A New Approach to the Fifth Starter

As we saw yesterday in my introductory post, there is really is no point in wasting time and resources trying to enter the season with five starters earmarked for 25+ starts. No. 5 starters, in the traditional sense, just don’t exist. The best bet is to focus on securing four starters that can make 24 starts or more. In the fifth spot in the rotation, a three-man job-share could then be developed and it would break down like this:

1. A long reliever who would serve as the seventh arm in the ‘pen and be expected to make eight to 10 starts on the year. Ideally, this would be a proven veteran who could stick at the MLB level all season.

2. A pitching prospect that projects to be a fringe No. 3 or 4 with two or three minor league options remaining. He would be introduced to the Majors in this low-pressure role over the next two to three seasons before officially (hopefully) graduating to the role of a reliable third or fourth starter. In this role, the pitcher would need to make about 10 starts at the MLB level each season.

3. A minor league “veteran” pitcher (somewhere in the 25-30 year old range) who has been unable to stick in the Majors – and still has at least one minor league option left – and can be relied on to make at least five starts on the season.

Let’s use a real team (The Toronto Blue Jays) to flesh out this example.
No. 1 starter: Ricky Romero
No. 2 starter: Shaun Marcum
No. 3 starter: Marc Rzepczynski
No. 4 starter: Brett Cecil

The job-share in the fifth spot would include:
1. Brian Tallet, as the long-man (0 options)
2. Brad Mills, as the prospect (2 options)
3. Lance Broadway, as the minor-league vet (1 option)

If all goes well, we can expect this group to make 23 starts out of the fifth spot, which more than most teams can hope for… and it leaves a little room to exceed expectation if one pitcher performs better than expected. But we should also have a safeguard because injuries and other unexpected situations always arise. In triple-A, Toronto would have a reliever capable of going 3-4 innings (Sean Stidfole, 3 options) to replace Tallet, along with two emergency starters that could fill in to replace Mills (Rey Gonzalez, 3 options) and Broadway (Randy Boone, 3 options).

This may seem like a lot of moving parts or a lot of resources to tie up in this situation but we know that a team is going to struggle to find five reliable starters each season (four will be tough to find for some), so it makes sense to plan ahead. Seven teams used 10 or more starters each that made 23 or fewer starts last season. Most of those pitchers will filling in the gap in the fifth hole in the rotation.

The best part of the Jays plan above, is the cost efficiency. Tallet will make $2 million in 2010 for his role as a spot starter and long man out of the bullpen. Mills and Broadway would make just over the minimum MLB salary when in the Majors, while the other three pitchers – if needed – would make exactly the minimum. In other words, you’d be filling a minimum of 23 starts on the year for less than $3 million. As well, by ensuring that everyone but Tallet has options remaining, the club will not risk losing any of these players in 2010.

Obviously this plan is not going to score a team 20 quality starts, so the goal would be to expect a league-average perform from the job-share when the 23 starts are averaged out. Even so, that’s solid value for less than $3 million. And if a significant injury strikes Tallet, Mills or Broadway, it’s not a catastrophic situation; you can have the other two pitchers pick up the slack, or bring up one of your back-ups. If you look at some of the pitching performances received from fill-in starters or supposed “No. 5 starters,” you find that the performances are most often well below league average. Scrambling to fill starts can also result in throwing millions of dollars at pitchers like Sidney Ponson to make five or six disappointing starts before a release is mercifully made.

So, to recap… This job-share plan is good because…
A) The inevitable pitching injuries will have a lesser (negative) impact
B) It will help train young pitchers for an eventual larger role
C) It’s cost efficient

It’s at least worth a try because we know the currently model is flawed for most – if not all – clubs.

Print This Post

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

52 Responses to “A New Approach to the Fifth Starter”

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

    I would add one other category to that: starting pitcher recovering from injury, such as Tim Hudson last season, or numerous other guys this year (Jordan Zimmermann) who will spell some innings in September.

