NL East Rotation Showdown

So a mini-argument broke out on Twitter today. No one got hurt, fortunately, but feelings may have been bruised. It all started when radio host Kent Covington suggested that the starting rotation in Atlanta is better than the one in Philadelphia, which is not an outrageous claim in and of itself. However, he suggested that the Phillies’ duo of Roy Halladay and J.A. Happ is only marginally better than the Braves’ duo of Jair Jurrjens and Tommy Hanson.

As a proclaimed radio host, I’m sure his intention was just to ruffle some feathers. He did so successfully, as his Twitter feed is mostly a litany of responses to outraged Tweeters. I don’t write this to call him out, however. It was a thought-provoking claim and it broke up the monotony of the baseball-less winter.

Still, I was curious. Is the Braves’ rotation better than that of the Phillies? Which rotation in the NL East is best? To answer these questions, I took the PECOTA projections from Baseball Prospectus (using WARP) and each team’s starting five according to the playing time projections from Heater Magazine. Here’s what I found:

  • Phillies (Halladay-Hamels-Blanton-Happ-Moyer): 11.7 WARP
  • Braves (Hudson-Lowe-Jurrjens-Hanson-Kawakami): 10.2
  • Marlins (Johnson-Nolasco-Volstad-Sanchez-Miller): 8.1
  • Nationals (Marquis-Lannan-Olsen-Martin-Stammen): 5.4
  • Mets (Santana-Pelfrey-Maine-Perez-Nieve): 4.7

PECOTA thinks the Phillies’ rotation is tops, about 1.5 WARP better than the Braves’ five. Additionally, Roy Halladay and Cole Hamels are the top two pitchers in the NL East according to the projections.

A nifty chart for you visual learners (click to enlarge):

Note: Pitchers are ordered by WARP, not by probable spot in rotation.

Of course, these are just projections and the baseball season will never pan out exactly as the projections say they will.

Kent lost me when he attempted to put Jair Jurrjens on the same pedestal as Roy Halladay. In several Tweets, he said:

Didn’t forget a/b Halladay. Jurrjens posted 2.60 just ERA. Halladay has never posted an ERA that low in a full season.

Jurrjens has yet to prove he can be quite the innings horse that Halladay is, but my point is, Halladay isn’t 1995-Maddux.

Jurrjens had better ERA, opposing average, & fewer HR’s. Now then, I agree that Halladay had a tougher road in the AL East…
…& I’m not even saying that Jurrjens is better than Halladay. But when you look at numbers, it’s hard to make a credible…
argument that Hallday’s in a completely different class. The facts just don’t support that claim.

The only reason I highlight this is to stress the use of BABIP in analyzing pitchers. Frequent readers of this blog are likely tired of me harping on this, but it is so important. BABIP does not correlate from year to year. When that BABIP significantly deviates from .300 (the league average), barring an obvious explanation, we attribute that season to luck and label the season an outlier. Both Hamels’ 2008 and ’09 seasons were outliers thanks to the respective .270 and .325 BABIP.

Jurrjens last year had a .273 BABIP, which is nearly as low as Hamels’ in ’08. What makes it worse for him is that he didn’t have the favorable strikeout and walk rates (and subsequent K:BB ratio). Thanks to the depressed BABIP, Jurrjens was also able to strand an abnormal amount of runners — nearly 80% to be specific.

Jurrjens struck out a meager 6.36 and walked 3.14 per nine innings, a K:BB ratio of just over 2:1. It’s decent, but nothing awe-inspiring.

Halladay, meanwhile, struck out 7.83 and walked 1.32 per nine innings last year, which gave him a K:BB ratio of nearly 6:1. He had a BABIP of .313. As noted here, Halladay has compiled his eye-popping numbers in the most offensively-potent division in the Majors: the AL East. It is very reasonable to expect his numbers to improve given the weaker level of competition in the NL East.

