Remaking Fireballers: Is Dave Stewart Building the Rotation in His Own Image?

It’s been just six months, but the new front office regime of Tony La Russa, Dave Stewart and De Jon Watson has dramatically altered the D-backs depth chart at starting pitcher. With Archie Bradley already on the verge of the majors, the team has added four more pitchers with high ceilings and spotty results: Robbie Ray, Allen Webster, Rubby De La Rosa, and Jeremy Hellickson. Young pitchers who are high risk but also high reward are valuable to any organization, but they may have been especially valuable to this team, which in the last two years has been almost completely starved for upside. The D-backs can’t afford to sign a Max Scherzer, but if they give themselves enough chances, they just might be able to develop a cheap one.

The new Ray/Webster/RDLR trio has been noted for their velocity, and that is absolutely a real similarity (and Bradley really is similar in that way, as well). They’re not all fireballers, exactly (Ray is more of a 92-93 mph guy), but that’s a common thread. And it would make sense to look at last season, and determine that gas was in short supply.

Because it was. The D-backs pitching staff averaged just 90.0 mph on fastballs as determined by PITCHf/x, ranking them 30th out of 30 pitching staffs. And it really wasn’t that close. The average of the 30 pitching staff’s averages was 91.8 mph, and that’s a bigger difference than it sounds like.*

To the extent that any of the Bradley/Ray/Webster/RDLR quartet pitch in the bigs in 2015, they’ll help to bump up the woeful staff velocity. Bradley recently averaged about 94 mph** in Arizona Fall League (two games had PITCHf/x). Ray averaged about 91 mph as a starter, but more like 94-95 mph as a reliever. Webster dipped from 94 mph in 2013 to 92 mph in 2014. And De La Rosa averaged 93-94 mph with the Red Sox.

Dave Stewart: Exceptionally Exceptional?

The beginning of Stewart’s career is eerily similar to that of Rubby De La Rosa. Like RDLR many years later, Stew started his pro career as a well-regarded Dodgers prospect, shuttling between the rotation and the bullpen for his first two teams and lighting up the radar gun with a fastball that was often surprisingly hittable — especially when he pitched as a starter.

Things had started to look bleak for Stewart when the fastball didn’t seem to work even in relief for the Rangers, and a tenure with the Phillies interrupted by bone chips in his elbow ended with his release. But he was picked up by the Athletics on a minor league deal in his age 29 season in 1986, and things immediately started to look up as Stew pitched to a 3.74 ERA in 29 games pitched, including 17 starts.

From 1987 to 1990, Stew rattled off excellent seasons in which he started 36 or 37 games, and won at least 20 of them. In that four-year span, he totaled 1,067.2 innings (!) with a 84-45 record and a 3.18 ERA. That’s great by any standard, but the checkered stat sheet Stew had in his 20s makes it much more remarkable.

Writing for the Washington Post in 1988, David Justice pointed to Stewart’s forkball as a possible cause for his success. In 1985 with the Phillies, Stewart’s manager, Doug Rader, was not in favor of Stew working on it in major league games. From that Justice piece:

Stewart said Rader called his forkball a ridiculous pitch and ordered him to put it in his pocket.

But it was the development of that forkball that a lot of baseball people credit with turning Stewart’s career around. Dave Duncan, Oakland’s pitching coach, noted that Stewart had terrific velocity, but that his fastball was straight as an arrow.

“To tell him to stick with the fastball, that’s one of those macho National League attitudes,” Duncan said. “Over there, they see a guy throw 94 m.p.h., and that’s all they think he should throw.

“The thing was, Dave didn’t have great control or movement just using that pitch. The forkball gave him an off-speed pitch he could control. He can throw that thing for strikes almost whenever he wants.”

Scouting reports told opponents to wait for the fastball, which was hard but true. Yet, soon after Stewart returned to the majors, he suddenly had a forkball that also was hard, but dipped sharply as it approached home plate.

In the same piece, Stewart also gave credit for his turnaround to a change in management, saying of Tony La Russa, “I’ve got a manager who respects my ability.”

La Russa and Duncan. Stewart, a man called “Smoke.” Maybe it was that Stewart was “one of the most intense competitors” that La Russa had ever seen. Maybe it was a National League philosophy and then elbow issues that had been holding Stewart back. But the scene is set; Stewart knows personally that this can be done, and that La Russa and Duncan also know that, and can even help make that happen.

If the Template is Followed, Then What?

If Stewart’s history has helped guide these offseason transactions (and there isn’t, necessarily, a reason that it shouldn’t), we’re left with questions. Has this been deliberate? Can it work? And what would it look like?

