What Are We Missing About Ian Anderson?

How is everyone dealing with the, uh, complete stoppage of major league baseball activity? Each person has a different method, I assume. As for me, I’m consuming both less and more baseball, strange as that might sound. The lockout has led me to invest energy into other hobbies, but baseball-related articles, podcasts, and videos have also been my lifeline in these trying times.

One podcast I owe much thanks to is Rates and Barrels, hosted by Eno Sarris and Derek Van Riper over at The Athletic. They’ve provided inspiration in the past, and I’m about to piggyback off of them again. In a recent episode about pitchers with bounce-back potential, Eno mentioned a quirk about Braves starter Ian Anderson that piqued my interest:

“The most interesting thing about Ian Anderson is he might be doing something with his changeup that my model can’t capture… it’s getting to the point where he’s demonstrated results on his changeup that far outweigh the grades these pitching models put on it.”

The model he’s referring to is Stuff+, which was developed in tandem with Max Bay and uses several variables to evaluate the quality of a certain pitch, or, in the aggregate, a pitcher’s entire arsenal. On the top of the Stuff+ leaderboards are names one would expect: Jacob deGrom’s four-seam fastball is otherworldly, Corbin Burnes’ cutter is unmatched, and Tyler Glasnow’s curve is as beautiful as his luscious hair.

But no model is perfect. Or I should say, no model is capable of capturing all the nuances of real life, and a few players are bound to slip through the cracks. And according to Eno, Anderson may be one of them. Despite a high whiff rate (34.9%) and good results on contact (.291 wOBA), Stuff+ is decidedly not a fan of his changeup. So what should we believe? Is Anderson a bonafide star, or does his seemingly underwhelming stuff suggest that what he’s achieved thus far is an illusion?

It’s time to start digging. During the podcast, Eno also surmised that Anderson’s unique release point might be contributing to his success with the changeup. Indeed, few pitchers throw a changeup from a similarly high and vertical arm slot, with elite extension to boot. But variables like release height and extension are baked into the Stuff+ model, and while a pitcher like Anderson is rare, the existence of those few others means the answer lies elsewhere. In fact, there’s a renowned righty whose changeup is nearly identical to that of Anderson’s. His name is Lucas Giolito:

You could have told me those are the same pitches and I wouldn’t have batted an eye. From spin direction to release point to movement, Giolito and Anderson share many of the same characteristics on their respective changeups. But look at those Stuff+ grades! Giolito’s changeup grades out as an above-average offering, while Anderson’s is below average.

What’s causing that discrepancy? There’s another important variable I omitted on purpose, and that’s velocity. Last season, Anderson averaged a whopping 88.1 mph on his changeup – Giolito, on the other hand, clocked in at just 81.5 mph. As one could expect, speed matters. Long ago, Harry Pavlidis of Baseball Prospectus found that speedier changeups tended to induce more grounders. Case in point: Anderson’s changeup averaged a launch angle of two degrees last season – far lower than Giolito, whose own averaged a lofty 19 degrees.

But Pavlidis’ findings also point us to the value of velocity differential. The bigger the difference between a pitcher’s fastball and changeup velocity, the more whiffs he accumulated on his changeup, the article states. Unsurprisingly, velocity differential is an essential component of the Stuff+ model. Between Giolito’s four-seamer and changeup, there exists a gap of 12.3 mph. For Anderson, it’s less extreme: 6.5 mph.

Yet, it’s not as if Anderson’s changeup was that inferior in terms of whiff rate (34.9% vs. 35.4%) or results on contact (.291 vs. .281 wOBA). Sure, the relationship between velocity differential and changeup whiff rate can vary some depending on the pitcher, but it still seems like we’re leaving out… something. So here’s a theory. It revolves around a concept whose popularity has waned in recent years but that is nevertheless relevant, though maybe not to the extent we thought previously. The t-word! Tunneling.

First, let’s actually see some pitches in action. Here’s Anderson with his changeup, which elicits a swing-and-miss from Bryce Harper:

Next is Giolito – same pitch type, same location, same result. But can you spot the difference?

Besides the difference in velocity, consider the general arc of each pitch. Anderson’s changeup veers downward gradually, whereas Giolito’s changeup fades suddenly, ending up at the bottom of the zone. The numbers corroborate this observation, too: After adjusting for pitch height, Anderson averaged a -6.8 degree approach angle on his changeup; Giolito averaged a -8.0 degree approach angle. The relative “flatness” of the former changeup is reflected in what we observed in the above GIFs. And no, this isn’t because one pitch generated more vertical movement than the other – according to Pitch Info’s measurements, the difference between the two is practically a rounding error (1.32 vs. 1.38 pfx_z). Identical destinations, alternate paths.

What does this all add up to? My hunch is that due to its shallower entry into the strike zone, the changeup Anderson throws resembles his four-seamer quite strikingly until, wait, it doesn’t! Cue the awkward swing and miss. This makes up for the absence of a large velocity differential – in fact, the zip of the changeup might be what keeps it line with the fastball, in conjunction with a high spin rate. There’s not much separation as a result, but maybe that’s the point.

Okay, here are some more GIFs, because they’re fun. In this at-bat, Anderson starts off with two curveballs – opposing hitter David Peralta took the first one, then fouled off the next. The third pitch is a fastball located down in the zone:

That’s not where Travis d’Arnaud had originally planned for it, but it gets his pitcher to two strikes – time to initiate the kill sequence. Anderson goes for his signature changeup and ends up way, way too low:

Not to worry, though, because the following pitch is an absolute blast:

Remember that low fastball from earlier? Well, the changeup ends up just beneath it, catching Peralta, who had the fastball in mind, completely off-guard. But a model like Stuff+ might look upon this with disapproving eyes, and perhaps that’s an inevitable oversight. Though teams have access to high-speed cameras that can quantify the benefits of pitch tunneling, we in the public are shut out from such information. It’s possible certain players are caught in a blind spot, Anderson included.

But ultimately, this is an educated guess. It could be that another attribute, not tunneling, explains Anderson’s success. Or worse, that there’s nothing particularly special about Anderson. There are indeed lingering questions regarding his performance – we should note the dip in strikeout rate last season – and he’s likely not an ace, but there’s lots to like about the right-hander. He wields three solid pitches, throws from a deceptive arm slot, and has room for growth. A changeup that goes beyond what is expected of it only adds to the intrigue.