Logan Gilbert’s new splitter might be just what he needed

Logan Gilbert had a phenomenal first two years, putting up 5.4 fWAR over 305 innings. But as good as he’s been, there’s been an exhaust pipe in his Death Star. Even as he’s racked up strikeouts and avoided walks, two pieces of ace-dom’s holy trinity, his contact management has left a lot to be desired. Check out the Statcast sliders for what happened when hitters got their bats on the ball.

Via Baseball Savant

Thus, when you zoom out from FanGraphs’ FIP-based WAR to Baseball Reference’s runs-allowed-based WAR, he drops from 5.4 down to 4.0. Baseball Prospectus, whose proprietary Deserved Run Average accounts for expected results based on quality of contact, pegged him as just 6% better than the average pitcher. The more you take contact into account, the worse he looked.

His first three starts of 2023 suggest that he may have found the solution.

This nice-and-easy groundout had an expected OPS of just .411 thanks to the new arrow in Gilbert’s quiver: a split-finger fastball, which he developed after googling every splitter he could find and then enlisting the help of the folks at Driveline.

Logan’s always primarily relied on his live-wire four-seamer, which comes in even hotter than its mid-90s velocity suggests, thanks to his elite extension giving batters even less time to react to it. But Logan’s secondaries haven’t been quite good enough to keep hitters off balance. By run value, his curveball is basically average. His slider was bad in 2021, but after working with Robbie Ray, he improved it to a solid option last year, getting whiffs on 30.4% of batter’s swings in the second half, up from 21.6% before the break. (League average was 34.3%.) A great fastball, solid slider, and average curveball is a recipe for a pretty good starter. But who’s satisfied with pretty good?

When Gilbert arrived in MLB, his fourth offering was a changeup, but I use that term loosely.

Even after changing grips for 2022, he still had no idea where it was going. Or maybe he did, I guess. I certainly didn’t.

Late in the year, chasing the hottest trend in 2022 Mariners pitching, he experimented with a sinker to complement his arsenal. Robbie may have inspired that addition, but the results were not great, Bob! Batters hammered it for a .403 wOBA, whiffing on just 9.1% of their swings.

The idea made a lot of sense though. A pitch like a changeup or sinker can be a great secondary for a pitcher whose best pitch is his four-seamer, as it is for Gilbert. Beyond the good velocity and elite extension, Gilbert generates a high-90s spin efficiency on his fasetball at a 12:30 axis. That backspin acts against gravity and stops the ball from dropping as much as it should, which can make the hitter swing underneath it. Here’s a totally randomly selected hitter who can’t figure out how high out of the zone this Gilbert four-seamer is going to be:

A changeup or sinker helps magnify this effect because those pitches look like four-seamers (whereas true breaking balls do not) but they move differently than a four-seamer—changeups drop down, and sinkers (despite the name) move horizontally or “run.” You can fool batters more often, on both the four-seamer and the sinker/change if that secondary is a good pitch, but Logan’s changeup and sinker were not good pitches.

Enter the split-finger fastball. Like a sinker and a circle change, its spin makes it look like a four-seamer pretty late in its flight path. Right until it doesn’t.

A splitter is thrown by lodging the ball back between your index and middle fingers. If you really wedge it back there, you could call it a forkball, but those tend to cause injury. Logan’s big hands help him get the same effect without having to shove the ball all the way to the webbing of his fingers. Asked about his grip, he said, “It’s just kind of natural for me. I just get around the ball really easily and the grip really takes care of the rest.”

Despite the nomenclature of a “split-finger fastball,” it’s actually a changeup, coming in meaningfully slower than a pitcher’s four-seamer, in Logan’s case by about 10 mph. And like a circle change, it drops at the end. A good one makes the batter swing over the top of it. Put it together with a four-seamer and you’ve made it impossible for the hitter to guess whether to swing higher than it looks like he should or lower. The batter has no idea where to aim the barrel: Both pitches should work better.

Here’s José Ramírez getting one that he can’t handle, even though it comes in middle-middle. It looks to me like exactly how this is supposed to work; I think JRam thought this was a four-seamer.

