The High Sinker Paradox

Peter Aiken-USA TODAY Sports

I thought that today’s article was going to be an easy one to write. Reading Alex Chamberlain’s post on the pulled fly ball revolution made me imagine the worst pitch a pitcher could throw: a sinker that ended up high and inside, an easy-to-contact fastball in the area of the plate that leads to the most damaging types of opposing batted balls. Then I extrapolated my idea out a little bit. Maybe I could look up the pitchers who throw their sinkers high in the zone most often. We could all laugh about how they’re called “sinkers” — so that’s clearly a bad place to throw them. Maybe we would gawk at a table of a few pitchers who do this bad thing, and then we could move on with life.

Well, I can do at least one thing. Here’s a table of the pitchers who threw elevated sinkers in or around the strike zone most frequently in 2023:

High Sinker Power Users
Pitcher 2023 Sinkers Up-In-Zone%
Michael Tonkin 785 42.3%
Alex Wood 762 41.2%
Ryan Yarbrough 441 40.1%
Steven Matz 1045 40.0%
Drew Smyly 933 39.4%
George Kirby 611 38.6%
Josh Hader 765 37.5%
Brusdar Graterol 405 36.8%
Aaron Civale 373 35.9%
Jhony Brito 465 35.3%

Haha, look at these chumps. Their pitch is supposed to go down, and they throw it up. What an obvious error! And look how frequently these guys are doing this obviously bad thing. They must be getting hammered on those pitches, right? Wrong:

High Sinker Power Users… Are Great
Pitcher 2023 Sinkers Up-In-Zone% RV/100
Michael Tonkin 785 42.3% 1.7
Alex Wood 762 41.2% 2.9
Ryan Yarbrough 441 40.1% 2.2
Steven Matz 1045 40.0% 2.4
Drew Smyly 933 39.4% -0.5
George Kirby 611 38.6% 2.4
Josh Hader 765 37.5% 3.4
Brusdar Graterol 405 36.8% 5.2
Aaron Civale 373 35.9% 5.3
Jhony Brito 465 35.3% 4.0

That’s Statcast’s measure of run value per 100 pitches, where higher is better for the pitcher. It’s formulated differently than the run values you’ll see on our website, with more attention paid to the context of a pitch. But regardless of how exactly the calculation works, all but one of them is positive. That seems weird. In fact, the league as a whole produced 479 runs above average on high, in-zone sinkers in 2023, 2.0 runs per 100 of them.

This makes very little sense on its face. It would be one thing if a few outliers were living large with high sinkers. Hader, for example, is an outlier on this list. He throws with a sinker grip, and so his pitch gets classified as such, but it behaves like a four-seamer. I expected to find Hader near the top of this list, and I know that his fastball performs well. But everyone?! In aggregate?!? What the heck is going on here?

My first guess: Maybe hitters just aren’t swinging at them. Here these pitchers are, lobbing up meatballs, and batters just let them go past. Believable? Maybe, but inaccurate. Here’s how those 10 pitchers have done when opponents swing at their high sinkers:

High Sinker Power Users, Results on Swings
Pitcher Whiff% BABIP wOBACON RV/100
Michael Tonkin 27.7% .259 .327 2.8
Alex Wood 33.3% .231 .326 5.7
Ryan Yarbrough 16.1% .250 .308 3.7
Steven Matz 31.1% .278 .323 4.4
Drew Smyly 17.4% .296 .460 -0.4
George Kirby 14.9% .258 .332 2.4
Josh Hader 28.0% .220 .237 5.9
Brusdar Graterol 8.5% .148 .131 7.4
Aaron Civale 12.8% .161 .201 8.5
Jhony Brito 18.8% .220 .254 6.0

These numbers are hard to believe. These pitches are supposed to be bad, and yet only one of the 10 pitchers to throw them most often allows a wOBACON above league average. Those whiff numbers look pretty juicy, too. Batters swung through Matz’s elevated sinkers more frequently than they did through either his curveball or changeup, the ones that are supposed to generate misses. This isn’t just cherry picking, either; across the majors, elevated sinkers did phenomenally well even when opponents swung at them.

