So How Many Starters Does a Team Need, Then?

Watching the Braves rotation grab appendages has been tough this spring. Kris Medlen has ligament damage in his elbow, Brandon Beachy has biceps soreness, and Mike Minor survived a scarred urethra only to encounter shoulder soreness. None of the three is a lock to make the opening day rotation. And this is a team that brought two veteran free agents in for depth and had extra youth at the back end of their rotation. They might be fine without Ervin Santana, but yet that team does inspire a question. How many major-league ready starting pitchers should a competitive team field in a given year?

Starting from a theoretical angle, the answer seems almost easy. Any given starter has a 40% chance of hitting the disabled list at any given time, so if they all happen to hit the disabled list at the same time, you’ll want nine starting pitchers — .4 times eight is 3.2 missing starters and we’re talking worst-case scenarios here. You might have a young team. Younger pitchers can have as little as a 35% chance of hitting the DL in a given year (every year adds 1%) — a young rotation might only need eight. Both of those numbers feel like a lot. It also seems very unlikely that one team would see three guys go down at one time, to say nothing about 3.2 or 2.4 guys.

Well, how about that likelihood of multiple injuries at once then. Thanks to Jeff Zimmerman‘s excellent work, we know that teams averaged 360 days lost to the DL from their starters, in six trips. Knowing that there will be six trips to the DL that average 60 days long, can we find out how likely it is that two (or more) starters will be on the DL at one time? Matt Murphy helped slice the data, and it looks like the following things are true of the average team in a given year, given the information above:

* There’s an 11% likelihood that two starters will be injured at the same time.
* There’s a 4% likelihood that three starters will be injured at the same time.
* Teams will need to use an eighth starter for a six games a season, on average.
* There’s even a 1% likelihood that *four* starters will be injured at the same time.
* Even if a team only suffers four 60-day DL stints, that team will only have a complete opening day rotation for 36 games (22% of the season).

Apparently I can’t read! The above is the case if only two starters are injured on a given team!! Here’s what the average team (six DL stints, 360 days) will suffer:

*10% likelihood that 4+ starters will be hurt at the same time (.082+.016+.001)
*32% likelihood that 3+ starters will be hurt (0.22 for exactly 3, plus the 4+ injured pitchers above)
*65% likelihood that at least 2 starters will be hurt at any given point in the season (105 games)

But we also know that year-long absences are skewing the data on some level. One 180-day stint from one pitcher is not the same as three 60-day stints. How about an empirical, retrospective angle?

We know from Jeff Sullivan’s excellent work that teams need about 32 starts from pitchers that weren’t in their top five. But how many people filled those posts? Answer: a lot. Since 2011, the average team has seen ten different pitchers start a game for them over the course of a single season.

The Rockies have needed 13 pitchers a season over the last three seasons, but they did try some different things with their rotation, so maybe they are an extreme case. It’s even more/less impressive that the Orioles have gone with a five-man rotation and needed more than 12 pitchers a year since the beginning of the 2011 season. But that might be about trying to find something that worked more than dealing with injury or depth concerns.

But it’s the Jays that show us a worst-case scenario when it comes to an injury-riddled staff. We know how many days they’ve lost to injury, and how they ranked in baseball last year, and now we know that they’ve needed, on average, 12.3 pitchers a year over the past three years.

The Braves had Mike Minor, Julio Teheran, Brandon Beachy, Kris Medlen, Alex Wood, David Hale, Gavin Floyd and Freddy Garcia lined up on their depth chart this year and it seemed like enough. Now it looks like it might not have been enough, even in a normal year where all the injuries didn’t happen at exactly the same time.

Print This Post

Graphs: Baseball, Roto, Beer, brats (OK, no graphs for that...yet), repeat. Follow him on Twitter @enosarris.

43 Responses to “So How Many Starters Does a Team Need, Then?”

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

    Thank God my Dbacks have Randall Delgado.

    Vote -1 Vote +1

  2. Iron says:

    Teams use 10 different starters in a season because a few times a year, inevitably, you don’t know until today that tomorrow’s starter isn’t going to be able to pitch and your best minor leaguer just pitched yesterday and the second best pitched today. So you dip to the #8 guy in your organization. Etc. etc. And in the scheme of things it probably didn’t much matter if 5-10 innings were pitched by #10 starter instead of #6 over the length of a season.

    This has little to do with the Braves situation. Santana is being brought in to pitch a lot of innings. The Braves obviously feel he is a significant upgrade from the bottom four names on that list. I don’t think it is that they think they need 8 or 9 or 10 guys. It’s that they weren’t comfortable giving lots of innings to whoever was #5 on that list.

