March 19, 2012

Prospectus Hit and Run

AL Rotation Rumble

by Jay Jaffe

Printer-friendly

Two years ago, the Rangers made a bold gambit that helped end nearly a decade of rotation-driven futility, shifting reliever C.J. Wilson to the starting five and bringing former supplemental first-round draft pick Colby Lewis back from Japan. Both pitchers did what Ranger hurlers of recent vintage had not: miss bats. In 2010, the two pitchers combined for 366 K's in 405 innings, helping the Rangers jump from 12^th in the league in strikeouts to fourth. Helped by other upgrades—shortstop Elvis Andrus keyed a defensive turnaround—they won the AL pennant, and last year they repeated the feat.

When Wilson jumped to the division-rival Angels as a free agent in December, the Rangers undertook another pair of bold moves designed to maintain a commitment to missing bats: They promoted young closer Neftali Feliz to the rotation—a move they toyed with last spring—and spent big money to sign Japanese free agent Yu Darvish.

How well does the revamped unit stack up relative to the rest of the league? Last week, I dug through our Depth Chart-based PECOTA projections to examine each NL team's selection of starters. Today, I turn my attention to the AL. Again, I've got my eye on something that functions as an entertaining argument starter, an instructive inquiry into the forecast system's inner workings, and a premature playoff preview, tallying the best Terrific Tandems, Big Threes, Front Fours, Fab Fives, and Six-Packs for each AL team. Because a substantial selection of starts comes from outside a team's top five (going strictly by total turns)—last year the number was 27 out of 162, 17 percent—I took into consideration each team's top six starters. In many cases an alternate with fewer innings (a swingman, midseason callup, or late-healing frontliner) forecasts to be more valuable than a pitcher with a higher innings total. It's worth noting that some of the pitchers whose totals are included into the six-man mix are swingmen whose relief work is incorporated into the valuations, but the low innings and WARP totals generated by that bullpen duty shouldn't drive us too far off course.

Before turning to that, and with Darvish in mind, it's worth a look to see which teams did the most to boost their rotations using trades and free agent signings.

Perhaps somewhat surprisingly, the Yankees top the list, with their mid-January trade of top prospect Jesus Montero to the Mariners for Pineda and their parallel signing of Kuroda. Note that the former's WARP total is based on a very conservative projection of 148 innings, down from 171 as a rookie last year; he could gain another half win if he matches that mark. The latter is one of five Asian players to make this list, but the only one with a stateside track record.

Ex-Ranger Wilson has a slight edge on new Ranger Darvish, but mostly that's a matter of estimating the two pitchers' innings. The former, whose track record as a starter is all of two years long, is projected for 208, while the latter, who is spending his first year in a new country and a new league, is set at 176; if he approaches 200, their values will converge.

Lower down, amid the mixture of the raw (Molina) and the Cooked, the hustle of new Orioles GM Dan Duquette, who brought in a pair of Asian imports, shows up in a surprising fourth place, though it's more a question of quantity rather than quality. We also see that PECOTA isn't particularly sanguine on some less-experienced players such as Noesi and Peacock, who took big steps forward in 2011, nor is it confident in the ability of geezers Colon and Millwood to sustain comebacks.

With that out of the way, we turn to the master table, ranked by the combined WARP totals of each team's top six starters, and showing the combined WARPs of smaller subsets as well.

Terrific Tandems: The honors for the top one-two punch go to the Angels, but Wilson is neither the one nor the two. Instead, the prize goes to Jered Weaver (3.14 ERA, 3.9 WARP) and Dan Haren (3.07 ERA, 3.9 WARP), both coming off fine seasons—Weaver was runner-up in the Cy Young voting—and virtually interchangeable as far as value is concerned. Running second, with much more of an imbalance between the two, is the Tigers' twosome of Cy Young and MVP winner Justin Verlander (3.06 ERA, 5.3 WARP) and Max Scherzer (3.85 ERA, 2.2 WARP). The heavyweight combo of CC Sabathia (3.24 ERA, 4.8 WARP) and Pineda (3.80 ERA, 2.3 WARP) rank third, again with the caveat about the latter's low innings total. The only other tandem within one win of the top spot belongs to the Red Sox, whose Jon Lester (3.54 ERA, 3.6 WARP) and Josh Beckett (3.60 ERA, 3.3 WARP) are both held back by conservative inning estimates themselves, 186 and 183, respectively—not unwarranted, particularly in the latter case.

