Keegan Thompson Made Sweeping Changes

Keegan Thompson
Kamil Krzaczynski-USA TODAY Sports

Pitches are hard to classify, and the interactions between them are even more fraught. Mix up a slider and a cutter, and you’ll make the aggregates look all — wait, I started too specific. Let’s back up. Baseball is a game played between two teams, each of whom take turns trying to hit a ball thrown by the other team using a bat. The object of the game is to advance safely — wait, now this is too general. Let’s see if we can hit a happy medium.

Keegan Thompson throws a lot of pitches. Depending on how you count them — and boy, will we get into how you count them in this article, don’t worry — he throws as many as six different pitches, all with distinct properties in one way or another. Need a grounder? Keegan has a pitch for that. Strikeout? Sure thing. Want to beat a lefty, or induce a pop up, or regulate local power plants? Thompson can do the first two of those, if he picks the right pitch for the right situation.

This year, it looks like that previous paragraph is only partly hyperbole. In his 2021 major league debut, Thompson mainly worked out of the bullpen and had mixed success. He struck out an average number of batters but also ran a walk rate of nearly 13%. He went into 2022 without a clear role on the team — we projected him for a few starts and plenty of bullpen work — but with a pressing need for improvement on the command front.

In the bullpen, Thompson largely relied on his four-seam fastball, cutter, and curveball. He mixed in an extremely occasional changeup — 12 of the 620 pitches he threw in relief — but for the most part, he stuck with his bread and butter. That makes sense. They can’t all be his best pitch, and if you’re only throwing one inning per game, you might as well stick with the bread and butter — in his case, the high cheese.

But tinkerers want to tinker, and in his new starting role, Thompson explored the studio space. At first, he did it in fits and starts. He stuck to his fastball/cutter/curveball strengths, mixing in a handful of changeups and sinkers in every start. Then, in his last three starts, he started throwing a slider once in a while. The slider doesn’t grade out particularly well by Pitch Info, and he doesn’t throw many of them, but fortuitously, his cutter and four-seamer got much better, and just like that, Thompson was an excellent starter.

Or, well, that’s how it goes if you look at various pitch classification sites. The reality isn’t quite so simple. The fact that he’s throwing a new slider is certainly true; MLB.com’s Jordan Bastian wrote an excellent story on the pitch this week. Thompson and the Cubs wanted a new pitch to finish at-bats, and it’s hard to miss the impact that sweeping sliders have had on the game. He learned the pitch during bullpens — side note, wow is that impressive — and started throwing it three starts ago, on June 17.

So far, so good! We’ve covered all that already. Bastian’s article also lays out some of the results: iffy location, but excellent bat-missing ability. Thompson mentioned the need to control the pitch better — per Pitch Info, it has a zone rate of 12.5%, so he’s not wrong — but praised its movement and sounded excited about using it in the future.

Now let’s get to the confusing part. In those same three games, Thompson’s cutter improved markedly. It started missing more bats. Contact against it got worse. The same thing happened to his four-seamer, at least somewhat; the contact data is mixed, but it’s certainly missing more bats.

We can leave the four-seamer aside for now. As best as I can tell, that comes down to better location. Take a look at this time lapse of four-seam location before his recent good stretch and during it:

Okay then! Higher four-seamers, profit. That’s easy enough to understand. And he’s locating his cutter closer to the edges as well — probably not enough to explain the extent of its improvement, but still a tailwind. Still, Thompson added a new pitch, and his old ones suddenly got better. Is there something to it?

Let’s try another theory. Is Pitch Info (which drives those location charts I just used) mis-classifying his new sweeping sliders as cutters some of the time, leading to a mis-match in what we’re measuring? As it turns out, basically no. One of the neat and under-appreciated tools on FanGraphs is the pitch graphs tab, where you can break out velocity, movement, and release point by pitch type. With perhaps one or two exceptions, the slider and cutter appear to be distinct and correctly classified pitches:

And, well, if that’s the case, then the cutter is more or less the same before and after Thompson added a sweeper:

Change in Cutter Shape? Not Really
Metric Pre-6/16 Post-6/16
Velo (mph) 90.5 90.37
HMov (in) 2.94 2.76
VMov (in) – Grav 4.44 4.23
VMov (in) – Total -21.17 -21.57

In fact, if you strip out the two potentially mis-classified cutters, there’s been no change in velocity and a negligible change in vertical movement. The only difference is slightly less horizontal movement, and that doesn’t sound like a great way to explain how the pitch got better.

