Skip to content

Latest research shows managers inefficient when taking pitchers out of the game

Feb 28, 2014, 8:15 PM EDT

Reds' manager Dusty Baker watches from the dugout as his team plays the Giants in Game 5 of their MLB NLDS playoff baseball series in Cincinnati Reuters

Peter Dizikes of the MIT News Office highlights one of the finalists in the MIT Sloan Sports Analytics Conference, currently being held in Boston. The paper, “A Data-driven Method for In-game Decision Making in MLB”, was developed by John Guttag and Ganeshapillai Gartheeban.

Guttag and Gartheeban’s developed a model that suggested when managers should take their starting pitchers out of the game by using data from the first 80 percent of the 2006-10 seasons to develop a model. Then they tested their model against the results in the final 20 percent of the season.

Dizikes summarizes some of the important findings:

The study finds that from the fifth inning on, in close games, pitchers who were left in games when the model recommended replacing them allowed runs 60 percent of the time, compared to 43 percent of the time overall.


Over 21,538 innings, the Guttag-Gartheeban model disagreed with the manager’s decision regarding his starting pitcher 48 percent of the time. About 43 percent of the time, the manager left the starting pitcher in when the model indicated he should be replaced. In just 5 percent of the cases did managers pull starting pitchers when the model suggested they should stay in the game

Guttag and Gartheeban do note that their methodology doesn’t consider every factor and recognize that the manager is considering a lot more than just the current game situation. Hypothetically, leaving Justin Verlander in with a 110 pitch count in a 2-1 game in the bottom of the eighth in late September when the Tigers are only in the lead by a game could be justifiable when most objective models would suggest taking him out.

That being said, many managers can improve simply by not having their pitchers throw unnecessary innings, like the seventh inning after their team has taken a 10-1 lead.

  1. jonrox - Feb 28, 2014 at 8:22 PM

    That photo was an excellent choice

    • donniev77 - Mar 2, 2014 at 3:28 PM

      Can someone explain how one of the worst managers ever received so many great opportunities?

      He managed to screw up the bat boy decisions in SF by putting his own son in harms way.

  2. chacochicken - Feb 28, 2014 at 8:28 PM


    • js20011041 - Feb 28, 2014 at 8:44 PM


      Forget baseball. They should be concentrating on the deep philosophical musings of Ogre. Such as the question, “what if dog were spelled C.A.T?”

    • chacochicken - Feb 28, 2014 at 9:36 PM

      For those uncertainly of my sincerity, I was quoting David Price.

      • js20011041 - Feb 28, 2014 at 9:39 PM

        Didn’t he go to Vanderbilt?

      • chacochicken - Feb 28, 2014 at 9:42 PM

        For at least half of the collegiate baseball season I’d reckon.

      • historiophiliac - Mar 1, 2014 at 12:26 AM

        Don’t explain your jokes, honey. Drop them and walk away. Don’t stoop to the level of the humor-impaired.

  3. agelardi - Feb 28, 2014 at 8:45 PM

    Should have been a pic of Grady Little.

    • js20011041 - Feb 28, 2014 at 8:50 PM

      Jesus. Screw one sheep and you’re a sheep porker for life.

    • Jack Marshall - Mar 1, 2014 at 1:50 AM

      No, I think it should be a picture of a manager…

    • mmeyer3387 - Mar 1, 2014 at 6:20 PM

      First, I want say that I’m not a sox hater. In truth, I’m not much of an Yankee fan. That being said, you guys are always on Grady Little. Over the years I’ve some say that they didn’t win the WS because Grady was a bad manger. Furthermore, you guys act like if he had made one move of taking his pitcher out of the game earlier, the sox would have won the WS that year. However, the main reason that they lost the game series that year is because they lacked the pitching talent when matched up against the Yanks that year (as most teams did.) After Pedro Martinez their pitching was average. Plus the pitching staff was worn down from a long season.When they went to go on and beat the Yankees and the WS, was because of better pitching talent. When they they inproved their pitching by adding another all star starting pitcher in Curt Schilling, along with a good closer they became a much better team. With new additions, they went on to win a couple world series. I was rooting for them against the Yankees and I enjoyed watching them win against the Yankees. However, the main reason that they won was because they became a more talented team, which allowed them to match up well against the Yankees.

      • agelardi - Mar 1, 2014 at 9:26 PM

        Believe it or not I’m not a Red Sox or Grady Little hater either. Just a fan of beating a dead unfunny horse.

  4. paperlions - Feb 28, 2014 at 9:02 PM

    The problem with the model is that it is only considering what happens when pitchers are left in, it does not consider what the bullpen does when the pitcher would be pulled instead or the fact that those innings have to be pitched by someone and teams are already carrying too many pitchers because of bullpen usage.

