The Untold Story Of LaRussa’s Bullpen Management

One of narratives leading up to Game 1 of the World Series emphasized the Cardinals’ and the Rangers’ heavy use of their bullpens in the Division and League Championship Series. We were told to expect pitching changes early and often, especially from Cardinals manager Tony LaRussa, who has a reputation for micromanaging his pitching staff to get just the right match-up.

In Game 1, LaRussa did not disappoint. Cardinals starter Chris Carpenter threw six innings, giving up five hits and two runs. After LaRussa pulled Carpenter for a pinch hitter in the bottom of the sixth, he used five relievers out of the bullpen to pitch the final three innings. The ‘pen acquitted itself well: three innings, one hit, no runs.

In the Division Series, the Cardinals’ starting pitchers threw 30 total innings to 14 for the bullpen. That’s 68% of the innings for the starters; 32% for the relievers. Over the five games played, LaRussa averaged 3.2 pitching changes per game.

In the NLCS, Cardinals starting pitchers were on the mound for only 24.1 innings to the bullpen’s 21.2 innings, a ratio of 53% to 47% over the six games played. LaRussa made an average of 4.67 pitching changes per game.

Despite LaRussa’s reputation, his bullpen management in the NLCS is the exception, not the rule, of his reign as the Cardinals skipper.

Let me explain.

LaRussa’s first year as Cardinals manager was 1996. In 12 of his 16 seasons in St. Louis, LaRussa’s bullpen pitched fewer innings than average for all National League teams. Only in 1998, 1999, 2002 and 2007 did the Cardinals relievers throw more innings than the National League team average.

Take a look.

Year Average IP by Starters per National League Team IP by Cardinals Starters Average IP by Relievers per National League Team IP by Cardinals Relievers
1996 970 1016 500 430
1997 978 1001 468 455
1998 988 936 464 534
1999 970 913 475 514
2000 976 1001 467 433
2001 960 969 481 467
2002 948 917 496 529
2003 951 979 494 484
2004 943 996 504 457
2005 973 1048 467 398
2006 942 942 504 488
2007 929 889 524 547
2008 932 955 514 499
2009 942 1034 502 437
2010 958 991 485 463
2010 972 999 481 463

* Numbers were rounded

Here’s a look at the numbers on a percentage basis.

Year Percentage of IP by Starters per National League Team Percentage of IP by Cardinals Starters Percentage of IP by Relievers per National League Team Percentage of IP by Cardinals Relievers
1996 66% 70% 34% 30%
1997 68% 69% 32% 31%
1998 68% 64% 32% 36%
1999 67% 64% 33% 36%
2000 68% 70% 32% 30%
2001 67% 67.5% 33% 32.5%
2002 66% 63% 34% 37%
2003 66% 70% 34% 30%
2004 65% 69% 35% 31%
2005 68% 62.5% 32% 27.5%
2006 65% 66% 35% 34%
2007 64% 62% 36% 38%
2008 64.5% 66% 35.5% 34%
2009 65% 70% 35% 30%
2010 66% 68% 34% 32%
2010 67% 68% 33% 32%

* Numbers were rounded

With LaRussa at the helm, the Cardinals made the postseason nine times (1996, 2000, 2001, 2002, 2004, 2005, 2006, 2009 and 2011) and advanced to the World Series three times (2004, 2006 and 2011).

In the 2004 NLDS (Cardinals won 3-1), Cardinals starters pitched 23.1 innings to 13.2 for the relievers. That’s 62% of innings pitched for the starters, 7% below the season percentage for the Cardinals starters. LaRussa made an average of three pitching changes per game.

But in the 2004 NLCS (Cardinals won 4-3), the Cardinals starters pitched 65% of total innings, closer to the season percentage of 69%. In the seven games of the series, LaRussa averaged 3.43 pitching changes per game. Fewer innings pitched by the relievers compared to the NLDS, but more relievers used.

In the 2004 World Series (Cardinals lost 4-0), Cardinals starters threw 17.1 innings with relievers on the mound for 16.2 innings, nearly a 50-50 split. LaRussa made an average of 3.75 pitching changes per game. Woody Williams‘ start in Game 1, in which he gave up 8 hits and 7 runs in only 2.1 innings, significantly affected the numbers in the four-game series.

In the 2006 NLDS (Cardinals won 3-1), the starters pitched 64% of the innings, compared to 66% in the regular season. LaRussa oversaw 3.5 pitching changes per game. In the 2006 NLCS (Cardinals won 4-3), the starters carried the load, pitching 69% of the total innings over the seven games. Even with fewer innings pitched, LaRussa still made an average of 3 pitching changes per game.

