Using Support-Neutral Winning Percentage to describe which team’s starters have done the most to produce victory for their teams makes ranking the best playoff rotations a relatively straightforward exercise. Here are rotations ranked by unit-wide performances on the year, using the expected 2009 playoff teams, and ranked by SNWP and SNLVAR, which adjusts for lineup variations and value-added over replacement level.

Team        SNWP    SNLVAR
Cardinals   .535     22.5
Dodgers     .534     22.5
Rockies     .515     19.8
Phillies    .511     19.5
Tigers      .511     18.4
Red Sox     .503     17.7
Angels      .499     15.8
Yankees     .492     15.5
Twins       .492     14.8

However, the postseason’s a very different animal from the full season’s marathon; depth gets you to October, but in theory sheer quality-or the convenient, random great game from an Anthony Reyes or a Jeff Weaver-propels you to a title. In the post-season fifth starters go to the bullpen, so whomever the club’s four likely picks wind up being for their playoff rotations, that quartet winds up being the men who matter. Then there’s the question of whether or not the Rockies will have the benefit of Aaron Cook, for example; if injury knocks him out of action for the post-season, are the Rockies anywhere as good as the full-season numbers suggest? Keeping all of that in mind, let’s simplify things by looking just at the combined SNWP mark for the probable front fours of each of these nine teams:

Team        SNWP   The Front Four
Cardinals   .575   Chris Carpenter, Adam Wainwright, Joel Pineiro, Kyle Lohse
Tigers      .562   Justin Verlander, Edwin Jackson, Rick Porcello, Jarrod Washburn
Red Sox     .561   Josh Beckett, Jon Lester, Clay Buchholz, Tim Wakefield
Phillies    .558   Cliff Lee, Cole Hamels, Joe Blanton, Pedro Martinez
Dodgers     .545   Randy Wolf, Clayton Kershaw, Chad Billingsley, Hiroki Kuroda
Rockies     .527   Ubaldo Jimenez, Jason Marquis, Aaron Cook, Jorge De La Rosa
Angels      .524   John Lackey, Jered Weaver, Scott Kazmir, Joe Saunders
Yankees     .520   CC Sabathia, A.J. Burnett, Andy Pettitte, Joba Chamberlain
Twins       .514   Scott Baker, Nick Blackburn, Carl Pavano, Brian Duensing

Essentially, that just tells us what we already sort of expected-the Cardinals are the class of the field, the Tigers, Red Sox, and Phillies make a tightly grouped second tier, and the Angels, Rockies, and Yankees are all roughly equal. Admittedly, this exercise credits fourth starters overmuch-they’ll only ever start one game in any post-season series, after all. However, that said, some notes about some of the teams, and to explain some of the selections and identify the impact of alternates:

  • Tigers: Whether or not Jarrod Washburn’s knee holds up actually ends up mattering very little, if only because a healthy Nate Robertson has been as effetive in his September spin in the rotation as Washburn has been on the entire year.

  • Red Sox: Substituting Dice-K for Wakefield without any adjustments for the value of a now-healthy Daisuke Matsuzaka would cut Boston’s front-four SNWP to .527, which is a bit extreme, but Matsuzaka was that bad. It seems safer to assume that he can be about as effective as Wakefield if he has to start in the knuckleballer’s place.

  • Phillies: It’s a big assumption that Pedro might start instead of J.A. Happ, but workload concerns and Happ’s recently being worn down contributed to the idea that Charlie Manuel rolls the dice and plugs in the famous person in the rookie’s place. If Happ’s starting instead of Pedro, the Phillies’ front four SNWP jumps up to .577, or just as good as the Cardinals’.

  • Dodgers: While Chad Billingsley (.502) and Hiroki Kuroda (.519) were the pitchers I ran with, swapping in Jon Garland (.502) and/or Vicente Padilla (.507) doesn’t materially change their ranking. While the Dodgers have depth, their hopes truly rest on the left arms of Randy Wolf (.561) and Clayton Kershaw (.597).

  • Rockies: If Cook isn’t healthy enough to answer the post-season bell, swapping in Jason Hammel or Jose Contreras only takes the Rockies down to a .521 SNWP.

  • Angels: The real wild card here is if Scott Kazmir continues to dominate as an Angel; his .725 SNWP in four turns is ridiculously good, and he’s been a much better pitcher since coming off of the DL at the end of June. With nine quality starts in his last 10, he’s gotten hot at the right time.

  • Yankees: While Joba Chamberlain’s miserable .456 SNWP drags them down, the ugly non-secret is that Andy Pettitte and A.J. Burnett have been barely better than mediocre, posting SNWP marks of .519 and .518 respectively.

