The Models Can't Make Sense of JP Sears

His elite VAA is confounding the computers

According to the computers, Oakland A’s lefty JP Sears is now a star. By the most visible “stuff” model, FanGraphs Stuff+, Sears ranks 19th out of 146 starting pitchers early in the 2024 season, around the same level as Grayson Rodriguez, Freddy Peralta, and Corbin Burnes.

JP Sears has also been one of the worst starting pitchers in baseball. So far in 2024, he’s allowed nine runs and struck out two batters. What’s going on? Why are the (statistical) models suddenly loving JP Sears despite his poor results?

Part of the explanation is a classic Stuff+ success story: Sears is throwing his best pitches more. In 2023, when his overall FanGraphs Stuff+ was closer to the middle of the pack, Sears leaned pretty heavily on his 93 mph fastball, throwing it more than 50% of the time. He also had two different sliders — one hard (the yellow dots on the plot below), one sweepy (the gold dots mostly clustered to the right of the yellow) — and the shapes often bled into each other. (The specs on his 2023 and 2024 pitches are at the end of this post if you’re curious.)

In 2024, Sears ditched the hard slider and went exclusively to the sweeper, and he’s now throwing that new sweeper more frequently than his four seam. (FanGraphs Stuff+ loves sweepers.) In addition, both the sweeper and the fastball improved their shape between 2023 and 2024; the sweeper added a few inches of glove-side break, and the fastball gained two inches of induced vertical break (IVB). And he started throwing a sinker with a ton of horizontal break inside to left-handed hitters.

Okay, so: moderate shape improvements + more sweepers + fewer mediocre fastballs = Stuff+ profit, right? Not so fast! Stuff+ doesn’t think the fastball is mediocre at all. In fact, it’s giving the four seam a 109 Stuff grade this season, significantly above the major league average for that pitch. What gives?

Though I cannot look directly inside the FanGraphs Stuff+ black box, I have a strong suspicion of what’s driving this above average grade: Sears’ superlative vertical approach angle (VAA) this season. Of the 249 pitchers this season with at least 25 fastballs thrown, Sears ranks 2nd in all of baseball in the flatness of his fastball, ranking only below the submariner Adam Cimber.

No starting pitcher comes close to his -3.5° VAA. Jared Jones, Joe Ryan, and Cristian Javier are somewhat close at -3.8°, and then there’s a huge gulf. Check out Sears all by himself on the far right of the average fastball VAA histogram, this one with a minimum 50 four seamers to limit the sample to mostly starters.

Combine his outlier VAA with roughly average IVB and velocity, and the computer spits out: good fastball. Sears, who stands at a modest 5-foot-11, gets to that steep approach angle through a combination of a lowish arm slot and being short, basically:

As Alex Chamberlain demonstrated in a series of articles for FanGraphs in 2021 and 2022, flat fastballs out of those low delivery slots are extremely effective at inducing whiffs and popups when located high in the zone. The theory: batters expect fastballs to come in at a certain angle, roughly -5°. When the pitch instead comes in at a much flatter angle, they tend to swing under the ball. If they manage to make contact, it’s often with the bottom of the baseball, resulting in a batted ball launched high and harmlessly into the air for an easy popup.

So how is Sears throwing this ostensibly good pitch and getting bad results? His start against the Tigers on April 5 tells the story.

Sears’ game plan against the Tigers was pretty simple. To righties, he wanted the fastball up and in and the changeup down and away. To lefties, he was looking to burrow the sinker inside to get ahead and finish them off with the high four seam or the sweeper.

The Tigers, unfortunately, were not biting. On the first time through the lineup, they mostly laid off the high fastball, instead focusing their attention low in the zone. Sears’ strategy was working well — they made good contact a few times — but delivered no concrete results.

That changed as soon as the lineup flipped around in the bottom of the 4th. After landing a surprise sweeper to Matt Vierling to get to 0-1, Sears tried to go upstairs with the fastball. Check out the target from his catcher Shea Langeliers:

If he hit that location, it would have been very hard to get the barrel on. Unfortunately for Sears, it was not located well.

Oops! Vierling crushed that pitch: 107 mph exit velo, solo homer, 1-0 Tigers.

This pitch explains the fundamental gap between the Stuff+ grade and the real-life results of the Sears fastball in 2024. If he hits his spot, he probably gets a whiff or some bad contact. Instead, Sears misses his intended location and pays the price. That elite VAA gives him a ton of juice at the top of the zone, but when he misses low — and he missed a lot against Detroit, including here to Vierling — it loses its powerful qualities and becomes pretty hittable.

It’s not a coincidence or bad luck that Sears gave up 34 home runs last year and ranked second among 2023 qualified starters in fly ball percentage. When hitters swing at their typical attack angle, designed to punish -5° VAA fastballs, they swing under the high four seam of Sears — but crush it directly on the barrel when it comes in below the intended target. And because the velocity is so-so, Sears does not have much of a margin of error.

