Toronto’s Splitter-Squad Doubleheader

The Rays have been really good this year, but their bullpen has not. And this has come despite them largely sticking to their strategy of deploying relievers with a wide variety of different release points: two sidearm righties in Kevin Kelly and Ryan Thompson, one sidearm lefty in Jake Diekman, an over-the-top righty in Pete Fairbanks, and an over-the-top lefty in Jalen Beeks. This strategy has succeeded in the past: from 2020 to ’22, Rays relievers have never posted an ERA higher than 3.37; their collective ERA of 3.31 in that span is second only to the Dodgers (whose bullpen also happens to be struggling this year); and their FIP of 3.71 ranks third.

This year, things are different. Even with some modest improvements of late, Tampa Bay’s 12th-ranked 3.83 ERA belies ugly peripherals, including a 4.43 FIP that ranks sixth-worst. The rotation has been ravaged by injuries, forcing the team to turn to the bullpen earlier in games; Rays relievers have tossed 1.26 innings per appearance on average this season, the fourth-highest in the league. But some of this is due to the usage of followers, which is nothing new; perhaps what’s behind the drop-off is simply too much of an emphasis on forcing opposing hitters to deal with different looks.

The Blue Jays, perhaps unintentionally, have taken this lesson to heart. Having signed one of the best splitter-throwers in Kevin Gausman prior to 2022, they traded for another this past offseason in Erik Swanson. Despite the league-wide increase in splitter usage due to its potential as a platoon-neutral offering for pronating and sweeper-throwing hurlers, only 2.2% of all pitches thrown so far this year have been splitters. Given the uniqueness of the pitch, the Jays should have Gausman and Swanson throw on different days to maximize the surprise factor, right? Just as the Rays might save two similar sidearmers for different days?

Well, that hasn’t been the case thus far. Gausman has made 17 starts, and Swanson has pitched in relief in all but five of them, including 11 of the last 13. What’s more, Swanson has been markedly worse when pitching after Gausman, regardless of his opponent’s wRC+:

I defined “right after” as when Swanson came in during the same inning Gausman left the game or the inning after. I thought about dividing the “different game” category into when Swanson pitched the day after Gausman and otherwise, but Swanson actually hasn’t pitched the day after Gausman once this year, and he’s only pitched later in the same series twice. Additionally, given the small samples, I included opponent wRC+ just to ensure that there weren’t any major differences in quality. However I sliced it, though, Swanson has looked a lot worse pitching in the same game as Gausman, specifically in terms of his strikeouts and walks. And most of this is because of his splitter, despite its changes in shape being within the margin for error:

Arguably the splitter’s best shape — the most run and solid drop — has come when Swanson has followed Gausman (albeit not immediately), though the former’s splitter has still seen worse results when it comes to swinging-strike and called-strike rates in these instances. By run value, the splitter has actually been more effective when Swanson has pitched after Gausman, but the important underlying indicators — swinging-strike and called-strike rate — indicate that this shouldn’t continue going forward, especially when Swanson enters right after Gausman (and it’s already showing up in the strikeout and walk rates).

Are the two pitchers’ splitters really similar enough for hitters to have a bead on Swanson’s whenever he relieves Gausman? The following is a graph of every pitch, grouped by type and pitcher, thrown at least 100 times by a right-hander this season. I’ve singled out Swanson and Gausman’s split-pieces:

At least by movement, the answer is no. In the grand scheme, the pitches dart and dive to similar extents, but Gausman’s drops and runs more, making the difference between the two offerings akin to that of a sinker and a four-seamer. So what about when it comes to the pitches’ other characteristics?

MLB.com’s Mike Petriello used six inches as a demarcator for release point uniqueness, and Gausman throws his sinker over a foot closer to third base than Swanson on average. A side-by-side look indicates that some of this is just based on differences in how the two set up on the rubber, but Gausman also throws from a noticeably lower slot:

This certainly explains why Gausman’s splitter gets more arm-side run, but it made me wonder all over again why these two pitches aren’t different enough to fool hitters when they’re used in concert. And that wondering brought me back to spin.

In my piece on baseball technology earlier this week, I wrote about how important spin axis is as a pitch-recognition cue. Batters identify pitches based on the orientation of the spin, perhaps more so than anything else once the ball is thrown. And what do you know: Gausman and Swanson’s splitters have remarkably similar spin. The pitches rotate with backspin around axes just over six degrees apart on average, and Swanson’s converts a mere 5.5% more of that spin into movement (i.e., active spin, which also explains why his drops less). Yes, the two pitches have distinct movement profiles, and Gausman throws his split 1.3 mph harder on average, but once the hitter watches a couple of Swanson splitters and calibrates to the movement and speed, it’s off to the races.

Gausman and Swanson are both excellent pitchers. Gausman will often keep the game close, if not win it outright for the Jays; in those close situations, they want to bring in one of their lockdown relievers, and Swanson fits the bill. Perhaps that’s all that’s gone into their strategizing and they haven’t given the splitter-squad doubleheader a moment’s thought. Or maybe they’re experimenting, trying to figure out if the two splitters can work together. But even if that’s the case, it’s time for the experiment to end.

Stats are as of end-of-day Wednesday, June 28.