  • Twins: Here, I simply cut 200 points off of Brian Duensing’s sparkling .715 mark in his seven spins through the rotation, because that seemed like a realistic approximation of what he’d be worth on a full season, and if the Twins catch the Tigers, looking at their as one that would be the weakest in the postseason doesn’t seem very extraordinary.

The takeaway? The Cardinals are the obvious class of the league, but the Phillies deserve to rank right with them when you consider that they’ll get to start a series with Cliff Lee and could elect to start Happ instead of Martinez or Joe Blanton. The Tigers and Red Sox are the class of the American League’s crop, but if Kazmir stays hot, he’d give the Halos a group almost as tough.

A version of this story originally appeared on ESPN Insider Insider.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
4 games out with 9 to play = no braves? no giants? i reckon both would fair pretty well by this metric, at least...
In sort of a half vein of "man its to bad" and the rest of "maybe its not too late". The Rangers rotation's SNWP was .511 overall (right in there with the Tigers and Phillies), but the top 4 starters (Scott Feldman, Kevin Millwood, Tommy Hunter, and Brandon McCarthy) top the list with a SNWP of .583

Whoda thunk?
This would a great time in the season to be able to rerun pecota on the remaining squads. We have a lot more data on all the players, and though there are lots of things you can do to try and treat these raw numbers from the season to be more accurate for this playoff situation, knowing which scott kazmir will show up in october is a job best left to a computer.
Pedro Martinez, the break out rookies (happ, bucholz, jackson, and in the reverse, joba), it'd be great to know where they trend, and I with my human mind don't know how to see through the noise.

I'm curious, Christina: the Yankees will likely have the option for the three-starter or four-starter schedule. Does the SNWP gap between a three-man and a four-man rotation point towards an obvious choice for them?
Frankly, I think they should go the three-man route, in no small part because anything that helps them get Sabathia starting two of the first four games in a DS is the best way of making sure it doesn't get to a fifth game and potentially mix things up too badly for their ALCS assignments.
Can we expect a similar exercise with the bullpens? (Please?)
I wish I could say, "your wish is our command," but actually Joe's planning on dealing with that very topic in Prospectus Today, which should run late this afternoon or early evening, after his chat. I'd press him there on making sure he follows through. ;)
Shouldn't you weigh the contributions by the expected number of starts? In a five game series the front guy probably averages 1.2 starts while the 4th guy likely expects less than 1 start. Likewise, in a seven game series the first two guys are far more important, since the first guy is probably going to get two starts while the 4th guy can't get more than one.

This would make the Cardinals look even more incredible, and the Dodgers would probably jump up a few spots.
There were two problems with that as I pondered it yesterday. First, you'd have to start making assumptions about the lengths of the various series, which might then also make the spread of assignments much more contingent. Second, let's face it, everybody benefits when you cheat towards favoring the front three and operate with the assumption that the fourth starter generally gets only one start. Some benefit more than others--the Yankees, certainly, but the Cardinals, Red Sox, and Dodgers as well--and other less (the Rockies and Twins). My purpose was just to present a general look at the strengths of the units to come up with a rough ranking. I think the suggestion that it breaks into three "groups"--the good, the very good, and the Cardinals--is really the somewhat simple takeaway here.
To some extent, the likelihood of the 4th guy getting a start (in a 5 game series) is a function of the strength of the top of the rotation, which makes computing those expected values of games a bit recursive.

Also, to address a comment in the main article that said the 4th guy only gets one start per series...if you are forced to go the distance in consecutive series (or fight right up to the end of the regular season to make it) that's not necessarily true. An advantage of winning early and easily.
While in theory if the Yankees choose to go to the 3 man starter series it helps them. BUT that also allows Verlander to pitch twice in the first four games in the series. I'm not sure what a simulation would say, but my first thought at that scenario was 'the Tigers win more series pitching Verlander in games 1 and potential 4 than the do in game 1 and a potential game 5.' If for no other reason, a game 4 is much more probable than game 5.
Why use SNWP and not adjusted K/9 as the sauce does?
SNWP describes the likelihood of a win for the team, which is sort of the ball of wax at that point of the season.
Thank you for not using the phrase "dangerous in a short series".
No mention at all of John Smoltz as the possible #4 starter for St. Louis? Odd.
Agreed, it merits mentioning, as Lohse has just two starts and 10.1 innings over the past month. He's got a .567 SNWP as a Cardinal, compared to Lohse's .450, so if the numbers were re-run taking that at face value, he'd widen the gap for St. Louis. However, if you take his Boston work into account, his overall SNWP is .431, which if included would close the gap between the Cards and Tigers.
I'd love to see these rankings with only the top three included. It's obvious to see how teams would move up and down, but I'd like to see the magnitude of the differences.