They apply a level of pressure to their throwing arms that is required of no other position in mainstream sports. To be at their best, they have to possess a certain level of intelligence, manipulation, craft and command that cannot be matched.
And yet, of all the pitchers in Major League Baseball, one stands out. The most interesting, unique and creative hurler in the game: Yu Darvish.
Simultaneously, Darvish is the pitcher that got shelled in Games 3 and 7 of the 2017 World Series, the pitcher that took a perfect game and a no-hitter to the ninth inning, only to lose both with two outs, and the pitcher that nearly won the Cy Young Award last season. Oh, and only the polydactylous among us can count his unique pitches on two hands.
After dominating in 2020 with the Chicago Cubs, Darvish was sent to the San Diego Padres, along with personal catcher Victor Caratini, in a seven-player deal, adding another ace to San Diego’s stacked rotation. Darvish compiled the lowest ERA of his career (2.01) last season, a 2.26 FIP – another career-best mark – and a 31.3 strikeout percentage. Darvish finished fifth among qualified pitchers in strike+ at 116, and tallied a 61 RV- and 116 whiff+ .
Darvish is notoriously selective, assertive and diverse with his pitches. There has been plenty of discussion in the baseball world over just how many pitches Darvish really throws. In a recent interview with Rob Friedman, known to his 260,000-plus Twitter followers as Pitching Ninja, Darvish said he throws 11 different pitch types.
We identify all of these within our classification system and took a deep dive into each pitch to understand the method behind Darvish’s madness. We’ll include 2020 usage and command data on each, along with a few hints on how Padres fans new to Darvish can ID some of the trickier pitches.
Before we dig in, here’s a look at the horizontal and vertical movement on Darvish’s vast assortment of offerings (this view is from the catcher’s perspective).
FASTBALLS
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| 4-Seamer | 203 | 91.3 | 98.4 | 18.7 | 61.6 | 84.3 | 84.0 |
Darvish uses his four-seam fastball second most among all his pitches, but his 16% usage rate was actually the fourth-lowest primary fastball rate in MLB last year. The pitch should be the easiest one to spot with the naked eye – it is very straight with excellent ride (“upward” movement due to backspin that fights gravity and makes the ball drop less than expected). Like most of his arsenal, the velocity range is wide, going anywhere from 91 to 98 mph.
For many pitchers, the four-seamer is the most easily commanded pitch that a pitcher will use to set up his secondary pitches or to get a strike when needed. But for Darvish, the opposite is true. He only threw his four-seam fastball 6% of the time when behind in the count, compared to a roughly 40% league-wide usage rate. However, his 84.3 command+ on the pitch is 11th lowest among 185 four-seamers in the league (minimum 200 pitches), making it easy to see why he prefers to go different routes when he’s behind in the count.
Instead of a get-me-over offering, his four-seamer is more of an attack pitch with an elite 18.7 whiff percentage, making it the best swing-and-miss four-seamer among all starting pitchers. In those situations, he uses it most often up in the zone against lefties and down and away against righties.
Here are two perfect examples of Darvish’s four-seamer. Both have outside targets from Caratini and both hit 97 on the gun. The first is commanded very well to the outer part of the plate, but the second one to Goldschmidt shows that even when he doesn’t command the pitch well, it is still very effective.
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| 2-Seamer | 106 | 90.6 | 95.0 | 7.5 | 82.6 | 92.0 | 88.0 |
Darvish’s two-seam fastball isn’t always visible on TV to the untrained eye depending on the camera angle, but it has very good run and sinks much more than his four-seam fastball. Impressively, Darvish is able to add this movement while losing minimal velocity from the four-seamer.
The two-seamer is his preferred fastball against left-handers, and Caratini will almost always set up high and tight for the pitches, hoping Darvish can start it right at the hitter’s front hip and run it back over the plate. It isn’t a pitch he’ll get a lot of swing-and-misses on, so he’s typically hoping to either get a called strike or weak contact after jamming the hitter.
An almost perfect front-hip two-seamer from Darvish. Caratini sets up high and Darvish keeps it right at the knees, but Sogard is completely frozen by running action back towards the edge of the plate.