The Cardinals starters worked even harder in the 2006 World Series, pitching 80% of the innings in the five total games (the Cardinals won 4-1). LaRussa used only 2.4 relievers per game on average.

In the 2004 postseason, LaRussa’s bullpen worked harder than in the regular season, but in the 2006 postseason the relievers worked less than in the 162-game schedule. And in both seasons, the Cardinals bullpen pitched a lower percentage of innings than did the average bullpen in the National League.

While this year’s NLCS is a poster child for bullpen micromanagement by LaRussa, the NLDS starter/reliever percentage split for the Cardinals was precisely the same as in the regular season: 68% v. 32%. And that split favored the starters more than the average for all National League teams this season: 67% v. 33%.

In baseball, sometimes the facts support the narrative. Sometimes they don’t.




Print This Post



Wendy's baseball writing has also been published by Sports on Earth. ESPN.com, SB Nation, The Score, Bay Area Sports Guy, The Classical and San Francisco Magazine. Wendy practiced law for 18 years before beginning her writing career. You can find her work at wendythurm.pressfolios.com and follow her on Twitter @hangingsliders.


20 Responses to “The Untold Story Of LaRussa’s Bullpen Management”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. MikeS says:

    This is nice analysis and a big sample size but it does not take into account the relative strengths and weaknesses of the starters and relievers. No matter a managers proclivities a good one* will use his best players more than his bad ones. So if LaRussa has historically had stronger starters than relievers that could skew the results. I’m not enough of a Cardinals fan to know if that is the case. This answers the question “does TLR use his relievers more than other managers?” It does not answer the bigger question which could be phrased as “if every team had the same staff would TLR use his relievers more than other managers do?”

    *I know not every one will agree, but 5 pennants and 2 world championships suggest he has something going for him.

    Vote -1 Vote +1

    • Wendy Thurm says:

      Agree that a more detailed analysis comparing strength of starters would show more. The fact that the Cardinals’ bullpen usage was above NL average in 4 years and the Cards made the post season only once in those four years suggests that he does use his bullpen more when he needs to do– when he has weaker starters. But I still thought the overall data was interesting.

      Vote -1 Vote +1

    • CircleChange11 says:

      So if LaRussa has historically had stronger starters than relievers that could skew the results.

      Ryan Franklin has been the closer for th elast 3 years. I’ll let you decide whether that indicates a talented bullpen or not. *grin*

      IMO, this year’s bullpen has been one of the best of the last decade in StL. The additionals of Scrabble, Dotel and the move of Kyle back to the BP and Motte to closer, while Jackson adds more quality IP to the starting staff have turned this team into a force. If StL plays the whole season with the team they have had over the last 2 months (including Furcal at SS and Jay in CF), they have the best record in the NL, IMO.

      Vote -1 Vote +1

      • Anon says:

        I won’t say they would have got the best record, but it is a likely possibility. I would take the Phillies 1st (would have won 106 if they won half of the post clinch 8 game losing streak).

        Vote -1 Vote +1

      • Spoilt Victorian Child says:

        Setting aside what the Phillies’ record might look like if they’d been in danger of losing HFA the last two weeks of the season — are you saying that a few more months of Jay, Furcal, Dotel, and Rzepczynski would account for 13 extra wins?

        Vote -1 Vote +1

      • Anon says:

        The Cards bullpen had 20 blown saves before the trade deadline. Using the save rate after the trade deadline, that would have been 12. Also, look at an improved rotation, deeper bullpen in earlier innings, and better defense for at least 2B, SS, and CF. Consider offense and health on top of all that, and 13 wins isn’t as hard to find as you might expect.

        Vote -1 Vote +1

      • Antonio Bananas says:

        Really Circle? I think they’ve played well, but even 20 games is a small sample in baseball. They play in a super weak division (2 worst teams, come on), so that had to inflate the numbers a bit already.

        Vote -1 Vote +1

  2. Richie says:

    You’re missing the story, so far as I know. I thought the thing on LaRussa was not how many innings he gave to his relievers, but how many relievers he uses to fill those innings. So how does his ‘relievers per game’ or ‘innings/batters per reliever’ compare to the league average?

    Vote -1 Vote +1

    • Wendy Thurm says:

      Good questions, both. Good ideas for a longer post over the off-season. Nevertheless, I still find the data presenting interesting given the narrative leading into World Series.