More deliciously, here he is swinging right over the top of one. Like Robinson Canó and a PED test, he’s failed on this pitch twice.

Of course, throwing pitches is a pretty difficult thing to do, biomechanically speaking. It’s why Logan Gilbert’s an MLB pitcher and you’re not. Even the most gifted athletes can’t simply pick up a new pitch. It has to work for their hands and their motions. And just like the changeup didn’t work for him, it wasn’t a sure thing the splitter would. But the splitter might be a perfect match for Logan’s mechanics.

He says it’s a natural fit for him. “It starts with the arm action and release point, which was always different on my changeup, so I think right from the get-go it helps me out that [the splitter is] very similar,” he said. “I’m basically just throwing the fastball with split fingers; I throw it the exact same way, it comes out the same way”

Gilbert has thrown 29 of them through his first three starts, serving it up to both lefties and righties, early in the game and late. And the results have been like Calvin Harris’s latest album: zero hits. That’s hardly been the result of good BABIP fortune. Hitters’ average exit velocity of just 80.1 miles per hour is supporting an expected batting average of .044 and microscopic expected slugging of .052.

Let’s take a look at how it’s changed those sliders on his contact stats.

Via Baseball Savant

With Logan, everything comes back to the fastball. Last year, he attributed the improvement on his slider to Robbie Ray’s advice to think more about his fastball when throwing the pitch. This year, he speaks similarly of the splitter: “It just feels like throwing a fastball, I really feel loose and like I’m just starting behind the ball but ‘flicking’ through, just like I would with a fastball, except instead of backspinning the ball it just slips out of my hand. But it’s the same thought process basically as my fastball, which helps a ton.”

The concern when you add a pitch designed for weak contact is that you can end up sacrificing strikeouts. But that hasn’t happened, at least so far. The splitter’s getting that weak contact, but it’s also getting more swings and misses than either of his breaking balls, with six whiffs on 19 swings, roughly in line with an average splitter’s whiff rate. It’s not the wipeout pitch of the splitter-dominant pitchers like Shohei Ohtani or Kevin Gausman, who run whiff rates in the mid- to high-40s, but it is fitting in nicely with the rest of his arsenal.

Overall, rather than losing strikeouts, he’s punching out more hitters than ever. His 29.9% strikeout rate is about a quarter more than his career average, despite having two of his three starts against the high-contact Guardians. His career K/BB+ has risen to 145, which is to say that his strikeout-to-walk ratio is 45% better than the average pitcher. The guys who can pull that off and limit the damage on contact are aces. Only the Cy Young candidates can do it all.

If there’s a red flag, it’s that a lot of the weak contact has been pop ups. I know, I know, pop ups are terrible contact! But that’s not really how splitters are supposed to work. Splitters are supposed to drop; hitters are supposed to swing over the top of them or mash some worms if they do make contact. Pop ups, by contrast, happen because a batter got underneath the pitch. The type of contact Josh Naylor gets here is going to register in the stat pages as bad: 76-mph exit velocity and a 41-degree launch angle do not make for extra base hits.

But this wasn’t the plan. Look at where Tom Murphy sets up, expecting the splitter to do what splitters do and fall away, and compare it to where Naylor makes contact.

This wasn’t an aberration either. Of the 13 times hitters have made contact, six have had a launch angle higher than 40 degrees. Four of those six have come on pitches at the top of the zone. Keeping the barrel off the ball is keeping the barrel off the ball, but I’m concerned that things aren’t as good as they seem.

So far, the splitter has helped Logan pitch like one of the game’s elites, but the pop-ups raise the possibility that the splitter might be just another one of Dr. Gilbert’s failed experiments. Either way, nobody should draw their final conclusions based on a 29-pitch sample, especially the first 29 pitches: on the one hand, the pitcher is still learning the pitch, and, on the other hand, the hitters haven’t been able to prepare. Tonight, we’ll see a Brewers club whose scouting and video departments surely spent all weekend on this, which will make it a real test. But these are the things I’ll be watching in his next few outings. On the whole, Gilbert’s splitter shows promise, especially since he seems so much more comfortable with it.