Now for something even weirder: These sinkers outperformed four-seamers thrown to the same location. I’m not just talking about our cohort of frequent high-and-sinky throwers, either; in aggregate, high sinkers allowed a .338 wOBA on contact last year, while high four-seamers checked in at .374. And it’s not just a contact advantage; sinkers outperformed four-seamers, even taking the inherent advantages of four-seamers (more whiffs, basically) into account. On all pitches thrown in the upper reaches of the plate or just off of it, four-seamers were worth 1.6 runs above average per 100 pitches. Sinkers, as you’ll recall from above, came in at 2.0 runs above average. Cutters split the difference at 1.8, in case you were curious:

Results by Pitch Type, Up-In-Zone
Fastball Type Whiff/Swing wOBACON RV/100
Sinker 17.4% .338 2.0
Four-Seamer 27.1% .374 1.6
Cutter 25.6% .351 1.8

So OK, fine, this effect sure seems to be real. High sinkers rock. They rock more than sinkers do overall, in fact: Sinkers thrown in or near the strike zone racked up 1.6 runs above average per 100 pitches in 2023. But what could be causing this counterintuitive result? The first place to look is for a selection bias. That’s basically the first place to look for everything in baseball analysis, to be honest. The pitchers throwing high sinkers are probably disproportionately good at throwing them. Otherwise, they wouldn’t throw them so much. Side note: If you want a sign that you’re on the right track on some research, Justin Choi asking the same question as you is as good as it gets.

To look at this, I started with a list of all pitchers who threw sinkers in or around the strike zone last year. Then I took a separate list of the run value that each pitcher produced on high sinkers. I used the first set to weight the second – in other words, I worked out how good high sinkers would be if each sinker-throwing pitcher threw up in the zone at an identical rate. But there’s no joy here. In reality, as you’ll recall, sinkers have produced two runs above average per 100 pitches; 2.047, to be precise. If you weight it based on overall sinkers thrown, rather than on high sinkers thrown, to remove the selection bias, high sinkers produced… 2.042 runs above average per 100 pitches. Yeah, those numbers are the same.

Other smart researchers are on the trail of this effect, and that’s where I looked next, with a little help from an anonymous internet friend. Last October, Eno Sarris wrote about pitchers deceiving hitters with up-in-zone pitches that don’t behave like four-seamers. Could that be what’s going on? I mean, definitely! That’s an effect that’s much harder to calculate, because you can’t just do some kind of easy weighting like I did. But I did make an attempt: I re-sampled the data for only pitchers who throw sinkers and four-seamers at least 15% of the time each. How’d they do on their elevated sinkers? They produced 1.96 runs per 100 pitches on high sinkers, as compared to 1.86 RV/100 on all sinkers. That’s less than the gap produced by all sinker throwers. Curiouser and curiouser – if it’s a trickery effect, it’s not one that shows up in an obvious way.

Is it a platoon effect? If you’re a Ben Clemens completionist, you might remember that I’m obsessed with sinkers and wrote about how the league is using them as a platoon pitch a lot these days. Maybe up-in-zone sinkers are disproportionately thrown to same-handed batters, and if we were to look only at, say, right-right matchups, the effect would disappear. Except… right/right sinkers thrown in or around the zone produced +1.7 RV/100 last year, and up-in-zone right/right sinkers checked in at 2.2 RV/100. Same effect, yet again. In addition, sinkers still outperform four-seamers up in the zone, even restricting the sample to right/right, by a score of 2.2 to 1.9. Try as I might, I just can’t seem to make these up-in-zone sinkers bad.

Well, OK, that’s not quite true. Here’s one place where up-in-zone sinkers do worse: when they miss. When pitchers have thrown high and not near the plate with fastballs, they’ve done quite poorly, of course. Batters don’t chase fastballs very often. Four-seamers that miss high in what Baseball Savant classifies as the “chase” zone – a place where batters sometimes chase pitches but where the pitches are almost never called strikes – have produced a pretty rotten -3.3 runs per 100 pitches. That’s because hitters generally take them. But sinkers in the same location are a good deal worse, -4.8 per 100 pitches.