In no universe, except the Bizarro universe, is Jurrjens comparable to Halladay. PECOTA puts Jurrjens at about 3 WARP in 2010; I will bet the under on that. Halladay is at about 4.5 WARP; I’ll take the over. And I’m not just being fannish — I’m the guy calling for a significant regression for J.A. Happ for reasons along the same lines as Jurrjens.

Tommy Hanson, on the other hand, had a .280 BABIP in 2010 and he’s an example of a pitcher we can reasonably expect to post a slightly below-average BABIP due to his high strikeout rate (8.2 per nine last year). Matt Swartz of Baseball Prospectus found the following:

Pitchers who are at least one standard deviation above average in strikeouts enjoy the benefit of even lower BABIPs than moderately high strikeout pitchers. These extreme power pitchers have overall BABIP of about .285, and this drop in BABIP is consistent among all four count types.

If I’m a Braves fan, I’m putting my eggs the basket of Tommy Hanson; not Jair Jurrjens. Anything that Tim Hudson, Derek Lowe, and Kenshin Kawakami give is dessert.

Leave a Reply

*

30 comments

  1. Padman Jones

    February 10, 2010 11:28 PM

    I agree that it’s silly to imagine that Jurrjens is on Halladay’s level. But it’s worth mentioning that Jurrjens posted a better FIP (3.59 vs 3.68) in 2008 than 2009 despite a significantly worse BABIP (.311 vs the .273 you mention). Jurrjens’ big strength seems to be limiting home runs, as he’s been at 0.53 and 0.63 the last two years.

    So but what simultaneously worries and heartens me as a Braves fan is that his FB% skyrocketed (26.5% -> 39.1%) betweeen ’08 and ’09…but he still managed to be very good. If that FB% regresses more toward his ’08 numbers, he should be the second-best starter in that rotation…and Tommy Hanson is nothing to sneeze at as a #3.

  2. Jay Ballz

    February 10, 2010 11:36 PM

    Good stuff as always.

    Phils have the advantage, in my book.

  3. Jurrjens4NLCY

    February 10, 2010 11:43 PM

    I have to disagree. Jurrjens was a better pitcher than Halladay last year. A pitchers job is to not give up runs, anything after that is just a secondary stat. You could argue Jurrjens was very lucky, but luck or not, he was better than Halladay last year.

    With Halladay the NL, it’s no question who will have a better 2010 (Halladay).

    But, If Jurrjens posts another year with slightly regressed stats, I will put him in the future ace column.

    Moyer is a god awful pitcher with an IP regression. Hamels is still a question mark. Blanton is very reliable and has good stuff. Happ is likely due a sophmore slump. And Doc is the Doc.

    Hanson has great stuff, should pitch about 170 innings. Jurrjens is due a regression, but not as much as Happ. Lowe will improve in 2010, but not by much. Hudson is a bit of a health risk, but if he pitches, he’ll pitch like an ace. Kawakami had a 3.33 ERA after april, he’ll do just fine being the best 5th starter in baseball.

    Halladay > Hanson
    Blanton < Jurrjens
    Happ < Hudson
    Hamels = Kawakami
    Moyer < Lowe

    I'd give it to the Braves

  4. Bill Baer

    February 11, 2010 12:01 AM

    Jurrjens4NYCY, I agree that Jurrjens had a better year than Halladay last year, albeit it in a much weaker division. However, that wasn’t the claim being made; the claim being made is that Jurrjens is on the same level as Halladay going forward. His 2009 was very much luck-based, meaning that we can’t reasonably expect him to repeat the same performance going forward since it was mostly out of his control.

    To elaborate, the standard deviation on pitcher BABIP is around .020, meaning that we would expect 16% of pitchers every season to post a .320 BABIP or higher and 16% to post a .280 BABIP or lower. Since pitcher BABIP doesn’t correlate year-to-year, we wouldn’t expect any pitcher to repeatedly under- or over-perform his BABIP.

    I flinch when people do the player-to-player comparison; it should be done in aggregate.