The sixth inning demonstrated the location-specific efficacy of the Sears fastball. Look at the location on this pitch to Spencer Torkelson, which Torkelson launched 59° in the air for a harmless popup.

It’s right at the top of the zone — great 0-2 execution from Sears. But two batters later, the execution was not as clean. The result: a titanic first pitch chuck off the bat of Mark Canha (107 mph exit velo, 29° launch angle, 414 feet).

The sweeper might also be grading out above its real-life level. It certainly has the raw characteristics of a great weapon: With that huge bending action, he can swing it front door for a surprise strike to a lefty or even throw it to the back foot of a right-handed hitter for a whiff. Here’s a great one that took Riley Greene by surprise:

But a potential problem for Sears is the differential in average release height between the sweeper, at 4.4 feet, and the fastball, at 4.9 feet. That might not sound like much, but it could be enough of a difference to mean he is “tipping,” or revealing, which pitch is coming. In other words, hitters might be able to tell if it’s a fastball or a sweeper before he’s even released the ball. The fastball arm angle is on the left; the sweeper arm angle is on the right, in a notably more sidearm position:

As Ethan Moore wrote in a primer on VAA in 2020, everything is connected in pitching. Stuff+ knows about the release differential between the fastball and the sweeper, but do the above average grades on these two pitches fully reflect that knowledge? It’s unclear to me if they do — or, at least, it’s unclear to what extent the Stuff+ grades account for this relationship.

That’s a lot of negativity about JP Sears, but there is a blueprint in here for a successful MLB starting pitcher. The changeup is a great pitch, and his command of it was good in his start against the Tigers; he consistently hit his spot low and away to righties. And the sinker is a great addition to his attack plan to lefties — when he gets the pitch on the inside part of the plate, it is hard for hitters to do much except get jammed. If he can match excellent changeup command with a similarly dialed-in fastball, Sears could see great results quickly. (For what it’s worth, FanGraphs Location+ likes his command this year, though it’s not meant to be used for small samples such as this — a post for another day.)

There’s a wave of decreasing four seam usage across the league, and Sears is riding it turbulently. It is not necessarily a bad idea to decrease overall usage — the sinker is potentially a better fastball option, and could be easier to throw from his sweeper arm slot — but as his arsenal is currently constructed, his four seam still needs to deliver weak contact and occasional whiffs, especially to righties. In the start against the Tigers, it was often way above the zone, allowing for an easy take, or dangerously low, opening it up to hard contact.

Look at the four seam locations in red, courtesy of Baseball Savant. A few are well located near the top line, but he is mostly having trouble hitting that up-and-in target to righties:

If he can’t consistently harness the best qualities of his super flat fastball by throwing it at or near the top of the zone, opposing teams will keep copying the Tigers’ game plan: Ignore everything high, focus on mistakes low. And without fastball command, his excellent FanGraphs Stuff+ will be a mere curiosity.

Questions? Suggestions for future posts? I’d love to hear from you — michaelrosen3@gmail.com

2023 PITCH SPECS:

FF (Four seam): IVB=14.7, HB=-11.2, Speed=93.1, Ext=6.6, VAA=-4.0, Rel Height=5.0
FF: 50.1%

SL (Hard slider): IVB=7.2, HB=1.5, Speed=82.1, Ext=6.5, VAA=-5.9, Rel Height=4.5
SL: 12.3%

CH (Changeup): IVB=3.3, HB=-16.7, Speed=84.1, Ext=6.5, VAA=-6.2, Rel Height=4.5
CH: 15.1%

ST (Sweeper): IVB=5.6, HB=9.9, Speed=81.1, Ext=6.5, VAA=-6.1, Rel Height=4.6
ST: 22.3%

SI (Sinker): IVB=7.0, HB=-14.7, Speed=92.1, Ext=6.4, VAA=-5.3, Rel Height=4.6
SI: 0.1%

2024 PITCH SPECS:

CH: IVB=2.5, HB=-18.3, Speed=84.8, Ext=6.5, VAA=-6.3, Rel Height=4.5
CH: 16.7%

FF: IVB=16.7, HB=-11.7, Speed=93.0, Ext=6.5, VAA=-3.5, Rel Height=4.9
FF: 35.6%

ST: IVB=6.0, HB=14.3, Speed=78.9, Ext=6.5, VAA=-6.2, Rel Height=4.4
ST: 37.9%

SI: IVB=7.3, HB=-18.9, Speed=91.3, Ext=6.6, VAA=-5.3, Rel Height=4.6
SI: 8.0%

SL: IVB=12.4, HB=-1.6, Speed=79.5, Ext=6.5, VAA=-5.2, Rel Height=4.4
SL: 1.1%