What is Michael King’s upside as a starter?

On Tuesday, Michael King tossed four scoreless innings. It was the first time he’d gone four frames in over two years, the latest step in the Yankees’ effort to assess if he can handle a starter’s workload post elbow fracture. King has mostly been solid in relief this season, and he’s seen his velocity jump since the All-Star break (though not quite back to last year’s peak), so it’s safe to say he’s feeling pretty healthy. In red are his per-game sinker velocities, while the blue line represents a smoothed five-game rolling average:

But King’s buildup is seemingly as much about his health as it is about the Yankees’ need to gauge their starting options for next year given the sheer amount of question marks they have in the fold.

So is King proving that he can be an option? Would he be an option on a less pitching-starved team? He’s always seemed like one given that he came up as a starter — reaching 149 minor-league innings in 2017 and 161.1 in 2018 — on account of a deep repertoire that has typically included at least four pitches. However, one major difference between present-day and prospect King is that he’s swapped his curveball for a sweeper that has roughly the same amount of velocity and drop but registers around 10 more inches of horizontal break. Per Stuff+, the pitch’s shape has graded out as 33 percent better than the average slider since King started throwing it in 2021, and it’s been 43 percent better this season.

This year, he’s turned to the sweeper more than any other pitch, at a 31.9 percent clip. However, sweepers have significant platoon splits, and managing hitters of either handedness will be crucial to King’s success as a starter. The right-hander throws his sweeper less than any other pitch against lefties, but he’s still throwing it 7.2 percent more against them this year than he did last season and without more success:

The past two seasons were small samples, but he’s saved just as many runs (0.7) as last year with the sweeper against lefties despite throwing 67 more of them. Most of its value against opposite-handed hitters has come as a “show-me” pitch: this year, 49.5 percent of his sweepers against lefties have been on the first pitch, 4.3 percent higher than his previous high. It’s generated called strikes on a solid 36.2 percent of these pitches (league-average called-strike rate on first pitches this year has been 30.6 percent), but it’s only a matter of time before hitters catch on and start hacking early and often.

After stealing strikes with the sweeper, King will usually turn to his four-seamer (which typically has more even platoon splits than the sinker) to continue getting ahead and then his changeup for a whiff. The changeup has been extremely adept at getting whiffs this year, with a 23.5 percent swinging-strike rate. But there’s reason to be skeptical of its success.

Firstly, Stuff+ grades its shape as 23 percent worse than league average. That’s likely because it drops 2.5 inches less than changeups at similar velocities. Additionally, the velocity differential between the changeup and four-seamer has been just 7.3 mph this year, while a 10-mph gap is a good heuristic for changeup effectiveness.

King has actually displayed reverse platoon splits this year: lefties have managed just a .236 wOBA against him while righties have put up a .316 mark. But on his career, those splits are almost even: lefties have a .288 wOBA and righties a .287. My bet would be that he’s been getting lucky stealing strikes with the sweeper, and the changeup is overperforming. The changeup locations have been good down and away, and his floor against lefties is still pretty solid. But I think King could improve his formula against them.

One way he could do so is by reintroducing the cutter he turned to occasionally in 2021. The cut-fastball is a platoon-neutral pitch that King could use to generate weak contact against lefties as well. It would be good to have a third weapon against opposite-handed hitters in the event that the changeup regresses and/or the show-me sweeper stops working.

What’s more, the cutter sat squarely between his sinker and sweeper in terms of velocity, horizontal movement, and vertical movement. This year, King has turned to the sinker-sweeper combo 81.1 percent of the time against righties, similar to the 85.6 percent clip he used them at last year. But hitters have caught on to this strategy, improving their wOBA by 69 points year-over-year. The cutter could help both pitches play up, bridging the movement and velocity gap between the two and making them indistinguishable.

As things stand, King seems to have enough to be a serviceable starter. But some tweaks could be in store once the league adjusts, at which point he’ll have to adjust back. We may not see any of those adjustments the rest of this year, but be on the lookout in spring training 2024, especially if King is still in the running for a rotation spot.