    As an example, an analysis of offense is poor if I compare the Phillies to the Yankees and conclude:

    – Ruiz < Posada
    - Howard > Teixeira
    – Utley > Cano
    – Rollins < Jeter
    - Polanco < A-Rod
    - Ibanez > Winn
    – Victorino < Granderson
    - Werth > Swisher

    There’s no room for degree; it’s a simple binary. According to this, the Phillies and Yankees are equivalent offensive teams even though the Bronx Bombers last year outscored the Phillies by 0.6 runs per game.

    I also disagree with your statement, “Jurrjens is due a regression, but not as much as Happ.”

    Here are their 2009 stats that are directly under their control along with their BABIP:

    – BABIP: Jurrjens .273 / Happ .270
    – K/9: 6.36 / 6.45
    – BB/9: 3.14 / 3.04
    – K:BB: 2.03 / 2.13

    My recommended ERA predictor, xFIP:

    – Jurrjens: 4.34
    – Happ: 4.33

    They’re pretty much the same pitcher except one throws from the left side. I’d say they’re due similar regressions.

    EDIT: Jurrjens, as you note, is better at inducing ground balls despite the 8.6% drop in his GB% from ’08 to ’09. I would definitely put Jurrjens above Happ but it is clearly not as wide a gap as you and many others seem to believe.

  5. Mike P

    February 11, 2010 01:02 AM

    I don’t frequently comment on the blog, but I just wanted to let you know how much it is appreciated. This was yet another great post.

  6. Phillies Red

    February 11, 2010 01:13 AM

    Wow. Those Happ-Jurrjens numbers are real eye openers. I do have to wonder about Jurrjens’ homer suppressing skills. Two consecutive years of that is a strong track record, and for what it’s worth, James, CHONE, and Marcel like him for another low HR/9.

    As for the analysis, spot on. I think PECOTA and CHONE (www.beyondtheboxscore.com/2010/2/10/1304375/looking-at-rotations-with-chone) have good reason for projecting the Phillies higher than the Braves. The Braves are relying on an injury riddled pitcher, an aging ground-ball thrower coming off a bad year, a young flamethrower, and a pitcher due for some regression. My suspicion is that these projection tools can’t help but knock down the Brave due to their dependence on something of a motley crue.

    That said, it sure seems that the Braves’ rotation has a lot of upside, more so than the Phils. If Lowe turns it around or just regresses, Hudson manages to pitch 170+ innings, and Hanson and Jurrjens continue to develop, they have four pitchers who could easily be 3.5-5 WAR pitchers. And Kawakami isn’t a slouch either. Lots of potential there.

    We’ve covered this angle before: if the Braves can stay healthy, they could be very good. But relying on Chipper, Hudson, Glaus, Wagner, and Saito is certainly a gamble.

    Oh, right. Halladay and Jurrjens. I think you covered this nicely.

  7. Phillies Red

    February 11, 2010 01:29 AM

    Dude, I’m a believer in stats and numbers as much as anyone. But I’m also a Philadelphia sports fan. And that last tweet of your likely just jinxed us for another 25 years.

    [Crashburn Alley EDIT: This is the Tweet he is referencing.]

  8. hk

    February 11, 2010 07:35 PM

    Bill, you compared Jurrjens’s xFIP to Happ’s FIP. Happ’s xFIP is a little higher than Jurrjens’s xFIP. Below is a list of each rotation by xFIP:

    Hudson 3.47 / Halladay 3.05*
    Hanson 4.03 / Hamels 3.69
    Lowe 4.19 / Blanton 4.07
    Jurrjens 4.34 / Happ 4.49
    Kawakami 4.61 / Moyer 4.74**

    * Halladay’s #’s were accumulated in a much tougher division and facing DH’s instead of P’s.
    ** Kendrick’s xFIP = 4.08.

  9. Doc | bss.com

    February 11, 2010 09:04 PM

    I was looking at Jurrjens numbers just the other day and wondering, “hey, why isn’t anyone calling for him to fall back a bit?” The k/bb ratio is what struck me most.

  10. Kevin

    February 11, 2010 09:51 PM

    Bill,
    Once again, amazing work.