    Yes, the model is correct in that teams would allow fewer runs during that one extra inning or so that a starter is left in, if he was pulled…but it doesn’t appear to be accounting for the fact that the worst pitchers in the pen will be the guys that are throwing more innings if starters are pulled more often…because they sure aren’t going to be putting in their best relievers to pitch the 5th or 6th.

    I would guess that within the context of an entire season, if the model was expanded to show likely results from pulling the starter early and giving more innings to the worst pitchers on the staff (which has to happen because someone has to pitch those innings), it would show that teams would actually give up more runs. The point isn’t whether or not starters tire and give up more runs the 3rd time through the lineup, of course they do….but they are still better than most middle relievers, and bringing in middle relievers that early isn’t doing anyone a favor but the other team.

    • js20011041 - Feb 28, 2014 at 9:25 PM

      I agree with this to an extent. There are just so many variables that go into (or, at least, should go into these decisions. I’m convinced that most managers don’t look past score and inning.) these decisions. Who’s coming up to bat, who’s available to pinch hit, who do I have in my bullpen, who needs work, who needs rest, do I want the platoon advantage. I’m sure there are more. It has to be almost impossible to create a model that takes all of these variables into account. Like you said, I just don’t think we’re gleaning that much information here. We know managers don’t do a good job handling pitching staffs. This is not news.

      My one nitpick however is that if managers are appropriating their innings more efficiently, it doesn’t matter if you give up more runs in a season. If you’re up 10-1 in the fifth and you decide to pull Justin Verlander to give him some rest and bring in the last man in the bullpen, who cares if he gives up a few runs? If managers can get better at leveraging their better pitchers by giving them the high leverage innings and giving their lesser pitchers the low leverage innings, it doesn’t matter if the overall runs allowed is worse. I think it would actually probably even out in the end, from a runs allowed perspective, but with greater results in the win column.

    • doctornature - Mar 1, 2014 at 10:49 AM

      Pretty easy to compare a starter’s ERA in each inning, after so many pitches, against certain hitters with men on base after so many pitches etc., and then compare them to the incoming pitcher’s stats against the next batter, or probable pinch hitter.

      There needs to be a software developed for managers with the ‘gut’ instinct, like Ron Washington, who wouldn’t recognize a stat if it bit him in the butt. Dusty and a few others come to mind.

      • paperlions - Mar 1, 2014 at 11:58 AM

        Not exactly so easy because there will be a re-distribution of the IP among the pitchers on the staff, which is not easily predicted. But, yes, some range of possible outcomes could be generated. Most late inning relievers are already pitching about as much as they can, which means that the IP by starters will have to go to middle relievers, generally the worst pitchers on the staff and guys whose effectiveness suffers more as they pitch in less ideal situations for their skill set.

        I guess what I mean is….it isn’t really as easy as saying starters would give up fewer runs if managers had quicker hooks and therefore the team will give up fewer runs….the important conclusion does not necessarily follow from the first.

    • thisdamnbox - Mar 1, 2014 at 1:23 PM

      PL, you had to go and kill the fun with your logic…tsk, tsk…

      Baker was a perfect choce of pic simply because he single-handedly destroyed Solomon Torres’ career while ending the Giants’ season a game short of the playoffs that year…Must-win final game of the season against the Dodgers and he gives a kid his second start in the Bigs to the tune of 8 runs in something like 2 innings. He’s brilliant in must win games! Ever heard of pulling a guy before he implodes when your season is riding on a game?

    • umrguy42 - Mar 1, 2014 at 3:51 PM

      Well, it *is* a limited model as you say (and they readily admit), but I just want to know why I couldn’t have thought of doing something like this for my class project back in grad school when I took a data mining/machine learning class. Would’ve been much more awesome than whatever obviously forgettable project I did wind up on 😀

  5. churchoftheperpetuallyoutraged - Feb 28, 2014 at 9:26 PM

    They used the first 80 percent of the games in those seasons to build a model of how pitchers fare over the course of a game, concentrating on Pitcher’s Total Bases (PTB) — an aggregate measure of hits and unintentional walks allowed — as the leading indicator of future performance. PTB, they note, is a more granular measure of pitcher performance than runs allowed.

    I was a little confused by what they meant, but the Scherzer example earlier in the article might have cleared it up. Basically, if a pitcher is performing poorly you should replace him earlier than usual. If a pitcher is doing well, you should replace him later than usual.

    Ground breaking work here…

    Oh and btw if you can’t trust your BP to not give up 4 runs over the last two innings in a playoff game, then maybe you should just stop playing.

