Slot Machine: Who’s Changed Their Release Point?

Though it feels like Opening Day was just yesterday, we’re officially a month into the 2023 regular season. On the macro level, that means the disappointing and surprising players are already starting to come out of the woodwork. More specifically (and importantly for writers like me), we’re at the point in the season when hitters are routinely cracking the century mark in plate appearances and pitchers are notching 35 innings.

Yet in some ways, this juncture is almost more maddening than Opening Day; we’re still in small-sample-size territory, but enough baseball has been played that we’re tantalizingly close to being able to take a hard look at some of the narratives being spun. For the time being, though, it still makes more sense to look at changes in approach rather than surface-level stats to predict rest-of-season production.

So I returned to a project I started this offseason — analyzing pitcher arm slots — to examine some hurlers who’ve made discernible tweaks to their release in accordance with early shifts in their performance. The equations I used to calculate these numbers can be found here.

Previously, I discovered that pitchers with more obtuse arm angles tended to generate more called strikes, their enhanced ability to freeze hitters perhaps due to the rarity of a sidearm or submarine look. So, case closed; if we find that a pitcher has become more sidearm-inclined this year, we should expect them to net more called strikes.

Not so fast. It’s one thing to look at pitchers who already throw from a certain slot and another to study pitchers who’ve changed their arm angle. So I conducted a new study, and this time I found that among the 312 pitchers who made 500 tosses in both 2021 and ’22, the most notable impact that arm-angle change had was on swinging-strike rate. Specifically, even after controlling for changes in zone rate, four-seam fastball usage, and velocity, each additional degree of average arm angle toward a submarine slot portended a .04% increase in swinging-strike rate — a statistically significant difference. Overall, these results indicate that if you’ve always thrown from a low slot, you might get more called strikes, but if you start throwing from a lower slot later in your career, you might generate more whiffs.

There have already been some fairly large changes in average arm angle on the young season, headlined by Kenley Jansen’s nearly 20-degree shift towards submarine-dom. With a change of that magnitude, a .04% increase in swinging-strike rate per degree could mean something. But there are a couple of caveats we have to consider first.

To begin with, that .04% number is an average. Not every pitcher who changes their slot will see that result borne out, especially since arm-angle changes alone only account for around two percent of the variation in swinging-strike rate according to R-squared. Second, the vast majority of hurlers alter their slots very little. Even if the .04% increase in whiff rate were set in stone, few would reap the benefits: Just 83 of the 312 pitchers from the dataset above changed their slot by five degrees or more; only 22 changed theirs by 10 or more.

At the same time, most of the signal does appear to be coming from hurlers at either extreme of arm-angle change. Thus, for that reason and for the sake of individual differences in how slot change impacts any given hurler, I decided to hone in on those with major changes in the early going this year. Below are the top five arm-angle increasers (i.e., moving lower to the ground) and arm-angle decreasers (moving closer to over the top) among the 320 pitchers who threw at least 500 pitches last season and have tossed at least 100 this season. For context, average arm angles ranged from the knuckle-scraping Tyler Rogers at 128.1 this year to the wholly-over-the-top James Karinchak at 2.9:

Here, the arm-angle increasers actually saw a smaller swinging-strike rate increase on average than the arm-angle decreasers. But two of the biggest arm-angle and whiff gainers — Kris Bubic (10.9 degrees, 5.8%) and Jordan Romano (10.8 degrees, 4.2%) — just missed the cut, as did two of the biggest arm-angle and whiff laggards: Tylor Megill (-10.8 degrees, -3.2%) and Domingo Acevedo (-8.6 degrees, -6.6%).