  11. Kevin

    February 11, 2010 09:55 PM

    is there any way you can clearly lay out what xFIP, BABIP, and all the other abreviations mean? I mean, I go and look it up every once in a while but I then forget it when it comes up again. It might help to explain it the first time it is used in each article. Just a suggestion. It would be more welcoming to those who don’t look at this quite as often.

  12. Bill Baer

    February 11, 2010 10:46 PM

    Kevin,

    The equations for FIP and xFIP are laid out in this article about Cole Hamels. In case you don’t want to click through:

    FIP = (HR*13+(BB+HBP-IBB)*3-K*2)/IP
    xFIP = ((FB*.11)*13+(BB+HBP-IBB)*3-K*2)/IP

    They are ERA estimators that give you an idea of a pitcher’s value on baseball events within his control.

    As for tRA, another ERA estimator:

    tRA (the FanGraphs version) “involves assigning run and out values to all events under a pitcher’s control and coming up with an expected number of runs allowed and outs generated in a defense and park neutral environment”.

    That may sound confusing, but every base-out state has an expected run value. You can find them here:

    baseballprospectus.com/statistics/sortable/index.php?cid=68778

    And you can also find more detailed versions in Tom Tango’s “The Book: Playing the Percentages in Baseball”.

    BABIP is batting average on balls in play. Pitchers don’t have very much control over it and it has no year-to-year correlation. Hitters tend to have better control over BABIP. I’d recommend the Matt Swartz article linked above for more on pitcher BABIP. Swartz also detailed hitter BABIP here:

    www.baseballprospectus.com/article.php?articleid=9533

  13. Dan

    February 12, 2010 10:27 AM

    Good work, and note the reason power pitchers have an advantage from BABIP is (I seem to remember) due to the pop ups that they induce. Take those away and their average BABIP is similar to most other pitchers.

    And I totally agree on the expectations of Halladay>>>>>>anyone else in the NL East this year. Well done.

  14. MG

    February 12, 2010 01:29 PM

    Really good post but the major caveat is assuming that each of these 5 guys makes most of the starts for their respective teams.

    Doubt Moyer play much of a role this year and it wouldn’t surprise me if the Phils get a little snake bit with an injury that forces at least 1 of their starts (Halladay/Hamels/Blanton/Happ) to the DL for at least a 1-month if not more. Phils have been pretty fortunate to avoid the injury bug to their starters the last 2 years.

    As long though as the front 4 of Halladay/Hamels/Blanton/Happ stay relatively healthy the Phils should have the best rotation in the NL East but it wouldn’t surprise me if the Braves or Fish have very strong results too. Braves may even end up with slightly better overall numbers too because they have a bit more depth than the Phils do.

  15. MG

    February 12, 2010 01:36 PM

    One last thing – PECOTA generally has been the least accurate of the major predictors of performance since they tweaked their formulas 2 seasons ago.

    Most interesting rotation if you look at it is the Fish. Lots of variation in what the likes of Bill James, Marcel, PECOTA, and CHONE are projecting for Volstad, Nolasco, and Sanchez.

  16. Bill Baer

    February 12, 2010 07:42 PM

    MG, I’m sure you meant to imply it, but the Braves have a much higher risk of injury with their starting rotation than the Phillies. The Phils are also known for having one of the best training staffs — Baseball Prospectus awarded them the Dick Martin Award:

    The Phillies three-year and five-year averages for both days and dollars lost to the DL are among the many reasons this team is taking home the trophy this season. In just the third year after taking over from long-time trainer Jeff Cooper, Sheridan’s staff is a relatively new one, but it’s already considered one of the best in the business.

    As for PECOTA, it did have a bad year in 2009, but its track record is impressive.

    vegaswatch.net/2008/09/evaluating-april-mlb-predictions-2008.html

    vegaswatch.net/2007/10/evaluating-april-mlb-predictions-again.html

    www.baseballprospectus.com/article.php?articleid=5609

  17. Undocorkscrew

    February 12, 2010 09:23 PM

    ‘But Hamels had a better FIP, DOP, BLIP, TUP, RIP, SIP, and HIP than any starter on the Braves staff.’