At the bottom of the pile, the Orioles can offer only Hammel (4.52 ERA, 0.7 WARP) and Wada (4.46 ERA, 0.6 WARP), and the Royals aren't much better, with Sanchez (4.45 ERA, 1.0 WARP) and Chen (4.62 ERA, 0.7 WARP). The latter is a chilly take on a pitcher who's been worth 3.5 WARP over the past two seasons during his KC renaissance, though both his 4.46 FIP and 4.55 Fair Run Average suggest he's been lucky to post that 3.96 ERA, and his 174-inning projection is generous considering he has totaled just 295 1/3 in that span.

Big Threes: Wilson's forecast (3.14 ERA, 3.5 WARP) puts the Angels more than one-and-a-half wins out in front of the field. Surprisingly, it's the Sox leapfrogging the Tigers and Yankees for second place here thanks to the sunny forecast for Daniel Bard (3.52 ERA, 2.8 WARP), one of several high-profile bullpen-to-rotation moves in progress this spring. The Rangers place third on the strength of Feliz (3.47 ERA, 3.2 WARP, in just 144 innings), Darvish (3.77 ERA, 3.1 WARP), and Lewis (3.87 ERA, 2.9 WARP).

The Yankees hold on in fourth thanks not to Kuroda but to Phil Hughes (3.87 ERA, 2.0 WARP), who thankfully appears to have come to camp in better shape this spring than last, helping to restore both lost velocity and lost promise. As with Pineda, his innings estimate is a very conservative one (141), but in light of his struggles over the past season and a half, perhaps more appropriate.

The Tigers slip to fifth; PECOTA just isn't much of a believer in Doug Fister (4.36 ERA, 1.1 WARP) due to a weak 4.9 K/9. He was at 6.1 last year, with a drastic split between the Mariners (5.5) and Tigers (7.1). Unsurprisingly, it's the Orioles (via Chen, 0.5 WARP) and Royals (via Luke Hochevar, 0.4 WARP) lagging behind the pack. The latter did have some late-season success to create some amount of optimism; after a 5.46 ERA, 1.2 HR/9 and 4.6 K/9 in the first half, he closed with a 3.52 ERA, 0.8 HR/9 and 7.7 K/9 in the second, a significant step forward befitting a former first overall pick.

Front Fours: The Angels maintain their spot atop the heap, but Ervin Santana's meager 1.1 WARP—again, a product of FIPs and FRAs in the mid-4s over the past three seasons—forecast is half that of Boston's Clay Buchholz, even with the latter estimated at just 156 innings, so Boston closes the gap considerably. The Yankees climb to third, but Kuroda's forecast (4.35 ERA, 1.6 WARP) is worth a quibble given his three-year FIP (3.57) and FRA (3.86). PECOTA sees the move from pitcher-friendly Dodger Stadium and the NL West to hitter-friendly Yankee Stadium and the AL East as rather drastic even for a ground-baller, but then again, this is a 37-year-old whose ground-ball rate declined and home-run rate rose considerably last year.

Via Derek Holland (4.59 ERA, 1.5 WARP), the Rangers hold onto fourth, while the White Sox pass the Tigers for fifth. Chicago’s best forecast comes from bullpen convert Chris Sale (3.4 WARP in just 120 innings), with Jake Peavy, John Danks, and Gavin Floyd bunched together between 2.5 and 2.2 WARP. Peavy (3.49 ERA, 2.5 WARP) has the best of those, albeit in just 144 innings, a level he hasn't come close to approaching since 2008; take the under on him before you do so on Sale.

Also creeping past the Tigers for sixth place are the Rays, who beyond David Price (3.4 WARP) and James Shields (2.5 WARP) are held back by particularly low innings estimates for Jeremy Hellickson (2.2 WARP in 156 innings) and Matt Moore (1.6 WARP in 138 innings). If the latter reaches last year's combined 174 innings (including the postseason) at a 3.59 ERA—even with an off-the-charts forecast for 10.5 K/9—he'd add another 0.4 WARP.

