Maybe Contact Management Pitching Doesn’t Work in Arizona

One of the first evolutions in sabermetrics was the idea that pitchers didn’t control nearly as much as was widely believed for a very long time. Some things pitchers clearly do influence heavily: strikeouts, and walks. And with those “defense-independent pitching” principles, we got the first generation of “ERA estimators,” statistics that derive most of their meaning from the outcomes pitchers influence most. Fielding-Independent Pitching (“FIP”) is probably the most simple of those estimators. And while different authorities calculate FIP differently, the idea is pretty simple: take strikeout, walk and home run rates, and rely on those for a best guess of what ERA a pitcher truly earned.

Like everything in life involving human beings, the world is more nuanced than that. FIP was never an attempt to say anything different; it was a reaction to us not knowing how so much of the rest of the game worked, and a measuring stick to use as a baseline. There have always been so-called “FIP beaters,” pitchers who perform better than strikeout, walk and home run rates suggest, in a way that seems sustainable. Some pitchers earn their Earned Run Averages in other ways. And the most significant of those: contact management.

Contact managers can succeed by limiting the number of very hard hit balls in play (or over the fence). And while it would seem to be another way of saying the same thing, pitchers can also gain an advantage by inducing a particularly high rate of very softly hit balls. The Paul Goldschmidts of pitching do everything: get strikeouts, limit walks, limit home runs, limit hard hit balls, and get softly hit balls. But there aren’t too many of those, and to fill out a pitching staff, a team must make choices. Maybe there’s a pitcher who’s good at some things, and not others; the trick is identifying the skills that translate into some of the only things that tend to matter long term. A control pitcher who doesn’t manage contact but has a strong strikeout rate in the minors offers little if he can’t bring that strikeout rate to the majors.

In choosing between pitchers, it’s typical for a team to use a strategy akin to one used in the Rule 4 draft, in which the players selected are so far from the majors that things like “need” aren’t very relevant, and things like “profile” can change. One takes the best player available, usually. And if that means a contact manager is a little better than a strikeout artist who doesn’t limit hard hits, that’s the direction you take.

Most of the time. Like in the draft, where the D-backs took college players with the vast majority of their picks in 2015, the D-backs do have a type. It seems like Tony La Russa has had a ground ball vision for the team, and it almost doesn’t matter whether it was inspired by the success of pitchers like Brad Ziegler or years of managing from the dugout. It doesn’t matter whether La Russa may have eyed the same strategy if he had taken over in Minneapolis instead of Phoenix, or if a ground ball staff appeared to be the way to build a passable pitching crew in a mandated time frame. Through repertoire changes, preaching location down, and aggressively promoting pitchers who took to the new model, the D-backs did lean into that approach. And when it came time to take the Extra Step for the Contention Window, the D-backs set their sights on two pitchers who happened to be FIP-beaters in 2015: Zack Greinke, who beat his very good 2.76 FIP with a majors-best 1.66 ERA, and Shelby Miller, who detoured around a pedestrian 3.45 FIP with a 3.02 ERA, which ranked 14th among 78 qualified starters.

Neither pitcher has done well to start the season. Miller has had lots of things go wrong; he’s had his mechanics out of whack or at least undergoing a significant change, and when you add in a repertoire change that hasn’t worked out well, you get a pitcher who in four starts has a higher walk rate (6.75 BB/9) than strikeout rate (6.14 K/9), a mind-boggling home run rate (3.07 HR/9), and a 8.40 FIP every bit as bad as his 8.59 ERA. Miller’s case is complicated, and given all he’s done to change, it should surprise no one if he showed up looking dramatically better as soon as tonight. But Greinke’s case is maybe not so complicated. And for a pitcher who outperformed his FIP three years running in Los Angeles while evolving as a pitcher, a 4.22 FIP looks as temporary as his 6.16 ERA.

But is there something wrong with the model? Even a casual observer must have noticed that so many hitters have done better than expected in Arizona, Jean Segura a particularly obvious recent example. By the same token, the overwhelming majority of pitchers have not worked out as well as expected, at least not for very long. Since the beginning of 2006, 26 pitchers have thrown at least 200 innings for the D-backs. Many have not done well, and despite the assumptions used for FIP, we can’t assume that whatever this phenomenon is, it doesn’t affect strikeout, walk and home run rates. For what it’s worth, only Brad Ziegler, Juan Cruz, Josh Collmenter, Joe Saunders, Doug Davis, and Livan Hernandez outperformed their FIPs in any meaningful way — and for the last three of those, that had a lot to do with regrettable FIPs in the first place. Only Ziegler, Cruz, Collmenter and the irreplaceable Brandon Webb have had ERAs under 3.50 in this span.