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| Hard Cutter | 53 | 89.6 | 94.7 | 9.6 | 78.3 | 73.0 | 82.3 |
The last of Darvish’s fastballs is his hard cutter. This pitch is one of his least used from start to start, and there was even a two-start stretch in August of 2020 where he scrapped it completely. It is a very traditional cut fastball with flat, subtle cutting action to his glove side, just a couple mph below his regular fastball.
He’ll almost always use this pitch to lefties, often throwing it up and in to play off the tunnel and action of his front-hip two-seamer. Much like the two-seamer, it isn’t a big swing-and-miss pitch, but because of the tunneling effect it creates, it still works as an effective pitch in its limited usage, resulting in an average exit velocity of just 82.3 mph, an excellent mark for any pitch.
The target is pretty much the same as the two-seamer above, but you can see the ball has slight cut instead of the running action of the two-seamer. Moustakas gets a poor swing on the ball and Darvish gets his weak contact for a quick out.
BREAKING BALLS
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| Slow Cutter | 442 | 80.5 | 92.2 | 13.9 | 71.8 | 111.6 | 83.1 |
Darvish calls his slow cutter his best pitch, and for good reason, throwing a little over 35% of the time. It is a cutter in name only, as its movement is much more similar to a traditional slider, with a few inches of glove-side movement and a wide range of vertical movement (and in fact, we treat these as sliders in our classic classification system). Movement is the key here though, because the velocity range is crazy. When he takes something off, it can get as slow as 80 mph. When he’s amped up in big spots, it can touch 92. And yet, it’s all the same pitch.
His confidence in the slow cutter is clear and obvious. With a 111.8 command+ mark, it is by far his best-commanded pitch, and he throws it a whopping 80% of the time when he’s behind in the count. It’s the only time Darvish is anything that resembles predictable.
Darvish will almost always throw his slow cutter glove side, both to righties and lefties. It’s a pitch he is most often looking to keep in the strike zone, because of both the aforementioned confidence and command as well as to help set up his slider as a swing-and-miss pitch down and off the plate. Speaking of setting up his slider…
Here is a great example of how Darvish can adjust the usage and velocity of his slow cutter. The pitches are very similar in movement, with tight spin and just a couple inches of horizontal movement. In the first clip, he has fallen behind in the count with no one on base, so he is comfortable just flipping up a low 80s cutter. In the second clip, the situation is much higher leverage and he has a pitcher’s count, so he can really attack O’Hearn with a 90 mph backfoot cutter.
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| Slider | 182 | 76.9 | 87.5 | 17.6 | 59.0 | 104.3 | 79.5 |
For the first time here, things start to get a little confusing. Darvish technically throws two different sliders, but the movement on each can blend together enough that it’s pretty much impossible to split the two up when classifying them, so for our purposes here we will label them all as a single slider.
The pitch itself has frisbee-like action that can get 10-plus inches of horizontal movement, and, much like the cutter, it has velocity range of more than 10 mph. There is a noticeable correlation between velocity, count and intent for Darvish’s slider. Early in counts, he usually keeps his slider on the arm side to drop it in for a strike around 78-82 mph. Ahead in the count, he’ll go more towards the glove side, sitting around 81-87 mph.
Hitters do a surprisingly good job taking this pitch – only his slow curve has a lower swing rate – but once they do commit, they only make contact on 59% of swings and have an average exit velocity of just 79.5 mph.
Much like the slow cutter, the differences in his slider by situation are clear here. The first is an outer third setup from Caratini that he often uses early in the count against lefties to keep the ball in the zone. In the second, you can see Caratini set up well off the plate and Darvish ramps the velocity up to 85 with that wicked horizontal movement that is much more extreme than his cutter, even though it comes out of the same tunnel.
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| Hard Curve | 122 | 74.6 | 85.3 | 18.2 | 71.8 | 106.9 | 8.25 |
The harder of his two curves, Darvish’s main curveball, thrown with a spiked grip, has sharp downward action. In many ways, the movement is similar to that of his slider, but it’s tilted 45 degrees downward. An excellent swing-and-miss pitch, it’s his preferred secondary offering against left-handers specifically.
His usage is pretty standard of a typical curve with more velocity. Caratini typically sets up square on the plate, looking for the ball down and sometimes below the strike zone to hunt for those aforementioned whiffs.