Fab Fives: Expanding to a full-blown rotation reveals the Angels' lack of back-end depth; via a sub-replacement level forecast for Jerome Williams (5.00 ERA, −0.7 WARP), they come back to the pack, with the Rangers (Matt Harrison, 0.6 WARP) leapfrogging the Red Sox (Felix Doubront, -0.9 WARP), assuming those two win their respective fifth-starter battles. The White Sox (Philip Humber, 0.6) and Rays (Jeff Niemann, 1.1 WARP) inch by the Yankees, because PECOTA sees Ivan Nova as a replacement-level proposition (5.07 ERA, 0.0 WARP), rather ungenerous given his improvement following his minor-league stint, from a 4.12 ERA via 0.9 HR/9, 3.6 BB/9, and 5.0 K/9 to a 3.18 ERA via 0.5 HR/9, 2.4 BB/9, and 5.7 K/9; all told, he was worth 1.6 WARP last year.

Six-Packs: It's here that the depth of the Rangers carries the day, as they turn a narrow lead into a substantial one on the strength of Alexi Ogando, who is forecast for a robust 2.3 WARP via 10 starts, 50 relief appearances, and 113 innings. Yes, I've introduced oranges into this apple-based comparison, because some of his value comes from his bullpen work (the breakdown is 1.3 starting, 1.0 relieving), but the same is true of many of the sixth men, and even some mid-rotation pitchers; Bard is projected for 23 starts out of 37 total appearances, Hughes for 24 out of 33. As it is, Ogando's projection ties him for 38^th among all pitchers and is the highest among pitchers classified as relievers (more than 50 percent of appearances out of the bullpen), 0.3 WARP ahead of the top pure reliever, Mariano Rivera.

Below the Rangers, just 0.6 WARP separates second place from sixth, with the Angels (Garrett Richards, −0.2 WARP), hanging on to edge the Rays (Wade Davis, 0.5 WARP) and Yankees (the just-unretired Andy Pettitte, 0.6 WARP on a 4.45 ERA in 90 innings gets the nod over Freddy Garcia, 0.2 WARP on a 4.82 ERA in 100 innings). The Red Sox (Aaron Cook, 0.1 WARP) and White Sox (Molina, 0.1) still hold on closer to the second-place Halos than the seventh-place Tigers.

One of the more disappointing forecasts belongs to the Blue Jays, for whom Brandon Morrow (3.98 ERA, 2.2 WARP) forecasts as the most productive pitcher, with Ricky Romero (4.37 ERA, 1.3 WARP) coming in well below the 3.60 ERA and 1.9 WARP he's averaged during his three-year career; his FRAs are tightly bunched from 4.61 to 4.67 so it makes more sense, but still, I'll take the under on the ERA and over on the WARP. Dustin McGowan's 1.3 WARP forecast comes on a wing and a prayer, as his 127 innings would be his most since 2007, several surgeries ago. Beyond Brett Cecil, there are particularly brutal forecasts for Henderson Alvarez (5.55 ERA, 1.0 WARP in 168 innings), who as recently as 2010 was putting up a 5.47 FRA in High-A, and Kyle Drabek (5.42 ERA, -0.4 WARP in 96 innings), who was torched for a 6.06 ERA last year.

As I did with the NL, here's a look at the raw rate stats for each Six-Pack:

Note that eight teams are forecast to strike out at least 7.0 hitters per nine, led by the Rays and Red Sox. Looking back at the NL batch, just two teams were, and quite frankly, that was a glitch, for which we apologize. According to Colin Wyers, pitchers' hitting (and all those extra strikeouts) wasn't being added into the totals, which didn't impact the WARP figures (which may have changed anyway due to updated innings forecasts), but did have an effect on the presented strikeout rates. Obviously, because of the difference between the designated hitter and the pitcher, the NL should hold an advantage; last year's Senior Circuit starters whiffed 6.9 men per nine, junior circuit starters 6.6. In the corrected batch, the NL had the edge, 7.15 to 7.01, but remember, that's just the top six starters on each team.

That's more like it, eh? The outcome of this lesson is that we have implemented safeguards to alert Colin and our stats crew when PECOTA produces a league rate in a category that might not stand up to scrutiny. As Ben Lindbergh described it:

The notification is an alarm in Colin's house that sounds like the one from “Lost” and can't be deactivated unless he slides down a pole into his subterranean computer cavern and keys a code into his calculator. Also, the alarm should sound on a frequency that only he can hear. We wouldn't want PECOTA problems to terrify his daughter. 

Back to the AL, the lesson is that the Rangers' depth gives them a slight leg up (once you subtract Ogando's relief work), but that it's anybody's ballgame among the other teams claiming the top rotations, so much so that the slightest injury or change in assumptions—particularly at the back of several rotations, where fifth starter candidates are still battling—could upset the order.