    Obviously, these guys have the talent to be top of the rotation guys, but by the time that they come off the DL, the rotation will likely be set, and their recovery regiment will curb their innings.

    Vote -1 Vote +1

  2. Tom B says:

    “Let’s use a real team”

    … and then he picks the Blue Jays?

    Vote -1 Vote +1

    • Big Jgke says:

      OH SNAP.

      But seriously, I think this plan would only really be acceptable to rebuilding and/or low income teams. The big boys have an expectation, it seems, from both fans and management to have that mythical 5th starter. Look at the rotations of the big-money teams and you can see that, even the mediocre contenders like St. Louis and the White Sox, have one guy set to make all the #5 starts.

      This seems like the closer vs. high leverage situation reliever debate. It’s clear that a more nuanced approach would be beneficial, but there just is no reason to believe that baseball orthodoxy would ever support such a move.

      Vote -1 Vote +1

  3. snapper says:

    I think this is an improvement over the current practice, but there is also another option that is worth considering.

    Going back to the 70’s, most relivers besides your “closer” and 1-2 setup-men, were basically “swingmen”. Failed or young starters who found themselves in the pen.

    If you carry 3 “swingmen” (plus closer and 3 short-RPs), you can patch together the 5th spot based on matchups, and also have the capacity to avoid using every RP on the team when an SP gets knocked out in the 3rd inning.

    Your AAA SPs would also serve as reserve “swingmen”.

    Vote -1 Vote +1

    • lincolndude says:

      Isn’t this basically what teams do right now?

      Vote -1 Vote +1

      • MikeS says:

        Yes, but teams still don’t expect them to go more than an inning or two, so when a spot start is needed they aren’t “Stretched out.”

        But yeah, there’s basically three kinds of pitchers:
        1) Starters
        2) Bad starters that became good relievers and pitch the 7th, 8th and 9th with a lead.
        3) Bad starters that became bad relievers and eat up innings in blowouts.

        Even today, very few guys get drafted as a reliever.

        Vote -1 Vote +1

  4. Adam says:

    If a team went through with this plan, how many pitchers would you have on the 25-man roster? I’m guessing 12, with the 4 SP’s, 5 RP’s, and this three-headed relief / #5 SP group; am I assuming correctly?

    Vote -1 Vote +1

  5. BD says:

    This is a classic case of a solution in search of a problem.

    This whole discussion is based on a flawed premise, i.e., that not having five 24+ start pitchers at the end of the season means that the so-called “5th starter” represents some kind of blight on the team that must be replaced. This is so illogical as to be laughable. As was pointed out in response to your previous post, a team’s “5th starter” ISN’T, by definition, the pitcher who provides the 5th most number of starts. “5th starter” refers to QUALITY, not QUANTITY. If teams are having trouble getting enough starts out of their rotation, the problem isn’t a failure of “5th starters,” per se, to take the ball; it’s just the normal attrition that occurs THROUGHOUT THE ENTIRE ROTATION. And the solution to such lack of starts is not to get rid of one starter and try to make up the difference with relievers and fringe players called up from AAA. (If anything, it would make more sense to ADD a starter in that situation.) Finally, it’s obviously senseless for a team to eliminate its 5th-best pitcher in favor of a committee comprised of perhaps its 9-10-11th best, which is what you are advocating. Take the Yankees, for example. According to your theory, due to the risk that ONE of their 5 starters won’t make 24+ starts, they should release/trade Phil Hughes now and try to make up his projected starts with a committee of long relievers and AAAA pitchers. What kind of nonsense is that?

    Please check your calendar. It’s not April 1 just yet.

    Vote -1 Vote +1

    • Bob says:

      Best post I’ve read in a week.

      Vote -1 Vote +1

    • Dan Lewis says:

      At first, I would have agreed, but I think there’s an odd factual correlation which suggests that while you are correct in theory, you are not in practice. Simply: The guys who, qualitatively (by pre-season perception at least), are 5th starters are the ones who quantitatively, are falling short of the 24 start plateau.