    Vote -1 Vote +1

    • Iron says:

      Let me put it another way.

      If you take out the top six starting pitchers on every team, what is the WAR difference from the best team to the worst. Zero point small number?

      32 starts divided up between your top five backups, many of which go to guy #6, and which also includes September call-up starts. How much does #7 or #8 or #9 matter? Less than 0.1 WAR?

      Vote -1 Vote +1

  3. Bob says:

    What am I missing? 360 days of DL equals approximately two pitchers for one year. 32 missed starts is one season from a starter. Where does the extra 32 starts go?

    Vote -1 Vote +1

    • Eno Sarris says:

      The starters that started because one of the top five sucked! That’s why I’d stick with eight or nine competent starters as my number. The extra two starters are for teams that didn’t field enough competent starters and are auditioning/scrambling.

      Vote -1 Vote +1

      • Eno Sarris says:

        Oops, I see what you are saying. Well, the methodologies for these numbers are all different. I’ll think on why there’s that discrepancy.

        Vote -1 Vote +1

      • Matthew Murphy says:

        To get to the 32 number, Jeff just looked at the five pitchers who started the most games on each team. In a number of cases, the #6 or 7 guys will make more starts than some of the guys who opened the season in the rotation. Also, this doesn’t take into account pitchers that suffer major injuries before the season starts (Harvey, Luebke, Medlen, to name a few recent examples). Lastly, because of days off, teams can skip their worst starter once every 2-3 times through the rotation if they want, so the number of starts given to the emergency starters will likely be lower than expected given the number of days missed.

        Vote -1 Vote +1

  4. randplaty says:

    So the answer is… 7-8?

    Vote -1 Vote +1

  5. Hurtlockertwo says:

    “Mike Minor survived a scarred urethra” What??? That just isn’t good.

    +7 Vote -1 Vote +1

  6. Mr Punch says:

    A further question is to what extent depth does much good. I can think of a number of cases in which teams with unusually healthy rotations have been successful – I’d ask how often a team succeeds despite an average or above-average injury rate. Or to put it another way: Do teams that use a 7th starter 36 times ever win?

    Vote -1 Vote +1

    • Pirates Hurdles says:

      Bucs used 12 SP and 9 made at least 5 starts last year on the way to winning 94. Its not just simply a function of injuries, its also the quality of the depth.

      Vote -1 Vote +1

  7. jfree says:

    Why not look backwards to find out? To the time when MLB teams actually had to carry most of their season starters on their own roster because minor league ball wasn’t purely a puppet for the MLB team.

    example 1955 (random year) – supposedly the ‘four-man rotation’ (when men were men and metrosexuals hadn’t yet introduced pitch counts and other nambypamby things). That year, the NYY had two pitchers with more than 30 starts (the only two with more than 200 IP as well), one with more than 20, four with more than 10, and two with more than 5. Every pitcher who started also relieved at least one game – and six pitchers started and relieved more than 5 games. Over in the NL, Brooklyn had virtually the same stats – one pitcher with more than 30 starts, two with more than 20 (one at 29), four with more than 10, two with more than 5. Every pitcher who started also relieved at least once – and five pitchers started and relieved more than 5 games.

    Seems to me that the answer to – ‘where do you find the extra starts?’ – is ‘in the bullpen’. The excess of roster slots devoted to one-inning-max relievers is the source of the problem here. And I suspect that this Berlin Wall that we’ve erected between ‘rotation’ and ‘bullpen’ is also the source of the excess injuries. We expect either too many IP’s (general durability) from ‘THE rotation’ or too many season appearances (muscle recovery time) from the bullpen.

    Vote -1 Vote +1

  8. Llewdor says:

    5 – 2003 Mariners

    Vote -1 Vote +1

  9. Baltar says:

    IIRC, the Rays had their #2 thru #5 starters out at the same time last season. (The #1 also was out for a fairly lengthy time.)
    The irony is that they won a higher percentage of their games when the regular starters were out.
    So the answer to the title question is “at least 9, if you want to contend.”

    Vote -1 Vote +1

  10. David says:

    I’d say you better have two full rotations. At one point last season the Pirates entire expected rotation entering spring training (Burentt/Rodriguez/McDonald/Karstens/McPherson) was injured (not on the DL as McPherson was in the minors and injured).

    Vote -1 Vote +1

  11. Erik says:

    The Angels are really asking for it this year. There are only 6 guys I’m comfortable giving ANY starts to right now.