    Not trying to discredit all sabermetric stats, but quite a few of them are pretty silly. Watching a player perform and using the stats MLB recognizes now has worked for me for years, and I see no reason to change that. I keep seeing sabermetric fans say things like, ‘RBI and hits aren’t relevant.’ I think when people start talking like that, they’re really losing focus of the game.

    Regarding who has the better rotation, it’s completely impossible to tell this early. The Braves are relying on Hudson to stay healthy, Hanson to not have a huge drop off, and Jurrjens to stay consistent. The Phillies are relying on Hamels to rebound, Happ to not have a huge drop off, and figuring out who their 5th starter is. In terms of depth, the Braves have the advantage. Jurrjens has seen success during the minors and majors despite his BB/K ratio. Hudson, if healthy, is a solid #2. Hanson has the potential to be one of the most dominant starters in the game. Lowe will be Lowe, not expecting much from him. Kawakami was probably the best #5 guy in baseball last season, but who knows if he can keep that up.

    In terms of offense, Phillies have a clear advantage. No argument could possibly be made here. But the Braves actually outperformed the Phils offensively after the break last season. This was due to getting rid of Frenchy, giving Diaz the full time RF job, and giving Prado the full time job at 2B. All of this despite Chipper hitting around .209 in his last 80 or so games.

    Obviously they have to rely on Glaus, Chipper, and McLouth to stay healthy. Chipper has stayed healthy enough over the past 3 seasons, no reason to think he won’t keep that up and bounce back. Glaus, if healthy, will help with the lack of power production from last years club. Escobar is still improving each year. Not to mention Jason Heyward, who could possibly have an immediate impact offensively.

    It’s gonna be a fun year.

  18. Bill Baer

    February 12, 2010 09:38 PM

    I keep seeing sabermetric fans say things like, ‘RBI and hits aren’t relevant.’ I think when people start talking like that, they’re really losing focus of the game.

    In honest analysis, RBI is one of the worst statistics you can use because it is less dependent on the player in question than it is on the offensive firepower that surrounds him, particularly the players in front of him.

    Ryan Howard has been able to drive in so many runs because he has had Jimmy Rollins (.329 OBP), Shane Victorino (.347 OBP), and Chase Utley (.379 OBP) getting on base in front of him.

    Would Howard drive in 140+ if he had three Pedro Felizes (.293 OBP) hitting in front of him? Not even close.

    It’s not a Sabermetric concept to discard RBI; it’s a logical concept. It just so happens that Sabermetrics takes a logical approach to analysis of baseball teams and players.

    Batting average is irrelevant because we have on-base percentage. Very rarely will you want to use batting average in lieu of OBP.

    As for the Braves, I do agree with you. I think that if they stay healthy, they’re a mid-80’s team in terms of win. With a little luck they could conceivably win the NL East.

  19. Undocorkscrew

    February 12, 2010 09:50 PM

    [quote]Batting average is irrelevant because[/quote]

    Batting average is used to show how good a hitter is on balls in play. Just because OBP gives you a better overall picture of how well a guy does while at the plate, it doesn’t mean AVG is irrelevant.

    Can you make an argument as to why ‘hits’ are also irrelevant?

  20. Bill Baer

    February 12, 2010 09:57 PM

    ‘Hits’ is a sum total, meaning that you don’t have a baseline (such as PA or AB). Comparing one sum total to another is meaningless without context.

    15 HR in 100 AB is much more impressive than 15 HR in 500 AB, for example.

    Batting average is used to show how good a hitter is on balls in play.

    Technically, that would be BABIP.