The model appears to be ground balls, at least on the surface. For the D-backs in this span, Ziegler has maintained a 70.1% ground ball percentage — and Webb recorded a 64.1% GB% that was every bit as extreme, in the context of starters. Ground ball percentage is one variety of contact management, because in addition to avoiding (very-hard-hit) home runs, one doesn’t just get lots of ground balls at those rates — one gets lots of weak ground balls, the way fly ball pitcher Tyler Clippard will undoubtedly rack up popups. But ground ball (and fly ball) extremes are not the only kinds of contact management. Launch angle is great; preventing hitters from reliably “barrelling the bat” is also great. And it’s the latter thing that the D-backs acquired in Miller, and what they paid for in Greinke. It’s what has made Johnny Cueto great and Mike Leake decent. It’s what’s given Ian Kennedy a career, and made Livan Hernandez a legend, and what now gives Chase Anderson hope.

But maybe it doesn’t actually work here. At least in terms of starting pitchers, only Patrick Corbin and Robbie Ray stand for the D-backs contingent of surprising pitchers in recent seasons. In sizing up Scott Kazmir as a free agent option this winter, we made that connection in exploring the idea of whether the common thread was that they were left handed. But I think it might be something different. Both Corbin and Ray succeed in large part because of pitches not put in play. Frequent called strikes through fastball command, frequent foul balls early in counts, and strong whiff rates would be (nearly) immune to the trials and tribulations of pitching in Arizona. Call them Immune Strategies.

For Corbin, a strong called strike rate of 17.3% on the fourseam helped in 2016. But his sinker had Immune Strategies written all over it; it wasn’t necessarily a contact pitch at all, with a middle of the road 55% ground ball rate and one of the most excellent called strike rates on the pitch in the majors, 20.6%. That got him somewhere, but what made him a near-ace, it seems, was that wipeout slider. In his magical 2013 season, Corbin had a ridiculous 52.7% whiff rate on that pitch, the highest by far of the 93 pitchers who threw at least 200 of them. Only 14 of the 93 had a whiff rate as high as 40% — and only Corbin, Tyson Ross and Derek Holland had whiff rates on the pitch over 46.0%. We got flavor of the same thing in 2015, when Corbin still ranked 9th of 87 pitchers. This year, Corbin is moving away from his slider and toward a changeup — and as Jeff observed, hitters have not been missing the change. Corbin’s change also had a mediocre swing rate in 2013, and a decrease in slider usage has snowballed so far — Corbin’s whiff rate on pitch is way down to 32.6%. If this Immune Strategies business is correct, we may be in for a bumpier ride.

Ray’s 2015 whiff rate on his slider (ahem, Slurge) was better than Corbin’s has been this year: 36.4%. A 32% line drive rate cut down on its usefulness, but Ray’s Immune Strategies story isn’t a slider story anyway. The cornerstone to the working Immune Strategy: all of those foul balls. Sure, it extended at bats and made it hard to put away hitters with two strikes. But it got hitters to two strikes, and it did it in such a way that did not rely on batted balls. In 2015, when hitters swung at Ray’s sinker, they fouled it off 42.2% of the time. The foul rate on his fourseam was even higher: 49.5%. Sixteen pitchers bested that last mark, but only four of them combined that with a whiff rate as high as Ray’s 18.8%. For what it’s worth, Ray’s foul rate on his fourseam is way down this year at 33.8%, one of the lowest rates in the majors thus far — something he’s accomplished by pitching more out of the zone, it seems. By the same token, his whiff rate has also skyrocketed to 27.9% on the pitch. With iffy “control” and hard hit rates only one step away from alarming, Ray is no ace — and yet his reliance on playing keep away with opposing batters does seem like a sustainable strategy.

Contrast with all of the contact managers who have been brought to Arizona, Shelby Miller included. Jeremy Hellickson seems like a particularly telling example. Trevor Cahill‘s ground ball rate was very high, but nowhere near Brandon Webb’s — he, too, is a contact manager with limited success in the desert. I really can’t think of anyone except the most extreme relief examples. This is feeling correct. And it’s going to change how I look at things — including how Rubby De La Rosa‘s recent traction has coincided with a skyrocketing whiff rate on his slider.

***

There’s so much more to do here, and once we figure out why Statcast exit velocity and launch angle data is missing so much for away teams at Chase Field this season, we have a very helpful data set to work with. More coming, I’m sure. But as recently as last week, I was pushing hard toward contact management strategies for D-backs pitchers, and I am now much less certain. The pitching carousel the D-backs have used this April may not be a bad strategy moving forward; we want pitching success, and it looks like pitching success means not pitching to contact, which piles up pitch counts. If you could clone Robbie Ray to fill out a rotation right now, you’d have to seriously consider it. But as far as we know, that would mean cloning a bullpen full of Tyler Wagners and Josh Collmenters, as well.