As mentioned, this is the target Darvish is most often throwing to with his harder curve. Even though the velocity sits right in the middle of his slider range at 82 mph, the movement here is much closer to 12-6 than you see with his slider.
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| Slow Curve | 36 | 62.5 | 74.1 | 11.1 | 66.7 | 82.2 | 77.2 |
Darvish’s least used breaking ball is his slow curveball. As mentioned above, the movement profile of this curve variation is similar to his hard curve, but he loses the spiked grip and really takes something off, dropping all the way into the low 60s in velocity sometimes.
It’s rare to see more than a couple in any given outing (10 of his 13 starts in 2020 had fewer than three), but when he throws it, it’s almost always within the first two pitches of an at-bat in an attempt to draw a take from the hitter and steal a strike.
Here is that same 12-6 movement of his hard curve, but Darvish really snaps his wrist and gets the velocity all the way down to 70 mph. As long as he keeps his sequencing unpredictable, hitters have almost no chance to be ready to swing at something like this so it’s a perfect strike stealer.
Breaking Ball Note: The situation often predicates how hard each of these pitches are thrown and it isn’t uncommon to see him start an at-bat with a 79 mph slider followed by an 83 mph cutter. Then once he’s ahead in the count, he could throw an 87 mph slider, a 90 mph cutter and an 83 mph curveball. So hitters really have no idea what type of speed or movement is coming, and it is impossible for a viewer to rely solely on velocity when trying to determine which pitch he is throwing.
OFF-SPEED
| Pitch | 2020 Total | Min. Velo | Max. Velo | SwStrike% | Contact% | Command+ | Avg. EV |
|---|---|---|---|---|---|---|---|
| Changeup | 36 | 85.8 | 92.9 | 11.1 | 75.0 | 73.9 | 84.1 |
| Splitter | 55 | 85.5 | 93.0 | 21.8 | 64.7 | 83.7 | 81.3 |
| Supreme | 12 | 85.3 | 92.9 | 16.7 | 81.8 | 96.0 | 91.3 |
It’s easiest to talk about all three off-speed pitches together, even though we can classify them all separately. His changeup is the easiest of the three to pick out. The circle-change grip is visible on a standard broadcast camera, and it gets consistent running action, similar to that of his two-seamer.
The splitter can give tracking systems and human eyes fits because of its cutting, glove-side action, but if you’re ever unsure, you can look for the split grip either in his mitt or at the bottom of his arm circle.
Last is the Supreme, a Darvish-specialty that got baseball Twitter excited pre-2020. The Supreme is a bit of a splitter/sinker hybrid pitch that is firm and has arm-side run. Admittedly, this is the hardest pitch to pick out because the grip itself is hard to tell apart on a broadcast, so among the 11 pitches we’ve classified here, this is the one we’re least confident in.
Darvish seems to use these three pitches relatively interchangeably. Of the 103 off-speed pitches he threw in 2020, all but nine of them came when he was ahead in the count, so they’re clearly viewed as out pitches in Darvish’s mind. He does prefer the changeup (4.2%) against left-handers and the splitter (6.5%) against right-handers, but, at least during his time in Chicago, he showed a willingness to throw both to anyone at various times. It is worth noting that the splitter has a much higher whiff rate (21.8) than the changeup (11.1), which could lead Darvish to increase his splitter usage in San Diego.
It might be tough to see, but the change grip is visible as Darvish brings the ball out of his mitt. This is the classic running, fading action that you’ll see any time he throws his changeup.
At first it might look like this is a cutter based on the glove side action to the pitch, but there are two keys here that make it clear this is his dastardly splitter: catcher setup and grip. First, that Caratini squat centered on the plate is something that only really happens on off-speed pitches or hard curves. The kicker is the grip though; the split in his fingers is visible in the mitt just as he starts his leg kick, and at the bottom of his arm circle you can see his pinkie and ring fingers tucked.
As mentioned before, it’s hard to be 100% sure that this is the Supreme, but it’s our best guess for one. The gap in his fingers can just barely be seen in his mitt while in the stretch, but he has more hand on the ball than he does on his typical splitter. The movement here is much more changeup-like with arm-side run instead of the dancing cutter shown earlier.
Lucas Haupt contributed. Design by Andrew Skweres.