    Anyway, using batting average, you would equate the following players equally:

    – Player A: 3-for-10, 3 singles
    – Player B: 3-for-10, 3 doubles
    – Player C: 3-for-10, 3 triples
    – Player D: 3-for-10, 3 home runs

    You would judge Player A higher than Player B here:

    – Player A: 3-for-10, 3 singles
    – Player B: 0-for-7, 3 walks

    “how well a guy does while at the plate” is more than just getting hits; it’s how many bases the player is advancing on those hits (SLG) and how often he’s reaching base (OBP; via walk, HBP, etc.).

    You will almost never need to use AVG when you have OBP and SLG. In fact, I can’t think of a reason why you would want to.

  21. Undocorkscrew

    February 12, 2010 10:21 PM

    Nicely put. Thanks for actually explaining why you prefer these stats, instead of just saying it.

    I do have a beef with your blog though, if I can change the subject here a bit. What’s with all of the ‘Braves whining’ garbage? I’m all for friendly rivalry, but it seems that you were a bit annoyed that the Braves got to Cliff Lee before any other team. You go out of your way to bash the Braves ‘for whining’ yet conveniently ignore the nonsense spoken from Phillies players and/or coaches. Every team has players and coaches that make excuses, as well as fans. Can’t exactly understand why you were singling out the Braves all year…

  22. Undocorkscrew

    February 12, 2010 10:25 PM

    ….and you’re taking McCann’s comments out of context. Is it not known that the park is one of the most hitter-friendly parks in all of baseball? Looked to me that he was just exaggerating to prove his point. I doubt he was ‘whining.’

  23. Bill Baer

    February 12, 2010 10:29 PM

    The “whining” is meant as sarcasm. Unlike a lot of Phillies fans, I actually don’t hate any teams, not even the Mets or Braves. In fact, I liked the mid-1990’s Braves as they were the epitome of a model franchise.

    As for CBP being hitter-friendly, it is, but not nearly as much as you may think. This article by Todd Zolecki explains why.

  24. Undocorkscrew

    February 12, 2010 10:32 PM

    Oh great, another made-up stat. But this time it’s park-related. Wonderful….

    Its three-year run index of 103 actually is tied with San Francisco’s AT&T Park and just one percentage point higher than Nationals Park (102) — both of which are considered pitchers’ parks.

    That’s all true, according to a made-up stat that holds absolutely zero credibility.

  25. Bill Baer

    February 12, 2010 11:02 PM

    Have you bothered to learn what it is?

    The park factor is a variation of an index. Indices are very common in statistics and used in many fields; it is not in the least made up. Give this page a read:

    www.baseball-reference.com/about/parkadjust.shtml

    You could use the same general calculation — (Designated / Base) * 100 — to rank city populations or company sales. If you call park factors “made up” then you are also calling the practices of scientists, economists, sociologists, businessmen, etc. “made up” as well.

  26. Phil Cherry

    February 12, 2010 11:13 PM

    All stats are “made up” by interpreting data via a set mechanism. When a Bill James stat doesn’t hold water, I question where the tidal waves, volcanoes and other apocalyptic staples are.

  27. Mike

    February 14, 2010 11:54 AM

    there’s no point in debating with some people. those people “know what they know” and damned if they’re going to listen to crazy things like “facts” or “logic” that might contradict their existing beliefs.

    it’s been shown pretty clearly that CBP is not overly hitter-friendly. The people who contend that it is never seem to have any facts to back up their assertion, and tend to focus only on home runs. Zolecki’s article sums it up pretty nicely. there is more than one way to score runs. it is easier to hit HRs at CBP, but harder to rack up extra-base hits. those factors tend to cancel each other out to some degree, resulting in the overall park factor.

  28. Unikruk

    February 14, 2010 07:45 PM

    Very good post.

    However, it’s become clear to me that we need a national requirement of two years of statistics, or statistical thinking, beginning in middle school. Some people just aren’t equipped.

  29. Bobo

    February 16, 2010 10:19 AM

    Not terribly relevant to the very interesting and informative discussion above, but wanted to point out that Jurrijens is apparently going in for an MRI due to shoulder soreness…

Next ArticleBDD: Frank Thomas, BBWAA, and Slam Dunks