      And the reason for this is, as Marc states it, because it’s really hard to find five guys who, if healthy, are deserving of 24+ starts.

      Vote -1 Vote +1

  6. rotofan says:

    (1) Last time I checked Gaston had Morrow, not Cecil, tapped to start.
    (2) Last year the Jays essentially cycled through prospects, which is also cost-efficient, and doing so turned up a guy you now pencil in at #3, Rzepczynski.
    (3) I would agree it doesn’t make sense for most team to invest a lot in a 5th starter, but might make an exception for teams with deep pockets where each marginal win is critical because they are on the bubble of winning a division or wild card.
    (4) How do you classify a signing like Boston’s of John Lackey? Lackey is clearly an ace on many teams and the second best or third best on the Red Sox but his signing also has the effect of bumping Wakefield to fifth best with Clay Buckholz ready to step in if Dice K isn’t healthy or someone isn’t productive. Isn’t Lackey’s signing predicated on the belief in Boston that they will go with a five-man rotation? And in Boston’s circumstance competing with with Yanks and Rays, does that make sense?

    Vote -1 Vote +1

  7. dave says:

    This post doesn’t really make sense. Each team begins the season with 5 starters. When one gets injured, they bring up a player from the minors or promote someone from the bullpen. You write as if finding four guys who won’t get injured is easy but the fifth guy will almost certainly go down at some point.

    What if Tallet doesn’t get injured? Then he’s a number 5 starter, right? He’d play the whole season, make 24 starts and be respectable, but below average. Sounds like a 5th starter to me.

    In your first post when you said that only 2 teams had 5 starters make 24 or more starts, you don’t take into account who was making the starts for the teams with less than 4. Braden Looper made 34 starts last year and had a -0.9 WAR. Sounds like a number 5 starter to me, yet the Brewers didn’t have 5 guys make 24 or more starts.

    It seems like this critique of the #5 starter still needs work. Obviously, you want good replacement players if your starters get hurt, but you also want to start the season with the best 5 starters possible. Period.

    Vote -1 Vote +1

  8. Matt S. says:

    I think there are already teams out there using a similar approach. (the Red Sox, Rays and Phillies come to mind) RD makes a good point about how the tag “5th Starter” refers to quality, but shouldn’t the issues of quality and quantity be linked. Giving one start every 5 to 6 days to a pitcher who is ineffective is a bad idea. In RD’s example the Yankees would not be cutting Phil Hughes, but rather slating Hughes in for the 10-15 starts, barring injuries to the guys above him and then a possible return to a bullpen role given ineffectiveness/ fatigue/ inning limits (he is unlikely to truly be able handle 25 starts given his experience) Aceves would then look to fill the second 5th starter slot, though in this case he would be less likely to go to the minors as he is effective in the pen. The final slot would be filled by a top prospect, Joba or, since this is the Yankees, an acquisition. That is very nearly what happen last season anyway.

    Vote -1 Vote +1

  9. Charles says:

    And what if a significant injury strikes your top 4 instead of your 3 headed monster? Not to mention you expect league-average performance in sum out of 3 parts that aren’t expected to be league-average? How? Why not just give most of the innings to the one guy out of three you expect to perform the best, and let the other 2 fill in where needed?

    Vote -1 Vote +1

  10. by jiminy says:

    I don’t quite agree with your reasoning.

    Yesterday you said most teams only get 24 starts out of four pitchers, so the fifth starter is not important. But that’s just because at least one of the top five is likely to get injured. And it’s very unlikely that the injury will be kind enough to strike your fifth best starter. In fact, it probably won’t.

    That doesn’t mean your fifth best starter isn’t important. It means he’s even more important, because he’s now one of your top four starters.

    So I think you’re misreading your statistics to conclude that your fifth best starter isn’t important. Chances are he will make more than 24 starts.

    If you know you’re probably going to lose at least one pitcher, that means you need six decent options, not four. If you decide, only four starters are going to make 24 starts, so I only need four solid pitchers, when the inevitable injury strikes, you only have three.