I have a guess, but I want to be clear that it’s just a guess. Basically, I think that adding a slider has freed up Thompson’s other pitches to be used in situations where they shine. Consider: before he added his slider, he threw either a cutter or curveball 50% of the time in 0–2, 1–2, and 2–2 counts, or, in other words, counts where he was hunting a strikeout and could afford to waste a ball. In his last three starts, that number has plummeted to 29% per Statcast, and it’s more like 23% when you account for classification issues.

That’s not a huge difference in the aggregate. Thompson has thrown 16 (or more, Statcast’s classification isn’t as clean as Pitch Info’s here) such two-strike sliders in these most recent starts; we’re not talking about an overwhelming number of pitches here at all. But I don’t find those counts conducive to cutters, at least not the kind of cutters Thompson throws. As assistant pitching coach Daniel Moskos told Bastian, “if you’re looking for swing-and-miss… to righties, you need something that moves a little more horizontally.”

Sure, that’s in response to a question about the new sweeper, so horizontal break was surely on his mind already, but if you’re specifically looking to miss bats, more movement is better. As an added benefit, having the slider as an option probably helps the cutter play up. Both pitches feature a ton of gyroscopic spin, which helps to disguise them out of his hand, and they’re only five or so miles per hour apart. Despite that, the slider drops nearly a foot more on its path home — thanks, gravity — and moves four or so more inches horizontally. When you’re seeing cutter after cutter throughout the day, a similarly-spun pitch that dives down and away can’t be easy to adapt to, and knowing the next pitch could be a sweeper does you no favors in hitting a cutter either.

Thompson hasn’t thrown enough sweepers for me to feel comfortable measuring this effect. I’m honestly not sure I could pick it up even if he’d thrown 10 times as many; deciphering the interaction between pitches is one of the toughest analytical challenges in the game. But I certainly don’t think it hurts. And even if that isn’t the case, using his cutter in better-suited counts can explain a lot of the improvement on its own.

If you give up contact in a two-strike count, you’re forgoing a chance to strike someone out. No matter how good you are at suppressing contact, letting a two-strike count pass by without a strikeout is a bad deal. Like the conundrum that sinker-heavy pitchers face, Thompson was suffering a strikeout shortage. He couldn’t keep his four-seamer high, and he didn’t miss enough bats with his cutter to make up for it.

That shortage has turned into a surplus in his last three starts; he’s struck out exactly one-third of the batters he’s faced. I don’t expect that rate to continue — his peripherals don’t support it, with a swinging strike rate of only 13.6% — but I do expect him to post above-average strikeout numbers the rest of the year.

Perhaps most exciting? Thompson has accomplished all of this without his sweeper being particularly good. He hasn’t refined its control at all; he’s thrown 24 of them, and 14 have been in the “waste” zone. Some of those have worked, but you can’t always face the Reds:

If he starts to hone his command of the pitch, better days might be ahead. Not better than his last three starts — I don’t think he’s a true-talent 2.45 ERA pitcher — but better than his mixed 2021, and better than the long reliever role he started the season with. As an added bonus, adding the slider makes the graph of Thompson’s pitch movement quite pretty:

For Thompson and the Cubs, 2022 is a season of discovery. There won’t be playoff games at the end of the year. There might be empty seats in September, and there will almost certainly be trades with an eye toward the future at the end of July. But finding something new is a perfectly good reason to play baseball, and with his new sweeping slider and suddenly improved command, Thompson looks to be on the right track.





Ben is a writer at FanGraphs. He can be found on Twitter @_Ben_Clemens.

4 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
lee170
1 month ago

Very nice article! It really shows how a new pitch can impact a pitcher’s repetoire. What is the base line (0 point on the charts) for vertical and horizontal movement? Point of release?

phoenix2042member
1 month ago
Reply to  Ben Clemens

I’ve never understood how a knuckleball with basically no spin seems to waffle in the air instead of being perfectly parabolic (with some slowing due to air friction).

Josermember
1 month ago
Reply to  phoenix2042

Ever stick your hand out a car window on the highway? Air imparts a lot of force at those speeds, and those speeds are knuckleball speeds (knuckleball velocity seems low relative to a fastball, but fastball speeds in a car in many places will get you arrested for reckless endangerment). So a seam on one side pushes the ball out of its parabola, and also rotates it a bit, which causes another seam to get into the airstream on a different side, where the air pushes on it in a different direction, and so on.