The Superlative Kyle Hendricks

You know it’s almost time for baseball season when all of the major projection systems forecast Kyle Hendricks‘ ERA one run per nine innings too high.

As much as this sounds like a knock on those who develop projections, it’s not. What Jared Cross (Steamer), Dan Szymborski (ZiPS), Derek Carty (THE BAT), and the folks at Baseball Prospectus (PECOTA) do is no small feat. If I weren’t too cowardly to even try to create my own projection system, I would be too stupid to design one that is half as effective as theirs. Glass houses and all that.

That said, I am just smart enough to know that projected ERAs ranging from 3.84 to 4.42 for Hendricks, who boasts a career ERA of 3.12 and has never finished a season with an ERA above 3.46 (except that dastardly 3.95 ERA in 2015), are too high. It’s easy to poke holes in the obvious outliers, but projections succeed by describing and then predicting the talents of most pitchers, not the ones whose talents deviate dramatically from expectation. Hendricks is every projection system’s known blind spot.

It’s not just projections that struggle with Hendricks, either. We, the sabermetric community, frequently use ERA estimators as shorthand to characterize a pitcher’s talent level. If you frequent FanGraphs, you’re familiar with Fielding Independent Pitching (FIP), expected FIP (xFIP), and Skill-Interactive ERA (SIERA). By virtue of how they’re constructed, each metric makes assumptions about the skills a pitcher theoretically “owns”:

• FIP: strikeouts, walks, and home runs allowed
• xFIP: strikeouts, walks, and fly balls induced
• SIERA: a complicated combination of strikeouts, walks, net groundballs (groundballs minus fly balls), and their squared terms and interactions with one another

While each estimator features a batted ball component, they focus on trajectory (launch angle), not on authority (exit velocity). This is a fair assumption, frankly. I have illustrated how a pitcher can influence hitter launch angle, operating under the assumption they bear little to no influence over hitter exit velocity. It’s not quite that bleak; certified baseball genius Rob Arthur found that the average pitcher’s effect on a baseball’s exit velocity: roughly five parts hitter, one part pitcher.

Still, that’s not much influence. For many pitchers, over the span of their careers and especially within a single season, that influence appears indistinguishable from random variance or “luck” such that it’s justified for ERA estimators to lack an exit velocity component. Again, generally this is fine — but it’s specifically not for Hendricks who, since he debuted, possesses the ERA with the largest discrepancy between its xFIP and SIERA, and the second-largest discrepancy with its FIP, all among pitchers who have compiled at least 1,000 innings since the start of 2014.

In 2017, Baseball Prospectus‘ Jeff Long, Jonathan Judge, and Harry Pavlidis employed modeling techniques way above my pay grade to investigate Hendricks’ propensity to induce called strikes, swinging strikes, and weak contact. Do all pitchers rely on those things to succeed? Yeah, they do. But do all pitchers rely on those things in precisely the way Hendricks does? No! They absolutely do not.

Weak Contact

In his piece, Arthur states:

[…] the pitcher’s effect is not negligible. While the best batters increase batted ball velocity by as much as 7-8 mph, the best pitchers suppress it by 1.5 mph compared with the average pitcher.

That’s a relief, because using altogether less-rigorous mathematics I came to similar conclusions that conveniently advance my pro-Hendricks agenda. Using Statcast data for the 2017 through 2020 seasons, I calculated the average exit velocity (EV) for every hitter each season. Then, I calculated the average EV for every pitcher-versus-hitter matchup. Afterward, I “backed out” the pitcher impact by calculating each hitter’s average EV absent the EV allowed by that pitcher. For example, 2020 Mike Trout would now sport a different average EV against 2020 Shane Bieber than he would against 2020 Aaron Civale. Let’s call this the “pitcher-neutral average EV.” Lastly, I took an average of the difference between the measured EV of every batted ball event (BBE) and the hitter-specific pitcher-neutral average EV. (Woof, that’s a mouthful.)

Ultimately, this approach makes some attempt of (1) identifying the “true” EV of every opposing hitter a pitcher faces and then (2) determining how much the pitcher makes the opposing hitter deviate from his “true” EV on average. It’s my way of crudely bypassing a complicated model I can’t assemble, let alone interpret, that would require some manner of random and/or fixed effects to control for things that need controlling.

It’s worth acknowledging that this approach imperfectly parses pitcher talent from other effects (hitter skill, luck/variance, etc.). The approach doesn’t explain how it happened — just that it happened, although these results align with findings by Arthur and with my other work, as I’ll discuss in a bit. Having admitted this, know that I’ve approached this analysis with the long-term in mind. Nearly anything can happen to nearly anyone in a single season. Only over time do we see distinct skills-based trends start to emerge, signals emerging from the noise.

