A niche for Yusei Kikuchi's new and improved curveball

Yusei Kikuchi and the Angels jumped the gun on a saturated starting pitcher market on Monday, agreeing on a three-year, $63 million pact. Los Angeles is betting that Kikuchi’s late-season revival was legit. The veteran left-hander reinvented himself with the Astros, following a simple new-age premise: He threw his best pitch (slider) more often, and slashed the usage of his worst pitch (four-seam fastball). 

I think there’s a little more nuance to Kikuchi’s restoration than that theory suggests. To explore the innerworkings, I want to take a look at his curveball, a pitch he mostly shelved in the immediate aftermath of the trade. 

Kikuchi’s curveball usage in 2024 by month

	Usage	wOBA
April	27.7%	.205
May	29.6%	.294
June	21.2%	.475
July	11.9%	.525
August	5.6%	.000
September	15.4%	.126

After all but eliminating the curveball in August, Kikuchi soft-launched the pitch again in September. In order to understand the logic behind these decisions, we ought to remember why Kikuchi began throwing a curveball in the first place. 

As Sportsnet’s Arden Zwelling explained in April, Kikuchi drew inspiration from then-teammate Chris Bassitt to “add an early-count breaking weapon he could use to steal strikes.” After toying with the idea of developing a sweeper (like Bassitt), Kikuchi and the Blue Jays settled on a curveball, a pitch that would cause less alterations to his release point and arm slot. 

But early success spurred a career-high usage rate, and the pitch eventually deviated from its founding purpose. In two separate months — September 2023 and May 2024 — Kikuchi’s curveball was his most-used pitch when he fell behind in the count. So much for an early-count strike-stealer. 

After tinkering with the rest of Kikuchi’s pitch mix, the Astros seemed to place an emphasis on the Blue Jays original idea. The curveball could still have significant value, when used right. And so Kikuchi became what we will call a “disproportionate curveball user.” 

Kikuchi’s Disproportionate Curveball Usage	0-0 Rate	Overall Rate	Difference
2024, With Toronto	26%	23%	3%
2024, With Houston	36.4%	10.1%	26.3%

While Kikuchi slashed his curveball usage with the Astros, he heavily relied on it to begin an at-bat. Sure enough, the pitch generates the majority of its called strikes on 0-0 counts. Intuitively, 0-0 would seem like an ideal count to ‘steal a strike.’ 

Ben Clemens explored the first-pitch curveball phenomenon for FanGraphs back in 2019. Clemens concluded that the success of first-pitch curveballs largely depends on the assumption that hitters aren’t going to swing; when the calculus changes, the pitch loses most of its value. Again, that adds up: You can’t steal a strike when a hitter puts a ball in play, and these purported strike stealers don’t want any contact. Because, as Clemens also found, the first-pitch curveball feigns tend to throw bad curveballs. 

But here’s the problem: Hitters have become far more aggressive. As Clemens quipped in 2019, “that free strike is becoming less free.”

Right-Handed Batters (RHB) vs. 0-0 Curveballs 

Year	Swing Rate
2019	21.7%
2023	23.4%
2024	23.3%

This holds true for Kikuchi: He induced more swings than the average MLB pitcher off 0-0 curveballs to right-handed hitters. But with Houston, the quality of contact generated by Kikuchi’s first-pitch curveballs improved vastly. That placed him at odds with the disproportionate curveball users that Clemens examined in 2019. 

RHB vs. 0-0 Curveballs from Kikuchi in 2024

Team	Swing Rate	wOBA	Avg. EV
Toronto	25.9%	.318	91 MPH
Houston	25%	.000	70.46 MPH

For other pitchers, stealing a strike is extremely important. It was that way for Kikuchi with the Blue Jays. But something happened with the Astros. He modified his curveball to guard against hitters’ increased aggression, doing his best to suppress the hard contact that befell other disproportionate curveball users. And these changes didn’t occur in a vacuum.

Location

Recall Kikuchi’s original motivation to throw a curveball: stealing strikes. To do just that, Kikuchi wants to place his curveball away from RHB, a spot that he hit routinely with Houston. 

As a Blue Jay, Kikuchi appeared to have less of an intent with his curveball. His called strikes are less concentrated, bleeding into different parts of the strike zone. We do, though, see the formation of a small cluster on the outer half, which would become Kikuchi’s sweet spot with the Astros. 

