Best Rotation of All Time?

Back in February, in an article I wrote for ESPN, I pondered the historical possibilities for the 2011 Phillies starting rotation. Last year, the four aces each finished the season with at least four Wins Above Replacement per Baseball Reference. Cliff Lee, of course, spent his time between Seattle and Texas while Roy Oswalt played with Houston and Philadelphia. With all four in the same starting rotation in 2011, though, they had the potential to become the third starting rotation with four pitchers posting individual seasons with four or more WAR.

How have they fared thus far in 2011? Roy Halladay has 6.2 rWAR; Lee, 6.1; Cole Hamels, 5.2. Due to time missed due to injury, Oswalt is only at 1.3. In nearly as many innings, though (about 110), Vance Worley is at 2.5. While the Phillies didn’t quite meet the stringent criteria used in my article from February, I still think you can make an argument that the starting rotation rivals that of the 1991 and ’97 Atlanta Braves.

In 1991, Tom Glavine posted 7.4 rWAR, followed by John Smoltz at 4.7, Steve Avery at 4.5, and Charlie Liebrandt at 4.3. In 1997, Greg Maddux led the way at 7.3, followed by Glavine at 5.0, Smoltz at 4.5, and Denny Neagle at 4.1. As yet, no Phillie is better than the Braves’ best starter in either year, but their #2 and 3 starters rate better.

On average, the Braves rotations were about a fifth of a win (0.2) better than the ’11 Phillies but if you combine Oswalt and Worley’s contributions (only about 20 innings more than Halladay has pitched), then the Phillies’ rotation comes out on top slightly, on average.

Going by defense-independent metrics, the Phillies sprint ahead. Their top three starters each have a strikeout-to-walk ratio at 4.0 or greater. All of the ’91 Braves’ starters were below 3.0. In ’97, Maddux had an amazing K/BB approaching 9.0, but Smoltz and Neagle were below 4.0. Per FanGraphs, this year’s squad has a 2.93 FIP, which is lower than the ’97 Braves (3.30) and ’91 Braves (3.55) by a significant margin. The 1971 Baltimore Orioles’ starting rotation, which had four — four! — 20-game winners, had a collective FIP of 3.60.

There is no way to definitively prove that one starting rotation was better than another, but the deeper you go with Sabermetrics, the more the 2011 Phillies’ rotation looks like the greatest of all time. With three starts left apiece, the Phillies’ starters still have time to move further and further ahead, and you know the Braves, Brewers, and Diamondbacks are watching, paralyzed in fear.

Leave a Reply


This site uses Akismet to reduce spam. Learn how your comment data is processed.


  1. Larry

    September 09, 2011 08:07 AM


    Are your numbers full season from those other rotations? If so, isn’t it likely that Halladay and Lee creep much closer to the 7 WAR mark with 3-4 starts left (adding likely close to 1 WAR each to their season total), making the case even stronger for this rotation?

  2. Bill Baer

    September 09, 2011 08:10 AM

    Yeah, the Phillies ’11 numbers are to-date, not prorated. The inequality is addressed in the closing sentence.

  3. Bill Pettti

    September 09, 2011 08:37 AM

    What happens if you league-adjust the K and BB rate numbers?

    I believe K rates have steadily increased over the past 10+ years. Not sure about BB rates.

  4. Larry

    September 09, 2011 08:45 AM

    Don’t know how I missed that. Thanks for the great work as always!

  5. Bill Baer

    September 09, 2011 08:50 AM

    @ Bill Petti

    Not able to do it now, but I’ll get to it later. Or, if one of my lovely commenters wants to take up the task, that would be great.

    You could come up with a K and BB index (like ERA+) by dividing the pitcher’s rate by the league rate and multiplying by 100.

    e.g. Player 7.9 K/9, League 6.2 = (7.9/6.2)*100 = (1.274)*100 = 127 K/9 index.

    You could also use K% if you are so inclined, which would be more accurate than K/9.

  6. Buzzsaw

    September 09, 2011 08:52 AM

    A really cool website like this will have daily DVD prize updates.

    If i did my math right, Hamels FIP score last night was 6.31

  7. Buzzsaw

    September 09, 2011 09:07 AM

    And if you add in the 5th starter we come out father ahead.

    91 Braves 5th starter was Pete Smith (10 GS 1-3 record 5.06 ERA)

    97 Braves 5th starter was Terrell Wade (9 GS 2-3 record 5.36 ERA)

    11 Phillies 5th Starter Roy Oswalt (19 GS 7-8 record 3.72 ERA)

  8. John Franco

    September 09, 2011 09:08 AM

    If you multiplied them by their rotation spot, measuring depth (1×1, 2×2, etc), you’d get:

    team depth
    ATL91 47.5
    ATL97 47.2
    PHI11 44

    If you did the opposite, measuring star power (1×4, 2×3, 3×2, 4×1), you’d get:

    team stars
    ATL91 57
    ATL97 57.3
    PHI11 56

  9. Rob

    September 09, 2011 11:48 AM

    yeah, not a good start to the FIP contest for most of us.