It helps that my results pass the smell test. Among active starting pitchers who incurred at least 1,000 BBE since the start of 2017, here are the top 15 in what I’ll call “pitcher influence on EV” and how their ERAs differ from their FIPs, xFIPs, and SIERAs during that span:

All things considered, this is a who’s who of high-achieving pitchers. Even the surprising names here hold their own: Sánchez compiled a 3.39 ERA in roughly 300 innings during the 2018 and ’19 seasons, and Minor owns an impressive 3.84 ERA since the start of ’17 despite last year’s sprint-season dud. Maeda, long a pitcher whose parts exceeded their whole, was always good enough but pulled it together last year with a change of scenery. Only Syndergaard defies the trend; his oddly imbalanced ratios (high BABIP, low home run-to-fly ball ratio) probably merit their own investigation.

For funsies, here’s the bottom 15:

Perennial under-performers Ray, Márquez, Archer, and Pivetta join half a dozen other pitchers with one foot out the majors’ door, several of whom I omitted because they did not pitch in 2020 but were signed to minor league contracts or something like that. Bumgarner continues to defy expectation, but his peripherals have backslid tremendously since his dirt bike accident. The reigning American League Cy Young winner might also seem like an odd inclusion, but Bieber’s approach prior to 2020 left him more vulnerable to hard contact than 2019’s 3.28 ERA let on (Statcast suggests he outperformed his ERA by 59 points).

I’m getting distracted, though. The beneficial (albeit daunting) part of working with Statcast data is that we can conduct this same analysis at the pitch-specific level. Hendricks grades out handsomely in aggregate, but how does this analysis look for individual pitches? Which pitchers’ pitches exert the most influence over hitter EV?

I fully expected Hendricks to show up on this table, but I did not expect him to utterly dominate like so:

Per my crude research, this amounts to a comprehensive list of pitches (1) thrown by starting pitchers; (2) that have incurred at least 500 BBE since the start of 2017; and (3) that exhibit negative influence upon a hitter’s EV. Keep in mind this table is biased in favor of both durable pitchers and oft-used pitches. Fragile arms and overstocked arsenals might be unfairly excluded. But there’s Hendricks’ changeup, standing head-and-shoulders above the field.

In another brilliant article (although I may be biased by its subject and content) aptly titled “Location, location, location: Kyle Hendricks is winning with control,” Arthur further elucidates how Hendricks suppresses hitter contact quality: with pinpoint command — and also below-average velocity. Blissfully unaware of Arthur’s work prior to conducting my own, I came to similar conclusions (not coincidentally using Hendricks as my inspiration), finding that pitch location (specifically) and lower velocity (generally) explains a significant chunk of the variance in weighted on-base average on contact (wOBAcon). Hendricks grades out exceptionally well there, and better yet with controls for pitch velocity.

It’s fitting, then, that not only does this short list feature two of Hendricks’ pitches, with one of them leading the pack outright, but also that an expanded view of this table, with the qualification threshold relaxed from 500 BBE to 250 BBE, features off-speed pitches almost exclusively:

Moreover, the bottom of the list, which I originally truncated to save space, is populated exclusively with fastballs:

Chris Archer? You mean the Chris Archer who once looked like an ace but broke in 2016, suddenly allowing uncommonly high BABIPs and ratios of home runs to fly balls? The one who sits at the wrong end of the ERA-minus-FIP leaderboard that Hendricks leads? That Chris Archer?

It’s one thing to suppress exit velocity and to do so skillfully, reliably. It’s another thing to do it routinely, deliberately. Hendricks not only silences opposing hitters (to the furthest possible extent that a hitter can be silenced) with his changeup. Per the “Pitch Type EV Influence Leaders (2017-20)” table above, he also induces contact via his changeups more often than any other above-average EV-influence pitch except for Trevor Williams’ four-seamer (which is nearly average in its influence) and Hendricks’ own sinker/two-seamer.

In other, less-verbose words, Hendricks zigs while everyone else zags: whereas most pitchers seek to avoid contact all together (adhering to the fielding-independent theories that underpin the Big Three ERA estimators), Hendricks practically welcomes it.

You can’t force a hitter to swing, though, so it makes one wonder exactly how, among 209 individual pitches thrown at least 2,000 times since the start of 2017, Hendricks’ changeup leads all pitches in swing rate (Swing%). Fortunately, hitter swing patterns are fairly predictable by ball-strike count on average (I realize hitters exhibit varying degrees of aggression/passivity overall), especially in those where the pitcher is ahead — a situation that, as Arthur notes and Statcast confirms, Hendricks finds himself frequently. (Per my numbers, Hendricks ranks in the 93rd percentile of “true” first-strike rate, which ignores balls in play, focusing only on pitches that get to 1-0 and 0-1 counts.)

What’s more, among that same sample of pitches, only three are thrown in the “heart” of the strike zone (defined by Tom Tango and his team in their swing/take work) less often than Hendricks’ changeup. Put two and two together, and Hendricks gets into counts where hitters are more likely to swing and induces weak contact by commanding his low-velocity changeup everywhere but the heart where hitters thrive, thereby suppressing hitters’ EVs by not only an elite magnitude but also at an elite frequency.

Called Strikes

Incidentally, coercing an inordinate number of swings comes with the territory — namely, not stealing many called strikes. Hendricks’ changeup doesn’t! In fact, it features the second-lowest called strike rate since the start of 2017, behind only Masahiro Tanaka’s splitter.

