Fifth Starters Don’t Exist

We read a lot of scouting reports and hear people comment that a pitcher’s potential is that of a No. 5 starter. Teams spend millions of dollars and thousands of hours crunching data to build a successful five-man rotation, but it’s all in vain. The truth of the matter is that these mythical creatures don’t actually exist.

If we look back to the 2009 season, only two teams had five starters on their pitching staffs that made 24 or more starts: the Chicago Cubs and the Colorado Rockies.

  • All 30 teams had at least one pitcher make 24 or more starts.
  • Twenty-six teams had two pitchers make 24 or more starts.
  • Then the number drops to 22 teams that had three pitchers make 24 or more starts.
  • Then we hit a cliff. Only nine teams were able to rely on four pitchers to make 24 or more starts.

    This may not be the most scientific way to look at the situation, but it’s quick and dirty and gets the point across. Only about a third of the teams in Major League Baseball had four reliable starters. Less than 10% of teams could make a claim that they actually had a “No. 5 starter” last season. Sure, you could say that some of these teams had another “quality guy” earmarked for the No. 5 role but injuries created the gap. But we know injuries in the starting rotation are inevitable each season, so it makes sense to start planning for that likely scenario. San Diego, Washington, Seattle and Cleveland could not even claim to have a No. 2 starter. Each of those four organizations had just one starter that made more than 24 starts on the season.

    Every season, most of the teams in baseball scramble to fill holes in their starting rotations and most of the headaches come from trying to fill the gaping hole in the fifth spot in the rotation. In ’09, seven teams used 10 or more pitchers to fill the black holes in their starting rotations.

    Tomorrow, I will suggest a new approach that some teams might want to consider for their starting rotations.

    * * *

    For interest’s sake, here are some other articles discussing No. 5 starters:
    1. A classic by Jeff Sackman from December 2006
    2. R.J. Anderson touching on the subject at FanGraphs
    3. FanGrapher Matthew Carruth discussing the issue at Lookout Landing
    4. Chuck Brownson tying it all together at The Hardball Times

    I’m sure there are other interesting articles on the subject… and these were just the tip of the iceberg that I discovered.




  • Print This Post



    Marc Hulet has been writing at FanGraphs since 2008. His work focuses on prospect analysis. Follow him on Twitter @marchulet.


    20 Responses to “Fifth Starters Don’t Exist”

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

      The “5th starter” label is almost as bad as “oh he’s the #3 starter so he’ll be facing #3 starters all year long” comments

      Vote -1 Vote +1

    2. grandbranyan says:

      I would love to see the Brewers go with a piggyback approach to their fifth spot in 2011 with Jeffress and Rogers. Depending how things go for them this year they could both be in line for about 120ish IP by then without too much concern and getting them both 32 games of 3 to 4 IP would get you there with a lot more upside than your traditional 5th starter fodder.

      Vote -1 Vote +1

      • Jack Straw says:

        This would make sense for the Mets, too. Neither Maine nor Niese should be pushed this season. Finding a combination that allows them both to get 120 innings with 150 as the top makes a lot of sense. Unfortunately, the FO didn’t put together a rotation of four other starters that would allow them to try this.

        Vote -1 Vote +1

    3. stevej says:

      I came of age in Baltimore when Earl Weaver was the last 4 man rotation holdout. His brain told him intuitively what we quants learn directly. I recommend to all to read his book Weaver On Strategy, and rule #7: “It is easier to find four starting pitchers than five.”
      If he were a new manager today, he probably couldn’t pull off a 4 man. My guess is he would make a rigid every-5th-day routine for his top 3, and fill in with a 5th only when the schedule demands it. That would be 35 to 36 starts for the top 3, 50 or so innings shifted from replacement level pitchers to quality pitchers.