      Vote -1 Vote +1

    • CircleChange11 says:

      Yeah, that’s always the issue with TLR. I think I brought this up in another thread, but he’s essentially given credit for coming up with the positions of “setup man” and “loogy”, as Rick Honeycutt served well in both roles.

      The 8th inning for a cards fan in often a nightmare. At the stadium, you’ll spend $50 that inning alone. At home, you’ll make 4 trips to the kitchen (if you stay awake). On a hot day game in July, you will seriously consider bringing a sniper rifle in case TLR decides he needs that 4th reliever in 2 innings.

      The cardinals are far too often in the top 5 worst of the blown saves category, and I believe that counts leads lost in the 8th inning.

      TLR can work a bullpen, figuring out who should bat 2nd and/or play 2B is another story altogether.

      Vote -1 Vote +1

  3. Anon says:

    The more telling comparison would be average IP per relief appearence. LaRussa micromanages the bullpen, but lets a good starter go deep in the game.

    2011 NLDS game 5: TLR lets Carpenter bat late in the game because of his performance rather than go to the pen.

    Vote -1 Vote +1

  4. GMH says:

    Wendy,

    Excellent analysis and writing. You made a winning argument for me.

    Perhaps the better narrative for this World Series is the blind hatred people have for Tony LaRussa. Every maneuver the guy makes is deemed “over managing” or “old school,” even if the holy trinity of Earl Weaver, Billy Beane, and Bill James would have made the same move in the same circumstances.

    This narrative seems to be reflected by the comments so far. They seem to be very quick to argue that your analysis is lacking: “A better article would delve further . . .” In other words, Wendy, you need to keep scouring data until you prove our belief that LaRussa is an oppressive micromanager who costs his team wins (even though we have sworn to uphold the mantra that managers really don’t matter much). Their confirmation bias is showing.

    Vote -1 Vote +1

    • Nathan says:

      That’s disingenuous.

      It goes both ways — if you really want to prove or disprove the “common knowledge” of the stat geek (I say this affectionately) community about TLR over-managing, you have to have a sound argument. And while I found Wendy’s article interesting, well-written, and an important first step in that discussion, that’s just it — it is a first step.

      Whichever position you are arguing for, I don’t see how you can come to a full and proper conclusion without analyzing TLRs habits both in isolation and against the league average, as well as normalized against the quality of the pitchers involved, as that is what will tell the story best.

      Vote -1 Vote +1

  5. J-Doug says:

    Wouldn’t pulls be a better measure of LaRussa’s short hook than innings pitched by the bullpen?

    Vote -1 Vote +1

  6. Corey says:

    As someone who has grown up watching LaRussa manage the Cardinals I can tell you he leans on his starters heavily during the season. There are many games he will let his starters take a pounding and labor through 6 or 7 innings just to save the wear on the bullpen. Once he goes to the bullpen, he seems to play the matchups more than other managers. An analysis of the average number of pitchers used per game over the years compared to the league average would be interesting.

    Vote -1 Vote +1

  7. CircleChange11 says:

    I’m also looking at Jackson in the rotation and how he’s better than McLellan, and how Kyle is better than ….

    The bullpen and Theriot’s defense were big team weaknesses.

    I’m saying with a full season of the current Cards are the NL’s best team.

    Vote -1 Vote +1

    • CircleChange11 says:

      Here’s what I’m looking at …

      [1] Jackson, a 3.8 WAR pitcher, is better than McClellan (-0.6 WAR).
      [2] McClellan as a reliever is Augustine, Miller, Bautista who relieved because McClellan was a starter.
      [3] Dotel has been worth 0.9 WAR in 29 games as a Cardinal.
      [4] Scrabble has been worth 0.4 WAR in 28 games as a Cardinal
      [5] Motte (1.5 WAR) is much better than Franklin (-1.2 WAR).

      But mostly I was looking at how many blown saves would realistically have not been blown saves.

      So when you look at just WAR and extend it over a full season, that’s as much as 11 WAR.

      When you factor in more at bats for Jay (~3 WAR) and Craig (~3 WAR) instead of Rasmus (~1 WAR), that’s probably another WAR. Furcal has outperformed Theriot in just 50 games (0.9 WAR to 0.7), so maybe another 1-2 WAR there.

      Punto has also outproduced Schumaker 1.8 WAR to 0.6 WAR in less than half the plate appearances, so another 2 WAR.

      In the end, you’re looking at a much, much, better team.

