 5.6 Drew Pomeranz 4.8 Trevor Bauer 6.3 Mike Minor 4.9

More elite pitchers are in the non-diverse rankings, especially for the 2021 fastball group. Stuff dominates performance.

I found no measurable advantage in plate discipline or limiting hard contact when moving a pitch around the strike zone. If anything, it might be ideal to attack one location with fastballs and another spot with his breakers. For now, I have another dataset to use for pitcher evaluation and maybe someday I’ll figure out its usefulness.





Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.

6 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Nathanmember
2 years ago

Did you control for how undesirable the middle of the zone is and that more diversity could mean missing middle middle more?

Nathanmember
2 years ago
Reply to  Jeff Zimmerman

I wonder how correlated the variety you found was with distance to center of zone. If it is, can you correct for that to hone in on the effect of variety independent of just bad positioning.