    • historiophiliac - Feb 28, 2014 at 11:26 PM


      • js20011041 - Feb 28, 2014 at 11:33 PM

        It’s gone! It’s into the bullpen! This game is tied! This game is tied! David Ortiz! David Ortiz! David Ortiz!

      • historiophiliac - Feb 28, 2014 at 11:35 PM

        Stay classy, buddy.

    • NatsLady - Mar 1, 2014 at 7:21 AM


  6. markofapro - Feb 28, 2014 at 9:28 PM

    “Baseball is a red-blooded sport for red-blooded man. It’s no pink tea, and mollycoddles had better stay out. It’s a struggle for supremacy, survival of the fittest.”

    Stay away, nerds. You are not welcome.

    • churchoftheperpetuallyoutraged - Feb 28, 2014 at 9:32 PM

      2003 called, they want their argument(s) back.

  7. tfbuckfutter - Feb 28, 2014 at 9:29 PM

    So what you’re saying is….we need more RBIs?

    -Ruben Amaro’s analysis of this information

    • Kevin S. - Mar 1, 2014 at 9:34 AM

      AnalyticsPlant is ON IT.

  8. tanzkommandant - Feb 28, 2014 at 11:46 PM

    Black managers are traditionally poor at such decisions according to certain members of the SABR circle. No real mystery why this picture was chosen.

    • dinofrank60 - Mar 1, 2014 at 12:56 AM

      Seriously??? Some people would say that???

    • dan1111 - Mar 1, 2014 at 2:02 AM

      Dusty Baker is a particular manager who has often been criticized for his handling of pitchers. But don’t let that stop you from baseless, slanderous accusations of racism.

      • tanzkommandant - Mar 1, 2014 at 2:51 AM

        Libelous but thanks for playing.

    • themanytoolsofignorance - Mar 1, 2014 at 8:06 AM

      Hey. look at this jackass. People who say “the PC Police are ruining this country” need to pay attention. Jerks like this still exist and they feel perfectly fine with their petty racism.

      Thanks for letting us all know you and your ilk are still out there, jackass

      • tanzkommandant - Mar 1, 2014 at 8:42 AM

        If you disagree with someone’s statements, you should always respond in anger & insults rather than providing evidence why you disagree with their stance.

      • themanytoolsofignorance - Mar 1, 2014 at 9:09 AM

        You mean like you did? You waltzed out a stupid, racist statement, without a shred of fact and want ME to justify myself? GFY

      • js20011041 - Mar 1, 2014 at 8:54 AM

        That’s actually kind of rich considering you provided a baseless assertion that analytically minded people are racist.

    • paperlions - Mar 1, 2014 at 9:31 AM

      Feel free to provide even one fact to back this up. Just one. Otherwise, just shut up.

      • tanzkommandant - Mar 1, 2014 at 9:42 AM

        Prove that it is false. Burden of proof lies on the accusor.

      • paperlions - Mar 1, 2014 at 9:45 AM

        So, no facts then?

        BTW, you are the accuser, not me. You accused an entire industry of something for which you provided no facts. One. Just one fact, one quote, anything. Go ahead.

      • themanytoolsofignorance - Mar 1, 2014 at 9:45 AM

        This isn’t a court. Its the internet where when you assert something you can get called on to prove it. So prove it dummy. You can’t otherwise you would. Hence, you’re an idiot racist jackass

      • js20011041 - Mar 1, 2014 at 9:57 AM


        Not only have you made a completely baseless assertion, you’ve also got the burden of proof wrong. The burden of proof lies with the person making the positive assertion. To say otherwise would be like me making the assertion that unicorns exist and putting the burden of proof on you to prove me wrong. That’s not how logic and reason work.

    • paperlions - Mar 1, 2014 at 9:59 AM

      The worst part of this fabrication is that the last segment f the baseball population that is likely to generalize in this fashion are people that base opinions on data (i.e. the “SABR circle”), because people that form their opinions in such ways would not maintain such an opinion if the data showed it was untrue….compared to more “traditional” baseball people who are willing to maintain opinions in the face of mountains of evidence that they are wrong.

      • churchoftheperpetuallyoutraged - Mar 1, 2014 at 11:03 AM

        Also he confuses SABR (Society for American Baseball Research) with sabremetrics which is about statistics. The former is far more interested in studying that history of baseball, like disproving the Abner Doubleday myth, discussing the old rules of pre-1900 baseball, etc.

        They wouldn’t make any absurd claims like tanzfurher up there…

Leave Comment

You must be logged in to leave a comment. Not a member? Register now!

Top 10 MLB Player Searches
  1. J. Baez (2418)
  2. C. Correa (2382)
  3. H. Pence (2331)
  4. B. Crawford (2330)
  5. B. Harper (2127)