    Someone isn’t going to make it through the season. But if you define your fifth starter after the fact as “the guy who got hurt,” or flames out for some other reason, then sure, you’d have been better off without him. But no one ever knows in advance who’s going down.

    Vote -1 Vote +1

  11. Bradley says:

    I see this being one of those clubhouse-prohibitive moves, like batting the pitcher 8th. It makes good sense, but it hurts feelings, so managers would never do it.

    Vote -1 Vote +1

    • walkoffblast says:

      I have a similar line of thought but I do not think it is hurting feelings that is the problem. I think the problem is that sliding guys in and out of the rotation, bullpen, minors etc. is going to result in them putting up starts that are not to their full ability. In a video game or on paper this is a great idea, but MLB players are a little more finicky than this and even teams like the red sox sacrifice some of the good ideas by the numbers in the name of consistency and comfort for their players.

      Like some people mentioned yesterday all this really points out is the importance of viable depth at starting pitching. If you only sign 5 guys that are remotely usable at sp that would be a bad thing. However, I am pretty sure all teams know this anyway.

      Vote -1 Vote +1

    • Jason B says:

      Batting the pitcher 8th makes good sense?

      Vote -1 Vote +1

  12. BD says:

    “Giving one start every 5 to 6 days to a pitcher who is ineffective is a bad idea.”

    Well, the OP is advocating giving starts to “a minor league ‘veteran’ pitcher (somewhere in the 25-30 year old range) who has been unable to stick in the Majors.”

    “In RD’s example the Yankees would not be cutting Phil Hughes, but rather slating Hughes in for the 10-15 starts, barring injuries to the guys above him and then a possible return to a bullpen role given ineffectiveness/ fatigue/ inning limits (he is unlikely to truly be able handle 25 starts given his experience) Aceves would then look to fill the second 5th starter slot, though in this case he would be less likely to go to the minors as he is effective in the pen. The final slot would be filled by a top prospect, Joba or, since this is the Yankees, an acquisition.”

    Sorry, but I fail to see how this makes any sense. Why would the Yankees only “slate” Hughes for 10-15 starts rather than 30 (apart from a possible innings restriction that is beyond the scope of the OP)? If the Yankees, or any team, can get 5 quality MLB pitchers who are each theoretically capable of making 30-35 starts, why should they arbitrarily get rid of one of them in favor of using a troika of lesser pitchers to make those starts. It defies all common sense.

    Vote -1 Vote +1

  13. Tyler says:

    Given that you’re likely to expect a lot of poor to mediocre/average performances, it wouldn’t hurt to also use your 5th slot (at least in the first half of the season) to evaluate pitching prospects. It’s almost worth it to not commit to 3 specific pitchers for your “job share” and instead just rotate in and out whoever you please, with one or two guys getting the major share of those starts while you have a look at who has what and might deserve a longer look at another time.

    And, of course, for giving starts to guys coming back from injury. The noticeably-absent Dustin McGowan comes to mind; what happens in May or June when the Jays start really thinking about whether he’s healthy enough to pitch, right? He’s not going to jump in as our ace or #2 or whatever, and we’ll want to minimize the impact of him flaming out, if it happens.

    And what about someone like Henderson Alvarez, or Kyle Drabek and so forth? Guys you might want to bring up for a week or two (or even in September, since we won’t be making the wild card) and our roster expands again?

    The general notion that a team should focus on 4 starters as a true core (or even 3, really) is pretty solid, but that 5th slot should be a little more flexible than just 2 or 3 guys who can give you decent innings, IMO.

    Vote -1 Vote +1

  14. BD says:

    Come to think of it, this is a great idea whose underlying wisdom can be applied in other areas of life. Statistics show that automobile tires occasionally blow out or become flat, such that, over the course of a cross-country trip, there’s a good chance that not all four tires will last the entire journey. Clearly, the “classic model” of a four-tire automobile is “flawed.”