      Vote -1 Vote +1

      • Charles says:

        The every 5th day routine is pretty tough, especially for 3 or 4 guys. It’s going to work sometimes, but sometimes that off day is going to come on your #1’s 5th day. You might be able to squeak out another start or two, but you’re not getting your top 4 guys 36 starts each

        Vote -1 Vote +1

      • Steven Gomez says:

        It’s possible to get your top guys 36 starts, but not likely. Injuries, off days and multiple factors will often delay starts for most pitchers.

        I don’t necessarily think this phenomenon calls for a hard wired strategy a la Weaver’s 4 man rotation. But it does indicate a team struggling to find five quality starters could get away with four set SPs, and perhaps rotate the 5th spot between multiple swingmen, long men, AAAA men, etc.men.

        Vote -1 Vote +1

    4. CJ says:

      The “5th starter” label means something if we are talking about the ceiling or likely performance level of a pitcher, i.e., basically meaning that they may have enough pitches to be a starter, but they aren’t likely to be very good). What is the probability that an average starter in the major leagues will have an injury which sidelines them for a substantial amount of time? I wouldn’t be surprised if it is close to 20%. (If someone can reference a study along those lines, please do.) And that would explain a lot about the big drop off in IP for individual 5th starters.

      Vote -1 Vote +1

    5. Chris in Dallas says:

      “Fifth starter” is a euphemism for the back end of the rotation as a whole. Like a “middle of the order hitter” means a 3-5 hitter. Teams break camp with 5 starters. Somebody has to go last.

      Vote -1 Vote +1

    6. BD says:

      It’s definitely true that the term “5th starter” speaks to the qaulity of the pitcher, not the number of starts he is likely to make. Same goes for all of the rotation slots. Just because Johan Santana gets injured and makes only 23 starts doesn’t mean he isn’t a “no. 1 starter.”

      The real myth still requiring a good busting relates to the quality of pitching deserving of the “no. 5″ label. In the AL, the average starting pitching ERA is something like 4.50. Yet, people persist in thinking of a pitcher with a 4.50 ERA as basically a no. 5. That pitcher is really more like a no. 3. The overall depth of pitching just isn’t nearly as great, statistically, as people seem to imagine.

      Another point, related to the subject of the OP, is simply the number of starters used by each team in a season. I would imagine teams use, an average, about ten different starters a year. Therefore, the preseason emphasis on having a solid 5-man rotation is a bit exaggerated. Ideally, teams will have at least 7 fairly solid guys available (including in AAA).

      Vote -1 Vote +1

      • Jack Straw says:

        Agreed. Even on sites that draw more informed fans it can be nearly impossible to have an intelligent discussion relating to the quality of starting pitching. As a Mets fan I endured a whole lot of “Pineiro sucks, he’s a back of the rotation guy”, and “Marquis is garbage! CHONE projects him at 192 innings with a 4.45 ERA, so what the hell do we need another fifth starter for???”

        Vote -1 Vote +1

    7. Mike Solomon says:

      I do not think you are framing your argument well. I think what you might be doing is perhaps explaining the importance of your 6th-10th options for a rotation. Saying that there is no such thing as a 5th starter is the same as saying there is no such thing as a pitching prospects. There is a kernel of truth there wrapped in a ton of overstatement.

      Vote -1 Vote +1

    8. Marc Hulet says:

      This is certainly not meant to be any earth-shattering statement, as mentioned in the article, it was a quick and easy way to state the basis of my discuss. The meat of the topic will post tomorrow.

      Vote -1 Vote +1

      • rotofan says:

        I’m curious to see tomorrow’s post, your premise being that teams should stop chasing the myth of a fifth starter and find better way to allocate innings.

        I’m not sure what you will propose, but it seems as a starting point you might begin with four starters. How then, should a team use its pitchers when those four starters aren’t available?

        Rather than use what end up being a series of failed fifth starters, team could instead start with what might be a strength, a group of middle relievers who can pitch well for two or three innings at a time, basically once through the lineup. Then plan on using two or three of them to get you to the eight inning and hand off to the setup man and closer.

