It isn’t every year that ball clubs with undeniably shaky pitching rotations approach October as odds-on favorites to make the World Series. Yet this is where the Yankees and Cardinals find themselves with two weeks remaining before the start of the postseason. Before the two best teams in the majors book their tickets to each other’s city, they will face some difficult decisions in assembling their postseason rotations.

The two clubs’ quandaries are distinct. The Cardinals’ two most reliable starters in previous seasons, Matt Morris and Woody Williams, haven’t been as effective in 2004, while newcomers Chris Carpenter, Jason Marquis, and Jeff Suppan have thrived. Still, there’s the feeling that the latter three are doing it with smoke and mirrors and might not remain effective in October; Carpenter’s strained biceps problem raises further doubt.

As for the Yankees, they have what should be an imposing collection of proven arms in the rotation. But almost all of them have underachieved or lingered in various states of disrepair this year. Joe Torre has generally preferred to go with a four-man rotation in recent years, but without the reliable trio of Roger Clemens, Andy Pettitte, and David Wells, the Yankee skipper may be wise to adopt a three-man rotation this year. Indeed, only Orlando Hernandez, Mike Mussina, and John Lieber have shown themselves worthy of playoff starts in recent weeks.

Let’s break down the Yankees’ pool of rotation contenders for the ALDS:

The three middle statistics in the chart above are commonly known as “peripheral indicators.” These numbers show what a pitcher is doing independent of the defense behind him and however luck may favor him on any given day. The runs/game number simply indicates how many runs a pitcher gives up per nine innings; it’s a far more informative statistic than ERA, which is too accommodating toward the pitcher following a fielding error.

Although Hernandez came out of nowhere this season and could regress to the lesser, more injury-prone El Duque of old, there’s still no sensible way to argue against his being in the post-season rotation. His peripheral stats don’t necessarily justify his sparkling 2.51 runs allowed per game, but he’s easily been the most effective Yankee starter this season.

Mussina also belongs. For much of the season he’s been undermined by injuries, misfortune, and a spotty team defense (as evidenced by his opponents’ .320 batting average on balls in play). Still, his strong K/BB ratio and recent run of success (1.20 R/G and 11.0 K/BB ratio in September) mean he most assuredly needs to be the no. 2 man.

As for the third ALDS starter, what’s most apparent is that it shouldn’t be Brown. He has the lowest K/BB and second lowest K/9 of the Yankee starters, and he’s certainly an unknown quantity for the immediate future given his injury status. Brown’s broken hand may be sufficiently healed for him to make one final regular-season start, but that’s not much of a tune-up for the Division Series, where the margin for error is small. Because of his health concerns, probable rustiness coming off the disabled list, and middling performance this season, Brown shouldn’t be in the rotation.

The Yankees will no doubt be tempted to rush Brown back into the fray, but that figures to be a mistake. The bold decision – and the one the Yankees should make – is to give Lieber a start in the Division Series. He has immaculate control, superb command (the sixth-best K/BB ratio in the majors), and he does a better job of keeping the ball in the park than does Vazquez. He’s also been about as dominant as Mussina in the month of September (23 K, 2 BB in 27.1 IP).

Vazquez, meanwhile, has been wildly inconsistent this season. While 16 of his 30 starts have been quality starts (defined as surrendering three runs or fewer and pitching at least six innings), he’s also had six “disaster starts,” in which he’s given up at least a run per inning pitched. Vazquez hasn’t logged consecutive quality starts since July. Torre will probably end up sending him to the mound during the postseason, but it’s a risky move.

As for Lieber, pitchers with 5.0 K/BB ratios or better are rare finds, and the Yankees shouldn’t make the mistake of squandering him in long relief. He deserves to join Hernandez and Mussina in the post-season rotation.

Now let’s take a look at the Cardinals’ potential starters.

Carpenter is the no-brainer here. He has a stellar strikeout-to-walk ratio of 4.0, a solid strikeout rate, and the lowest runs-per-game mark of anyone in the rotation. If merit carries the day, he’ll be the Game 1 starter in the NLDS, health permitting (he’s due back from injury for his last start of the regular season, September 30).

Williams should probably be the no. 2 man. He has a solid K/BB and K/9 and leads the field in lowest home run rate. He also brings a veteran presence and a history of strong postseason performances.

The temptation for St. Louis manager Tony La Russa will be to install Morris, who has a record of mostly impressive performances in the playoffs, as a starter in the post-season. This would be a serious mistake. Morris has a grisly home run rate (he’s on pace to give up 35 homers in just over 200 innings), and challenges Vazquez in terms of inconsistency. Thus, Marquis should be the Cardinals’ third starter.

Of course, given the shaky nature of both teams’ pitching staffs, a larger question needs to be asked: Will either the Yankees or Cardinals even make it to the World Series?

This article was provided by Baseball Prospectus.The Sun will run exclusive content from Baseball Prospectus throughout the 2004 season. For more state-of-the-art baseball content, visit www.baseballprospectus.com.