    Vote -1 Vote +1

  12. SABRphreak says:

    If it’s 360 DL days spread over 6 pitchers (60 days each) – representing the starting 5 plus one replacement – potentially equals the need for 11 pitchers available at the same time. This is a near worst case senario. Also, keep in mind that the replacement starts are not given to the same replacement pitcher(s) due to a lack of effectiveness, developmental concerns (e.g. a minor league call up) and/or durability concerns (e.g. relief pitcher that makes a spot start)… which does not include the possibility of additional injuries. Most teams have an SP depth chart of around 10 starters at any given time (some stashed in the pen or triple A). Clearly, Santana was a superior option than going further down this chart for the Braves. Most teams need extra starters to replace ineffectiveness in addition to injury. My initial glance at some data indicates most teams give substantial starts to 6 or 7 pitchers with a couple additional ones to a few others. Again, this varies depending on the needs of the team and their luck staying relatively free of najor injuries. GMs know they need a 6th starter (a/k/a “long man”) in the pen and someone ready to be called up from AAA if needed. The cold weather at the start of the season tends to create issues/needs as well, thus the need for extra depth during Spring Training.

    Vote -1 Vote +1

    • Matthew Murphy says:

      The likelihood of actually needing the 10th or 11th guy is pretty low (the average team should spend 3 days during the season needing more than 9 starters). The 6th and 7th starters should get lots of use, as the updated numbers show that teams should spend about 2/3 of the season with at least two injured starters. Once you get to the 8th guy, you’re still looking at significant portion of the season, but the ability to skip him when a day off if he’s really bad might mean that he only makes 7 or 8 starts on average, and 7-8 starts of a replacement player versus a league-average starter is only a fraction of a win.

      Vote -1 Vote +1

  13. Josh says:

    Appendages, I see what you did there…

    Vote -1 Vote +1

  14. Yehuda Hamer says:

    As long as you can beat up on the bad teams, four starters should suffice.

    Vote -1 Vote +1

  15. Vince says:

    “* There‚Äôs an 11% likelihood that two starters will be injured at the same time.
    * Teams will need to use a seventh starter for an 36 games a season, on average.”

    36 starts would lead the league most years. Does he mean that the team will be missing 2 guys from the rotation for 36 days and actually start the 7th guy ~5 times?

    Vote -1 Vote +1

    • Matthew Murphy says:

      I think the 37 may have been meant to be total number of starts from the guys 7th or lower on the depth chart (obviously not #7 guy alone). Regularly scheduled off days might drop this number down a bit closer to 30, depending on timing.

      Vote -1 Vote +1

  16. Damaso's Burnt Shirt says:

    All of them.

    Vote -1 Vote +1

  17. Jason B says:

    It’s like last year when the Dodgers entered the season with like 8 very reasonable rotation candidates I think? Then Billingsley got hurt, Beckett got hurt, Harang did something or ‘nother, Ted Lilly turned to dust, etc, etc.

    You can never really have TOO many.

    Vote -1 Vote +1

  18. FeslenR says:

    Answer? Can never have enough pitching depth

    Vote -1 Vote +1

  19. tz says:

    I say if you’re going to keep a bazillion guys in your bullpen, why not have the two guys furthest back in the pen pecking order fill in for injured starters using a 3 IP + 3 IP strategy?

    If they’re worth having on your roster to begin with, maybe you should just trust them to actually pitch meaningful innings, instead of just blindly calling the Livan Hernandez hotline whenever a starter goes down.

    Vote -1 Vote +1

  20. stan says:

    When I was young teams generally carried 4 starters and had a 5th guy to act as a swingman. Now it seems wise for them to carry 5 starters and always have a 6th guy on the major league roster to act as a swingman.

    Vote -1 Vote +1

  21. Nick says:

    I think SP depth is a little bit of an inefficiency – teams are still reluctant to sign/trade for a 4th or 5th guy when they have a young prospect #5 that they think can do well. Nationals are a good example this offseason – the fact that they already had 4 great starters doesn’t change the fact that signing Santana or Burnett would have provided a swing of 2-3 wins over their #5 options, not to mention the value that’s added with depth. I feel like a lot of teams see their rotation as one unit, rather than as 5 discrete parts that can each be improved.

    Vote -1 Vote +1

  22. John Lannan says:


    Vote -1 Vote +1

  23. Snowman says:

    So let’s see:

    It’s looking more and more likely that both Medlen and Beachy will need second TJ surgeries and will miss the season (and along with Venters, this is three 2-TJ pitchers for the Braves in under three seasons). Minor will miss at least the first two weeks of the season with his penile shoulder injury.

    So they’ll start the season with three starters on the DL, and Wood, Teheran, Santana, Freddy Garcia, and probably David Hale as their 5-man rotation. On opening day, they’ll already be looking at having their 8th starter in the rotation (Hale).

    There really is just no such thing as too much starting pitching.

    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>

Current ye@r *