      So, while the Cards are going to gt called “lucky” for making it to the World Series, and especially if they win it, my point is that the team currently on the field is much better than the team that was on the field for much of the season.

      If you wanted to argue that the players’ performances would be decreased over a full season, that would be a valid point. If you wanted to say that this current team would still slightly trail the Phillies in record, I wouldn’t make a big deal about it.

      Really, I’m just saying the current lineup, rotation, and bullpen is one of the best in MLB, if not the best .. and it would likely be reflected in the standings had this current team played together over the course of a full season.

      Vote -1 Vote +1

      • KDL says:

        Wouldn’t the easiest way to see if La Russa micromanages the ‘pen be a simple relief appearances/relief innings pitched. Then compare this to other managers and/or the league. It would get us looking at how managers use the ‘pen once they’ve decided to use the ‘pen – which is a different question than when to start using the ‘pen.

        The raw number of innings pitched by relievers is more likely to tell us about starting pitching or the runs an offense puts up – or at least the data is noisy as heck for what we are trying to get at. If anything it tells us how quick a hook the manager has for starters, which is 1) not La Russa’s reputation and 2) not what we’re purportedly talking about with this article.

        Vote -1 Vote +1

  8. Cuban X Senators says:

    Yes, the issue has never been whether LaRussa uses relievers for inoptimal numbers of outs, nor that he makes inoptimal numbers of changes, the question has always been does carrying the number of relievers necessitated by his preferences on a roster allow one to have optimal resources available as reserves/relievers.

    Is it better to go ’80 Orioles with a 9-man staff, but have Terry Crowley, John Lowenstein, Benny Ayala & Lee May available to hit, or is it better to go recent LaRussa and carry 13 arms? Or somewhere in the middle?

    Vote -1 Vote +1

  9. LaRussa has 2 World Series rings. From how many playoff trips? And Bobby Cox: 1 in how many trips?

    I love sabermetrics & appreciate your work, Ms. Thurm. But I’ll always hate LaRussa & Cox based on a small sample size: 2 games. Am doing this from memory so what I’m writing is virtually true, i.e., I may have some facts wrong, but the essence is true.

    6th game of the great, see-saw WS between the Twins & Braves. Alejandro Pena was masterfully retiring the Twins. Pena had been out of the Show either earlier that year, or the prior one. A reclamation project from the Mexican League having a nice renaissance as the Braves set-up man. By then he looked like Bouton’s description of Fred Gladding: “a grocer who’s been eating up a good deal of the profits.” The by-then utterly annoying Vin Scully kept informing us that Pena had not pitched over 2 innings that season. Cox’s closer was probably unavailable b/c Cox, like LaRussa, insisted upon 3.9 pitching changes per playoff game.

    But old fat Pena was getting them out. It was the World Series. Me, I get other arms ready. But I Ieave Pena in there as long as he keeps getting them out. Or his arm falls off. Micromanager Cox lifts Pena after 2, brings in Charlie Leibrandt. HR, by Puckett, if memory serves. Braves lose the 6th game & the next night, the Series.

    LaRussa in a playoff series w/ AZ, the year the D-backs won it all. A crucial game. Darryl Kile was pitching for St.L. Kile may have been only a slightly-better-than-average 4th starter by then.

    But he was getting the D-backs out. Inning after inning, through 6 or so. Then LaRussa maybe saw a chance for 1 of his beloved double-switches, or the D-backs had some lefty hitters due up, so LaRussa brings in Ken Dayley. If memory serves. I’m sure Dayley had 45 or so appearances that season. I’m sure St. L wouldn’t have made the playoffs without Dayley. But I’ll bet if Dayley had 45 appearances, at least 5 of them were bad. That appearance against AZ certainly was. Retrosheet could confirm my hazy memory but I wouldn’t be surprised if micromanager LaRussa made 3.9 pitching changes in that inning alone. The game & the momentum for the series was lost. Dayley just didn’t have it that day.

    And Kile had been getting the D-backs out.

    Don’t be Grady Little & have no relievers ready until the game is lost. But LaRussa & Cox earned all their playoff losses. Would you rather have a platoon advantage?

    Or would you rather leave a pitcher out there who’s getting them out?

    I’ve been following Game 4 in only a peripheral fashion as I wrote this. But LaRussa just made another pitching change. & Napoli promptly hit a 3-run HR. Go Rangers. Keep pricking the myth of micromanagers.

    But that Murphy from Texas – he’s sure a rally-killer. Why did Wash move him up in the lineup instead of out?

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>