    To solve this, let’s simply remove the fourth tire before embarking on the trip. Just get rid of it completely. In its place, let’s use a combination of a bicycle tire, a hula hoop, and an especially large bagel. This makes sense because the fourth tire was unlikely to last the whole trip anyway. Also, it’s cheaper. And, as an added bonus, we get to evaluate how well a bagel performs at highway speeds!

    +19 Vote -1 Vote +1

  15. Tyler says:

    I think that’s missing the point, BD.

    I think the general notion is that you shouldn’t overpay for a 5th starter because you can find equivalent value at a lower cost than signing, say, an over-the-hill guy who used to be good or spending a lot on a more mature talent to fill that slot when long relievers can be stretched out well enough to give you 5 or 6 solid innings and keep you in the game. Consistent 5th starters aren’t common and, your hyperbole about blown tires aside, most teams don’t have a consistent 5th starter, as the study showed.

    There’s got to be a better way to arrange the rotation than to sink money into a guy who’s going to end up not starting 25+ games for you. There are certainly exceptions, the Red Sox are a great one, but take Toronto for example. If we HAD a guy who was good for 25 or 30 starts and pitched really well, he’d shuffle the entire rotation so that he’d be a 3rd or 4th starter, because that’s what Toronto’s rotation needs. If McGowan comes back and stays healthy, he’s instantly one of the top 3 starters, which totally invalidates the need for a committee at the 5th slot. But if he does, it’ll also make true the notion that we don’t have a specific 5th starter, because some guys are going to pitch while he’s out, then get shuffled out if/when he comes back and for the rest of the season, different guys will be pitching their because of the talent shuffle. And because it’s damned expensive to keep 5 starting-caliber pitchers on your roster unless you’re Boston or New York, who have all the money they desire at this point. The Yanks, for example, can throw Sabathia, Burnett, Pettitte and even Joba out there. Last year, they didn’t have a 5th guy who really stacked up. They had 4 guys with 30+ starts and NO ONE else with 10+ starts. Mitre and Wang started 9, Hughes started 7, Gaudin started 6, Aceves started 1 and no one else on the team had a start.

    And they won the World Series. Yeah, they had epic offense and the biggest payroll in the league, so it may not be a strategy all teams can follow, but the idea that the 5th starter is this critical piece is misleading. They had 5 guys rotating in and out of the 5th slot taking up those last 32 starts and it didn’t affect them a bit, even though Wang was awful, because the other guys basically pitched to about .500.

    The BoSox went a similar route. With Dice-K pitching in only 12 games, they tried to add Penny and Smoltz, Paul Byrd, Junichi Tazawa, Justin Masterson, etc, and that didn’t work out so well for them. Fortunately, Buchholz pitched reasonably well, and they had Beckett and Lester and a decent ‘pen.

    So those are the outlier teams, payroll-wise, and they both won a lot of games doing it both ways. Good offense, 3 or 4 starters and then whatever the Hell was left in their bullpen became their starter for those other nights.

    So let’s look at another team, the Angels.

    4 guys with 20+ starts. Two guys with 10+ and three others with 4+. So it starts to look a lot the same, right? 4 starters and then whatever’s left. Worked for them, they won 95 games and beat the Sox in the ALDS, then took the Yanks to 6 games. They had two 16-game winners, but Saunders wasn’t very good by any real measure, their offense bailed him out a lot (6.5 R/G is nice…). The point is, again you see 4 semi-reliable starters and then pitch-by-committee after that. The 5th starter? Not really important to the teams winning the divisions and wildcard so far.

    So let’s look at Minnesota. 87 wins, payroll not astronomical. Couple of good offensive players. 7 guys with 12+ starts and then guys with 9 and 5 (and 1, but that’s not really important). They spread it around, big-time committee pitching. Only 3 guys with 20+ starts.

    But who else? Where’s that 5th starter on a good team? Detroit? 4 guys with 25+ starts, no one else with more than 8… but 7 guys with 3-8 starts. The Rangers? 8 (!!) guys with 10+ starts, though only 3 with 20+ (Millwood, Feldman and Holland), though 3 guys had between 17 and 19.