        The downside would be to put more strain on middle relievers, but perhaps that is a strain that is shared among enough relievers as to not make it overly burdensome.

        There would be some other benefits to the approach:

        (1) Managers could mix left and right handed relievers to force opponents to use up bats on-the-bench before the late innings.
        (2) Managers could mix up hard and soft throwers to keep batters off-balance.
        (3) Middle relievers might feel a greater stake in what happens as their responsibility expands.

        Vote -1 Vote +1

    9. Rob says:

      If by “quick and dirty” you mean “meaningless and wrong” sure. Doug Davis made 34 starts for the DBacks last season, Brandon Webb 1.

      Vote -1 Vote +1

    10. logan3825 says:

      I tried something similar to the Jeff Sackman article from December 2006. I came up with these numbers for 2009. I used starters splits only.
      NL
      ERA/Innings per start
      3.19/6.3
      3.77/6.1
      4.24/5.8
      4.73/5.7
      5.65/5.2

      AL
      ERA/Innings per start
      3.27/6.4
      3.98/6.2
      4.47/5.9
      5.08/5.5
      6.47/5.1

      MLB
      ERA/Innings per start
      3.23/6.4
      3.87/6.2
      4.34/5.9
      4.89/5.6
      6.03/5.2

      Vote -1 Vote +1

    11. Ted Hoppe says:

      I’m incline to look at the innings pitched as a factor.
      How many pitchers in your rotation are giving you 200+ innings? 36 pitchers hit that mark last season
      If you translate this into starts, 6 innings per start X 36 = 216. Only 15 pitchers throw more than 216 innings last season

      …and this doesn’t count the post season games.

      So this makes the question, do fourth starters really exits? Because fifth starter roles are usually filled by committee.

      Vote -1 Vote +1

    12. Eirias says:

      I disagree. If fifth starters didn’t exist, Yankees beat writers would be out of a job. Well, most would be. Some would need to hang around for teh 8th.

      Vote -1 Vote +1

    13. B N says:

      I’m not quite sure how number of starts determines the 5th starter. So… if halfway through the season, your #1 guy goes down, and you get a full season worth of starts out of your 5th guy, plus a full half season out of the 6th guy brought in… you had no 5th starter? Whaaa?

      The issue with this methodology is not injuries. It’s that the distribution of injuries affects it so much. If all your starters had ailments that caused them to make just 23 starts, you’d have no starters at all! (according to this). Likewise, losing a top starter to a season-ender early in the season will make it seem like you had 5 guys all along (if the others are healthy). Losing the same guy midway will make it seem like you only had 3 starters, because both he AND his replacement will only get a half season’s worth of starts.

      I mean, look at the Red Sox. They probably don’t have a 5th starter last season according to this. Untrue. They could have had 2 or 3 options for that slot, if they wanted. Penny could have done it all season. Buchholz could have. They were going for the bigger-better pitcher, instead. So they traded Penny, and put in Smoltz. When Smoltz didn’t fare well, go to your young gun. While in truth, Penny and Buchholz were reasonable 5th starters for a whole season. For some clubs, these guys would have been in the 4rth slot all season. Turnover doesn’t mean derth- it can mean surplus also.

      The issue is not just about filling holes. It’s about always seeking improvement. If the Brewers, during their playoff run, had 5 guys who did 23 starts and then trade for Sabathia and push a guy out of the rotation- they no longer had a 5th starter? No. They now have 2 of them, but one isn’t playing. But this won’t show that.

      Vote -1 Vote +1

    14. Dave S. says:

      Fifth starter is a state of mind. It is a baseball abstraction of major-league tightrope mediocrity.

      Vote -1 Vote +1

    15. walkoffblast says:

      Interestingly enough one of the two teams with a “5th starter,” the Cubs, had a worse record than the Mariners even though they could not claim to have even a number 2 starter.

      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>