  10. LarryM

    September 09, 2011 02:11 PM

    If you compare the aggregate starter totals, the 1991 Braves aren’t really all that impressive; by ERA- the 1997 Braves and 2011 Phillies are in a virtual tie.

    To put this in perspective, the 2010 Phillies, in terms of ERA-, is essentially the same as 5 Randy Johnsons (using career numbers).

  11. LarryM

    September 09, 2011 02:16 PM


    Fangraphs has a stat which figures relative FIP and xFIP, FIP- and xFIP-. Phillies are 76 and 82, not as impressive as their 74 ERA-, but pretty darn impressive none the less. ’77 Braves are a little worse on FIP- (78), a little better on ERA- (73), with no data for xFIP- (insufficient fly ball data, no doubt).

  12. LarryM

    September 09, 2011 02:19 PM

    Actually, my bad on the Randy Johnson comp; I truncated his career accidently. Actually, a team of 5 Randy Johnsons (using career numbers) would be a tad WORSE (per ERA-) than the 2011 Phillies.

  13. Buzzsaw

    September 09, 2011 02:56 PM

    if i done things right, the FIPS DVD leaders at the moment are

    3.57 Buzzsaw
    3.72 Tom
    4.23 David
    4.34 SlackerJoe
    4.45 Proco

    I think we should just call it now…

  14. Jeff G.

    September 09, 2011 03:06 PM

    I’d wanna see something that shows K-rates have steadily increased while other rates have stayed the same before I accepted a league adjusted K-rate.

  15. jauer

    September 09, 2011 03:34 PM

    Between Questec and Eric Gregg’s retirement (seriously, the 1997 NLCS strike zone was embarrassing), it seems impossible that BB rates haven’t increased.

  16. Brad

    September 09, 2011 05:23 PM

    Why did you not even use the Braves best years to compare? On ESPN’s all time greatest pitching staff list they have #4 as the 1993 Braves and #1 as the 1998 Braves. I wonder how Philly matches up against those years?

  17. Bill Baer

    September 09, 2011 05:50 PM

    The ’91 and ’97 teams are the best by rWAR. The only two rotations in history with at least four members with four-plus rWAR.

  18. LarryM

    September 09, 2011 06:01 PM

    On an aggregate level, the ’98 team is comparable to the ’97 team and the 2011 Phillies, with the ’94 team and the ’91 team lagging behind.

    The season isn’t over, but absent a stretch of uncharacteristically poor starting pitching, IMO the Phillies will end up essentially tied with the ’97 and’98 Braves.

    Of course the Braves’ performance over a course of years will probably never be duplicated, and was just insanely good. From ’91 through ’98, their starters had an ERA- of 79. Basically five Cole Hamels for 8 straight years.

  19. Phillie697

    September 10, 2011 07:12 PM

    From 91 to 98 Braves starters only averaged 1007 innings pitched per year, or 201.5 innings per starter, assuming 5 starters. Phillies starters are already at 932.1 innings with 21 more games to go (Halladay 210.2, Lee 203.2, Hamels 194 already). So, not quite five Cole Hamels. When we talk about the overall performance of a pitching staff, quality AND quantity should matter.

  20. larrym

    September 11, 2011 01:06 PM

    “ONly 1007, average over 8 years? That’s actually pretty darn impressive. More later, only mobile access at the moment.

  21. larrym

    September 11, 2011 06:01 PM

    Stil haven’t had time to do as much research as I wanted, but what I’ve done confirms that the 97 Braves, 98 Braves, and 2011 Phillies (projected) are in a virtual tie.

    The IP actually works slightly (but not much) in favor of the 2 Braves teams. ’98 IP of starters was 1096.67, ’97 was 1074.67. This year’s Phillies projected to be 1071.67, but with players being rested for the playoffs, may end up a little lower.

    Both Braves teams are a tad ahead on ERA- (73 for each versus 74 for the Phillies). Phillies are ahead a little more significantly on FIP-, but a big divergence between FIP- and xFIP- suggests they have benfitted somewhat from good fly ball luck.

    More later if I have time. Again the 3 teams are so close that it remains a virtual tie.

  22. Phillie697

    September 11, 2011 07:32 PM

    Somebody has to pitch those 21 games. Even at 5 innnings per game for whoever the starters are, 21 games would add 105 more innings to 931, which would mean 30+ more innings than even the 98 Braves. That’s, again at 5 innings per, 6 more starts of elite quality pitching than the 98 Braves. I mean, it’s all very close I agree, but don’t underestimate this rotation’s contribution quantity-wise.

  23. Phillie697

    September 11, 2011 07:35 PM

    Woops, I did my math wrong, LOL. Forget everything I said above then ๐Ÿ™‚

  24. larrym

    September 11, 2011 08:59 PM

    The 2003 Yankees should probably be in the conversation as well.

Next ArticleMLB and Doublethink