Regardless, I promised to hone in on those in the table, and there are definitely some names that pop out. For Jansen, drifting all the way to three-quarters territory seems like a drastic change to be making in his age-35 season, but it may be paying dividends. Perhaps his lower arm slot is allowing him to reach back for more, as he’s now averaging a career-high 94.7 mph on his famous cutter, which hasn’t topped 93 since 2017 according to Statcast. Here’s what his trademark offering looked like at release last year:

Compared to this year:

I’m no mechanics expert, but Jansen’s shoulders look a lot more square to the plate, and his balance seems better, leading to a more fluid motion overall. He’s been sharp thus far, allowing just one run in 8.2 innings with a 12/2 K/BB ratio, though there may be some literal growing pains, as the right-hander dealt with back tightness this weekend.

On the other end of the spectrum, Fulmer’s move to a more over-the-top release seems to have aided him in the swinging-strike realm, but he’s also yielded nine runs in 12.1 innings. Yet while he’s been hit hard with a 6.51 xERA to boot, his xFIP is at a career-low 3.81 thanks to a career-high 29.6% strikeout rate. And as with his surface-level statistics, there is more to his tweak than meets the eye. It turns out that, while he’s most frequently throwing from an over-the-top 35-degree slot this year, he’s also occasionally dropping down north of 50 degrees:

Meanwhile, north of 50 degrees is where he lived pretty much exclusively last season:

A quick glance at Fulmer’s release points on Baseball Savant tells you that his changeup (in green) is the one pitch he’s still dropping down for:

If he’s tipping the pitch, though, the results bear no indication. He’s thrown the change 27 times this year, and batters have whiffed on seven out of 13 swings; the other six cuts have resulted in fouls. The velocity on his change is up over three ticks without too much change in movement, which might also explain its newfound penchant for whiffs.

Perhaps Fulmer’s issue has been commanding his other pitches from the new release point. His bread-and-butter cutter has been demolished to the tune of a .455 wOBA and .443 xwOBA after hitters managed just a .269 wOBA and .268 xwOBA against it last year. He’s clearly missing over the plate with the pitch too often:

Since it’s his most-used offering, you might be thinking that he has to throw it for strikes in order to set up his secondaries. But he turned to it even more frequently last season and still had much better command of it:

If you don’t believe me, ask Location+ and Pitching Bot’s command metric, both of which failed to register Fulmer’s cutter last season but are giving it unspectacular grades of 97 (100 is average) and 37 (50 is average), respectively, this year.

Ultimately, hitters might catch on to the different release on Fulmer’s changeup and start mashing it, too. It could be a subtle enough difference that he can get away with it, or perhaps improved command on his other pitches will allow him to weather a less-effective changeup. But the best remedy might just be returning to the changeup release point with his other pitches. And it looks like that’s the strategy he’s been leaning toward as of late, at least with the cutter, even if it’s still not reaching a 60-degree slot like the changeup:

Sure enough, in the three appearances since he returned to last season’s cutter release point, batters haven’t managed a hit against the offering on three balls in play.

One more note on Fulmer’s bimodal arm-angle distribution this year. I found that for the 45 pitchers with such distributions last season, when their two peaks were at least four buckets apart (as was Fulmer’s), chase rate tended to be higher; Fulmer’s O-Swing% has increased by 5.1% this year by my calculations. But for the 24 pitchers who added another peak (i.e., moved from unimodal to bimodal) going from 2021 to ’22, chase rate didn’t increase as much as it did for other pitchers (including the 16 who moved from bimodal to unimodal). These were all small samples, but again this underscores the difference between starting out with a particular strategy versus converting to that strategy.

The jury is still out on any sweeping conclusions about whether any particular arm slot is superior to another. In fact, the absolute value of arm angle change had a relationship with swinging-strike rate of a magnitude similar to plain old arm angle change. In other words, it might be that arm-angle changes in any direction — or other mechanical tweaks for which arm angle change is just a side effect, as in Jansen’s case — help improve swinging-strike rate rather than just drifting toward a submarine slot. As things stand, though, arm angle is still useful for identifying changes in approach, especially at this early stage in the season.

Stats are as of end-of-day Sunday, April 30.