    The Mariners? 9 guys with 10+ starts, 2 with 20+. Hell, 2 with more than 15, King Felix and Jarrod Washburn. And Tampa? 6 guys with 18+ starts and one guy with 6.

    The point is, among the good teams in the AL (including the division winners, the wild card and the guys contending legitimately for the wild card… and the Jays, who represent a bad team), there really isn’t a team that has just 5 guys hogging the starts.

    There is, however, a great proliferation of teams that choose to spread out the starts that would, in principle, be the 30-some starts a 5th starter would get among 3-5 other guys. Which is exactly what this article is talking about.

    Vote -1 Vote +1

    • Luke in MN says:

      All those winning teams also considered themselves to have 5-man rotations, more or less, even though the personnel shifted as the people they thought were their best 5 got injured or turned out to suck. So who needs the revolution?

      And if the point is just “don’t overspend on a 5th starter,” I’ll do you one better: Don’t overspend on anyone.

      Vote -1 Vote +1

      • ineedanap says:


        Especially in the case of the Angels, their “5th starter” was one of the pitchers that made 20+ starts. Its a good thing they had a long man, a prospect and a minor league vet in addition to their 5th starter or they probably wouldn’t have made the playoffs.

        Vote -1 Vote +1

    • Carlos says:

      The biggest flaw in your thinking here is that these teams are “choosing” to start all those pitchers. The only thing they chose to do is have enough arms available to cope with injuries and ineffectiveness. If their theoretical “top 5″ are all healthy enough to make 30 starts (pretty much never happens) they’ll be glad to let the remaining 3-4 arms sit rotting in reserve the whole year. The reason this article is getting so much backlash is because its silly to suggest a team should plan on having 3 guys share one spot of the rotation. Pick the best of the 3 and have the other 2 waiting back for the inevitable injury. This really doesn’t have to be so complicated.

      Vote -1 Vote +1

  16. Tyler says:

    Further, I think it’s just about managing the pitchers that you use for spreading out those starts and trying to keep cost down. If your cost-to-win ratio is too high at the 5th slot, then you’re burning money. That’s like, “if you set fire to a forest, frogs will die” simple, of course, but what I’m saying is, you’re not going to get a lot of high-end production from whoever you throw into your 5th slot… if you did, he wouldn’t be a 5th-slot starter.

    So why spend a lot of money on a guy who’s at best going to give you a league-average performance most of his good nights, when you could be using prospects and relief arms and whatever else you’ve got to get that same result?

    Vote -1 Vote +1

    • Carlos says:

      Because maybe those prospects and relievers will be well below league average? Shouldn’t the goal be having your best players on the field, period? This whole thing is so circumstantial is ridiculous to suggest a hard and fast rule like the author is proposing.

      Vote -1 Vote +1

  17. matt w says:

    I’m going to join with everyone else in expressing bewilderment. If a team reliably could “secur[e] four starters that can make 24 starts or more,” that would be great and a smart thing to do, but the problem is that teams can’t predict which of their starters won’t make 24 starts. (And if they could, it would probably be worth it to find such a fifth starter.)

    What I think the analysis shows is that it’s not worth much to focus on a fifth starter at the expense of your sixth and seventh starter. A lot of teams probably do carry a swingman and a minor-league vet to step into the rotation when they’re needed. It’d be interesting to see how many starts the #1, #2, … #5, #6, #7 starters averaged (counting by numbers of starts, not by quality) — that’d show how important your backup options are.

    Vote -1 Vote +1

  18. BD says:

    Teams already routinely use long-relief guys, AAA prospects, and AAAA has-beens to make starts when their regular starters go down with injuries, etc. What the OP is proposing, however, is that, rather than wait for an injury to happen that will force the team to rely on these backup/emergency starters, they should just start using them from the first week of the season, dispensing with one of the “regular” starters in the process.