In all, the curveball became a more competitive pitch — reflective of its refined purpose. In September, Kikuchi’s curveball averaged a Plate Vertical of 2.03 feet, its highest single-month average all season. Kikuchi’s misses are revelatory: With the Astros, he stopped burying the curveball, a reaction to a drop-off in both Swing% and Chase%. On top of that, when he buries a curveball, he can't steal a strike.

The curveball's In-Zone% in September still dipped well below early-season rates. But for Kikuchi, it’s less about the strike zone than it is the specific outer half of the plate. He simply began to pound the location where he knows he can steal strikes. 

But what happens if the hitter swings? Remember, as we demonstrated through the lens of Clemens’ exploration, hitters are increasingly more aggressive against 0-0 curveballs.

Luckily, Kikuchi made additional changes to succeed in contact situations, too. 

Shape

Command has never been Kikuchi’s strong suit: Since his MLB debut in 2020, he ranks in the 22nd percentile among qualified starting pitchers in Location+. In 2024, he cracked the 50th percentile for the first time, in part because of his curveball, which generated a 105+ Location+. That’s the highest of any pitch in his arsenal. 

The Angels evidently believe that Kikuchi’s command will stick. With the curveball specifically, I think that belief originates from its new shape. 

Kikuchi’s curveball looked vastly different in September than it did earlier in the season. It was slower, had more depth and, most importantly, more gloveside movement. Predictably, models liked it much better. 

• Curveball Stuff+, TOR — 98

• Curveball Stuff+, HOU — 102

Let’s break down the evolution of Kikuchi’s curveball shape through the lens of three individual months: May, when he used his curveball the most; July, when his curveball was hit the hardest; and September, when his curveball found the most success. 

	Average Vertical Break (Induced)	Average Horizontal Break	Average Velocity
May	-3.9 in.	-4.0 in.	82.4 mph
July	-4.2 in.	-4.0 in.	83.4 mph
September	-7.0 in.	-7.0 in.	79.8 mph

For the moment, let’s focus on the added horizontal break: Kikuchi generated an additional three inches of gloveside movement on his curveball after joining the Astros. What may that do? Well, with the Blue Jays, Kikuchi often missed off the outer half of the plate, creating a large cluster of balls — pitches that were intended to steal strikes.

Kikichi's CB misses with Toronto

If his curveball is going to break an additional three inches to his glove side — towards a right-handed hitter — then it would make sense that Kikuchi would miss off the plate less often. Sure enough, with the Astros, Kikuchi located with precision. In part due to its refined shape, that cluster of balls disappeared. I think this offers a valuable lesson in crafting the shape of a specific pitch to maximize the specific purpose of said pitch. 

Differential 

Kikuchi didn’t just add horizontal break to his curveball. He added nearly three inches of depth, while cutting almost 4 MPH in velocity. We can best visualize the differences through a 2D pitch plot generator of Kikuchi’s arsenal with the Blue Jays, compared to his arsenal with the Astros. 

What we see here is that Kikuchi’s new curveball shape allows the pitch to have a better relationship with the rest of his repertoire. 

Let’s look at the curveball and slider — Kikuchi’s best pitch, the one responsible for his restoration. These two pitches are close enough together on the movement plane to cause a state of confusion among hitters. But with the Astros, the pitches (on average) don’t move as similarly, thanks to

the additional drop on the curveball. Add in the velocity separation — from a 6 MPH difference in July to a 10 MPH difference in September — and the pitches profile differently. Ideally, that leads to weaker swings, especially considering the fact that Kikuchi also likes to cluster his slider on the outer half of the plate. Two pitches in the same spot with a greater velocity differential and a greater movement differential — albeit still appearing similar out of Kikuchi’s hand — will lead to better results. 

Recap

The ultimate question is, naturally, of sustainability. Can Kikuchi carry his late-season success further west? 

At least with the curveball, these changes should be pretty sustainable. Maintaining a new shape should help him offset any location issues, which should in turn lead to “stolen strikes” and, if a batter swings, conducive contact. As Houston correctly predicted, Kikuchi’s curveball can continue to serve a valuable purpose, albeit a niche one. He is a rare example of a disproportionate curveball pitcher with a good curveball. Maybe that lays the groundwork for an increased curveball usage in 2025, so long as he keeps the pitch’s original purpose in mind.