High Sliders: Junk or Genius?

In my time at Sports Info Solutions this summer, I scored both of Janson Junk’s 2022 major league starts. Typically, getting assigned to an Angels’ game, especially a Mike Trout-less one (he was injured at the time), would elicit a groan. However, come the second Junk start, I was admittedly a bit excited, because in his first turn, I saw a lot of this:

I kept the audio in that clip so that you can hear the announcer say “there you go, there’s another one” — specifically, another whiff on a high slider. I put the announcer’s assertion to the test by defining a high pitch as one in the highlighted part of Statcast’s strike zone:

In that start against the Royals alone, Junk threw 36 sliders, 17 of which were high. All four of his slider whiffs came on the high hard ones. In his next start against the lowly A’s, Junk didn’t fare as well, but the high slider wasn’t to blame. He threw 24 more sliders, eight of which were high. His only slider whiff came on a high one, and the two doubles he allowed off sliders were not off high ones.

Sadly, that’s all the data we have to go on, as Junk was demoted after failing to quiet Oakland’s typically silent bats. In Triple-A the rest of the year, he pitched to a 6.12 ERA and 5.10 FIP, making it unlikely he’d receive another nod in the majors this year. So I had to search elsewhere for a verdict on whether high sliders were truly effective in the majors. They certainly remain uncommon, with little change from last year to this year:

Using my Statcast-aided definition of high sliders, their usage has actually decreased from 18.0% last year to 17.3% this year, a statistically significant difference.

Besides usage, high sliders have differed crucially from their lower counterparts this season in terms of plate discipline and hitter outcomes. Revisiting Statcast’s strike zone from above, in addition to still defining high sliders as those in the highlighted portion, I defined middle sliders as those that landed in the three belt-high squares and low sliders as the rest.

Likely due to their uncommon nature, high sliders seem to sneak up on the hitter, leading to significantly fewer swings than all other slide pieces. Compared to the lowest sliders, high ones notch more called strikes, probably on account of the fewer swings against and because a lot of the lowest sliders are absolutely buried. High sliders do result in significantly fewer whiffs than the lowest ones, and not just because they result in fewer swings; their whiffs per swing number is also significantly lower. Their wOBACON (wOBA on contact) against is also meaningfully higher than that of the lowest sliders.

Despite the higher rate of called strikes for high sliders, I would award the point to lower sliders here. But this table is eye-opening when it comes to comparing high sliders to ones in the middle of the zone. Naturally, the ones in the middle nab more called strikes, but the two groups don’t differ meaningfully in swinging-strike rate (SwStr%), and high ones take the cake in terms of whiffs per swing and wOBACON. Perhaps it’s better to miss high with a slider than in the middle; maybe the hanging-est sliders aren’t the worst ones after all. And maybe they should even be intentionally hung depending on the situation.

Another way we can evaluate the effectiveness of high sliders is by looking at their proficiency at the individual level. While they are infrequent league-wide, this isn’t necessarily the case on a pitcher-by-pitcher basis. Below are this season’s leaders, among the 492 pitchers who have thrown at least 50 sliders, in high slider percentage (high sliders/total sliders). For reference, the average High% for this group was 17.4%.

Junk comes in at no. 3, and we know him. But what about the top two? Castellanos underwent Tommy John in early August; before that, his preferred slider target was at or near the top of the zone, and he threw more sliders high than either middle or low. He allowed the largest wOBACON against high sliders but also netted his best SwStr% and whiffs per swing with them.

Side-winding reliever Rogers is just below Castellanos on the leaderboard. He too throws more sliders high than either middle or low. In fact, high sliders make up more than a fifth of all pitches he throws. His high slide pieces allow a significantly lower wOBACON than his lowest ones, and while they have middling called strike numbers, they are tops in whiffs per swing.

Despite his large sample sizes, Rogers is probably not the best case study. His slider is the second-slowest on average in the league (interestingly, no. 1 is Hill, who ranks sixth in High% but has a more typical profile of results against his slider). It also ranks better than average in both vertical and horizontal movement. Overall, it has a weirder shape than most, especially given Rogers’ release point. Castellanos’ slider is above average in velocity but below average in both types of movement, even for slide pieces at similar velocities. What ties the two together, then? And going a step further, what ties together pitchers with high percentages of high sliders?

The answer might just be a lack of overall command. Over at The Athletic, Eno Sarris developed a metric called Location+, which evaluates a pitcher based solely on their pitch locations and the run-preventing value of said locations. Theoretically, the pitchers who are best at hitting their spots, or the pitchers with the best command, should have the highest Location+ numbers. Sarris also has a Command+ metric, but it did not meaningfully differ from Location+.

If you listen closely to the audio in the Janson Junk GIF above, you can hear the announcer say at the end, “I think he’s missed the target.” In lieu of Sports-Info-level catcher setup locations, Location+ is a worthy proxy for evaluating whether a pitcher is hitting his spots. And the proportion of a pitcher’s high sliders correlates negatively with Location+. In other words, pitchers who throw more high sliders tend to have worse Location+ numbers:

Let’s review. Burying sliders is the best way to elicit a swinging strike, but high sliders generate more called strikes than buried ones and still allow a lower wOBACON than belt-high slide pieces. There is pitcher-by-pitcher variation across these categories, possibly dependent on arm slot, but overall, high sliders seem to achieve their potency through their unexpectedness. A worthwhile experiment would be increasing their usage at the expense of belt-high sliders; only then would we be able to tease the value of a high slider apart from its element of surprise.

Statistics as of September 28.