What’s odd in this context (albeit not generally, given what we know about Hendricks and how he thrives) is that Hendricks is the king of called strikes. Among 98 pitchers who have thrown at least 6,000 pitches since the start of 2017, no one boasts a higher called strike rate (20.3%). Clearly, those called strikes don’t come from his changeup, so they must come from somewhere else.

Enter Hendricks’ two fastballs. Unlike his changeup, he weaponizes them in counts with low swing probabilities, resulting in absurdly low swing rates (fifth percentile for his four-seamer; for his sinker, 10th percentile) that, of course, produce elite called strike rates (95th and 99th percentile, respectively). It’s not just hitters’ tendencies, though. Hendricks commands each with painstaking precision. Both fastballs rank in the 98th percentile with regard to how often each finds the Statcast-defined “shadow” of the zone (aka the space surrounding the heart and encompassing the areas just inside and outside the edges of the zone). No one lives in the shadows more than Hendricks (45.6%), which, if nothing else, sounds cool as hell:

Which is all a long way of saying: Yes, Hendricks is a called-strike machine. Yes, he’s a master of inducing weak contact. Yes, his command is virtually unparalleled among his contemporaries. But it’s the diametrically opposed use of (and uses for) his pitches — a changeup that is built exclusively for swings and suppresses hard contact; two fastballs that move in dramatically different ways that steal strikes and yield advantageous counts — that get me. Every single feature of his arsenal is the best, or close to it, at what it is supposed to accomplish.

The way Hendricks dispatches his pitches, one might go so far as to call it predictable. Given how extreme everything Hendricks does is, I think it’s a fair characterization! Yet his location-based aptitude (an under-appreciated skill) combined with below average velocity (a characteristic too often perceived as a flaw) makes it such that a hitter who rightly anticipating a fastball in an early count or a changeup in a late count is tasked with squaring up a tough pitch regardless and remains off-balance.

Per Long et al. (it was Long who forwarded me this article after I wrote about vertical approach angle and pitch location recently), Hendricks also tunnels his pitches like a son-of-a-gun. And we’re learning more and more about pitch movement, spin mirroring, and how everything interacts via Statcast’s new spin direction leaderboard alongside more complicated, physics-related topics like Magnus force and seam-shifted wake. I’m eager to learn more about how the unconventional Hendricks defies every conventional expectation designed for the conventional pitcher, which he surely is not.

In the meantime, I can revel in Hendricks’ superlatives — how he not only is the best at earning called strikes or finding the shadows of the strike zone (while, at the same time, avoiding the heart of the zone) or suppressing hitter contact quality but also achieves those superlative outcomes with superlative pitches that, individually, are the best at what they do. Long et al. invoke Greg Maddux’s name, although they note Hendricks is reluctant to accept such praise. I am but one man/idiot, but I think he should at least consider it!

Unfortunately, while I celebrate Hendricks and his status as baseball’s contact management unicorn, I also lament how his value is defined and perceived. I’m a company man, but FanGraphs’ wins above replacement (WAR), just like FIP which underpins it, inadequately captures Hendricks’ value.

Compiling 20.8 WAR in 174 starts (roughly 3.6 WAR per 30 starts) is nothing to sneeze at — this may surprise you, but Hendricks ranks 15th in WAR since his debut. Yet a version of WAR that accepts his uncommonly good ERA as skills-based, rather than dismissing it through FIP as primarily luck-based, might net Hendricks another five or six wins, based on my back-of-the-napkin math. Being revisionist history’s ninth-best pitcher of his era won’t earn Hendricks a trip to the Hall of Fame, but having skills that could age very, very well over the next decade (exceptional command aided by already-below-average velocity) might very well bring him into the conversation.

To wit: Maddux sported an indexed ERA, or ERA- (which adjusts ERA annually for run-scoring context, where lower is better), of 86 through his first seven major league seasons and finished his career with a 76 ERA-. Hendricks currently sports a 76 ERA-. Hey, pal, it’s just what the numbers say.

This is not uncharted territory; much smarter people than me have explored the concept of contact management generally and as it pertains to Hendricks specifically. Let this serve as testament to Hendricks’ stature as one of the game’s best pitchers. He doesn’t pile up strikeouts. He doesn’t flash game-changing velocity or wipeout stuff. He’s never going to be GIF’ed by Rob Friedman.¹ But he succeeds despite it all using an approach that weaponizes certain pitches in very specific ways to produce elite, practically exaggerated outcomes — an approach that, very probably, only he (and another very famous pitcher, his name escapes me) can capably pull off.

Footnotes

¹ I know Friedman has GIF’ed Hendricks before. To bring everything full circle, here’s a delicious GIF of that shadowy changeup:

Kyle Hendricks, Wicked Changeup…and is impressed with his own filth. pic.twitter.com/sXuMe126Dj

— Rob Friedman (@PitchingNinja) August 12, 2020

This piece originally said Hendricks’ highest single-season ERA was 3.46 in 2019. It was actually 3.95 in 2015.