If the idea was that Stewarts can be made, then yes, what has happened thus far in this FO-by-committee’s time is what we’d expect that to look like. It’s very possible. But as noted above, for any team and especially the D-backs, adding high risk, high reward ingredients into a new magic potion is just a flat-out good idea anyway. It’s been pitched to us as an attempt to get back a lot of depth in that part of the roster, but this is a much better kind of depth than the kind the team had just two years ago. Seven or eight #4 or #5 starter types (a la Randall Delgado and David Holmberg) still makes for a staff of #4 or #5 starters. Seven or eight guys who might be #2s or washouts has the potential to make for a much more competitive staff, especially with some safety nets like Delgado and Chase Anderson and Vidal Nuno already in house.

There’s a logistics problem with having so many dicey candidates: they can’t all pitch in the rotation at the same time. There really is no scenario in which all of the risk/reward set of Bradley, Ray, Webster and RDLR are all in the rotation, even in May. But the D-backs know that. And that makes one wonder if the plan isn’t to play as many hands as possible, but to stack the deck.

Return to RDLR as our best example. The mostly-starter version of RDLR’s rebuilt arm that threw for the Red Sox relied a little bit less on a four-seam fastball — 55%-60% of the time, as opposed to when RDLR was with the Dodgers throwing it 72% of the time. His main secondary pitch has been a changeup, which has rated below average for his whole career to date (0.84 runs worse than average per 100 pitches, per PITCHf/x Pitch Values/100). And the slider, which he threw more frequently (12%-20%) with Boston? Very poor results there, as it’s been completely punished especially when he’s thrown it as a starter; in 2014, opponents had a .674 SLG against it.

But is RDLR an unfinished canvas, the way Stewart was? Now with Arizona, will Duncan coax him to air out something new, and see what happens? Heck, in RDLR’s case, the tweak might even be the very same. Stewart learned his “forkball” (we’ll call it a splitter) from some guy named Sandy Koufax. Maybe it’s Stewart’s turn to pay that grip forward.

Spring training absolutely will be a free for all, with some extra split-squad games and few practice mounds without occupants. But maybe the D-backs are gambling on making tweaks with Ray, Webster and RDLR before they ever step on the mound at Chase Field. For Webster, maybe it would be getting him to narrow his repertoire if there’s a pitch he doesn’t have as much confidence in (also known as a “total Duncan move”). Ray had already turned toward a slider, but maybe he’ll be persuaded to push his four-seam rate down below 60%, maybe in favor of a cutter (also known as a “classic Duncan move”). And considering that all three have had middle-of-the-road batted ball spreads, maybe a spell will be cast that will move any of the three toward ground balls (also known as a “Dave Duncan Effect”).

The rotation mix has two Chosen Ones in Josh Collmenter and Jeremy Hellickson (the latter of whom might even still fit this model we’re talking about). The rest of the rotation will be filled through some kind of spring training meritocracy — but it could be that we’ll be seeing some serious changes not just among pitchers, but for those pitchers. Hey, when things really aren’t working, change — even indiscriminate change — can be a good thing. It’s also going to be a blast to watch.

*We use standard deviation to get a feel for how much of an outlier an outlier might be, and the D-backs’s 90.0 mph average fastball velocity is 2.70 standard deviations from the mean. That’s a lot like saying that if there were 200 MLB teams instead of 30 and these fastball values followed the same distribution, the D-backs probably still would have had the slowest average fastball velocity in the league. Yikes! That figure was dragged down by a half season of 85 mph Bronson Arroyo fastballs, and in this figure, Josh Collmenter’s cutter (he doesn’t throw a four-seam) was included by PITCHf/x in the fastball bucket.

**Above in comparing pitching staffs, I used PITCHf/x numbers as reported at FanGraphs. Around here, we tend to use Brooks Baseball when possible, as those pitches aren’t sorted only by an algorithm; Pitch Info (a/k/a Harry Pavlidis) make sure pitches are classified the way they should be. That leads to some differences, as with Collmenter’s cutter. Another difference, however, is how velocity gets reported; in this piece, I’ve kept those readings as reported at FanGraphs. But those are calculated at the plate, whereas Brooks Baseball reports a velocity extrapolated from that information to show the speed out of the pitcher’s hand. That’s one reason why these numbers may look lower than you’d expect. Also, this forced me to guess with Bradley (hence the asterisk) — Brooks has Bradley’s fastball at two AFL parks averaging at 94.8 or so (but touching 97, for what it’s worth).