That just makes sense. Four-seamers draw more swings in general, and more whiffs when hitters do swing. Aim high and miss high with a sinker, and you’re phenomenally likely to end up with a called ball. At least a four-seamer that misses high leads to some ugly swings once in a while. And missing low isn’t good, either; over the dead center of the plate, exactly middle height and centrally located, fastballs of all types get absolutely demolished when hitters make contact. That means that whiffs are at a premium. Four-seamers outperform sinkers there, too:

Fastball Results by Location
Fastball Type Dead Middle RV/100 Up-in-Zone RV/100 Too High RV/100
Sinker 0.9 2.0 -4.8
Four-Seamer 1.2 1.6 -3.3
Cutter 0.2 1.8 -4.5

In fact, then, my gimmicky high sinker statistic is pointing at a different effect entirely: accuracy. Sinkers are really good when they’re located well, away from the red hot center of the strike zone but also close enough to the zone that they’ll at least get called a strike sometimes. That’s because sinkers draw far fewer swings than four-seamers, particularly high in the zone:

Fastball Swing Rates
Fastball Type Dead Middle Swing Rate Up-In-Zone Swing Rate Too High Swing Rate
Sinker 70.6% 54.1% 15.1%
Four-Seamer 76.5% 65.8% 18.7%
Cutter 78.9% 56.2% 10.9%

So yes, up-in-the-zone sinkers perform well, because hitters leave the bat on their shoulder too often. It’s harder to swing at a pitch with that shape up high than it is to swing at a four-seamer, apparently. Maybe that’s because batters have looked for four-seamers for a long time. I’m not here trying to prove causation or anything, but it’s undeniably the case; that huge advantage in swing rate (if you’re throwing in the strike zone, swings are obviously bad) is at its greatest up high but still in the strike zone, which makes that a good place to throw your sinker, so long as you don’t miss in either direction.

Don’t try to bet on this effect too much, though. Part of the reason that high sinkers are producing suboptimal decisions from opposing batters is that they’re rare. Pitchers threw four times as many up-in-zone four-seamers last year. Sinkers in general are still less represented than four-seamers; pitchers threw twice as many four-seamers as sinkers, period. And sinkers performed better overall, but only by a hair, with selection bias likely playing a huge part here. If you don’t have a good sinker, you just don’t throw it, but far more pitchers throw marginal four-seamers.

So in the end, maybe the answer is that pitchers with good sinkers really should throw them high in the zone. They should throw them low in the zone, too, and they do, 50% more frequently than they throw them up in the zone. They should also avoid grooving pitches down the center, and avoid missing the zone by enough that it’s an easy take. All of these things are obvious. I’m basically saying that good pitchers should hit the edges of the strike zone with their good pitches. We all knew that already. It’s just fun to illustrate it with the strange case of the high sinker, and its recent surprising run of dominance.





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

15 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
dbannonmember
3 months ago

I actually do think it’s a platoon thing, the key being that high fastballs > low fastballs, and R/R two seam > R/R four seam.

Essentially, the word “sinker” taking over for “2-seam” has made us concerned about the vertical break more than the arm-side run, which I think is doing the work.

Is it worth checking horizontal movement rather than pitch type? My guess is that that’s where the breakthrough lives

Nathanielmember
3 months ago
Reply to  dbannon

I had that exact same thought. A zippy two-seamer that breaks arm side at the last second might be really effective up in the zone. If you’re a same-handed hitter, it probably looks like it’s about to break your jaw.

The Fenway center field camera did a nice job of picking up this movement on Eduardo Rodriguez’s fastball (he throws a four-seamer but it used to have a lot of arm-side movement). It would either look like it was going outside to a lefty and then dip back to the zone, or look hittable and then go straight toward the hitter’s face.

Ivan_Grushenkomember
3 months ago
Reply to  Nathaniel

Yes arm side movement to same handed hitter would likely be tougher to hit high than low

Ivan_Grushenkomember
3 months ago
Reply to  dbannon

Yes a purely downward motion would be terrible, like maybe a split-finger fastball.

connjc
3 months ago
Reply to  dbannon

this intuitively makes sense. its why cutters are often used against opposite handed hitters and a horizontal moving two-seamer has a similar movement profile just in the opposite direction. its easier to jam a hitter up in the zone