    One argument that has somehow crept into the discussion is the idea that the aforementioned “long-relief guys, AAA prospects, and AAAA has-beens,” as a group, would either be BETTER (qualitatively) or at least a BETTER VALUE than a full-time “no. 5″ starter. But has anyone presented any evidence to prove this? As lackluster as the average “no. 5″ starter is, he’s still likely to be a better pitcher, start-for-start, compared to the same team’s mop-up reliever and AAA/AAAA call-ups. I also don’t see any particular reason to think the “committee of three” would provide better per-win value as compared to a full-time “no. 5.” One can assert such better value in the abstract, but it really requires some evidence to be believed. If a guy’s main purpose in holding down a roster spot is to provide 5-10 starts per year, he’s going to have to have a pretty low salary in order to be a better value than a “no. 5″ who can at least give you MLB 180 innings.

    (The money aspect is, I think, a little unrealistic anyway, because “no. 5″ starters are often not signed to BE no. 5 starters. They are often someone at the beginning or end of their careers. If the former (think Phil Hughes), they are cost-controlled guys who are hopefully in the process of developing into front-line starters. If the latter, they may well be getting paid like a no. 1-3 starter because that’s what they were at the time they inked their contract. )

    Vote -1 Vote +1

  19. B N says:

    I have a better cost-effective plan. Every season, only claim to have 3 starters on board. Sign everybody else as if they are swingmen or relievers, at the appropriate pay. Then make those guys start, even though you’re only paying them about 3 or 4 million, tops. Cost effectiveness at work! Except like… that’s not going to happen.

    But seriously. Let’s think this through. You got 162 starts to play with. You’d love to sign 5 guys that play better than AAA or AAAA level to play those spots, generally. You’d also love to assume that they’re healthy enough, in theory, to make 30 starts in a season. Since that’s not a reasonable expectation (probablistically), you say instead “I have 5 guys who should average 25 starts apiece.” So you take into account the impact of your aggregated injuries to different starters, etc. And then you fill in the remaining 37 starts with your backups (swingmen, farm guys, late season signings, etc).

    Now instead imagine that you come in with 4 guys who should average 25 starts. Now you need to have enough depth lying around to substitute in 62 starts from your swingmen, AAA, etc. 62 starts! Me gods. That is a LOT of starts to be pushing down into your system. Most systems don’t have depth/money to draw on enough guys to fill in those slots. Sure, the Yankees can always grab a new guy via a salary dump. And the Cardinals know black magic. But for most teams, the reason why you signed 5 guys who could put up 25 starts is that you can only reliably substitute for about 40 more. So instead of expecting to try to get 60+ starts out of your relievers and farm, you sign another guy expected to put up 25.

    And this guy you have just signed doesn’t even have to be better than the guys who put up those extra 40 starts (who, by the way, would be guys like Tallet generally). He just has to be better than the guys who would put up the next 20. You can certainly argue that many clubs are paying too much for their last “could make 25″ but to say that teams shouldn’t get one and should expect their depth to absorb 60 starts is crazy.

    Vote -1 Vote +1

  20. Michael says:

    This sounds a little too similar to the 2003 Red Sox’ closer by committee; and that certainly was not a success. I think it is a better approach to have a set 5 man rotation at the beginning of the season, but also carry a middle reliever who can fill in as a spot starter as injuries occur, a role I imagine Tim Wakefield will assume this year in the Sox’ case

    Vote -1 Vote +1

    • Jason B says:

      The Red Sox “closer by committee” approach didn’t mark the concept as a failure, any moreso than Tony Gwynn striking out once indicated he was a complete failure as a hitter. It’s still a worthwhile concept that some forward-thinking team will embrace and tune out the cacaphony of “WAIT…THIS IS NOT WHAT I’M USED TO! CHANGE IS BAD!!”

      Vote -1 Vote +1

  21. kingjd36 says:

    Isn’t it better to commit to the 5th starter to preserve the arms of those long relievers/AAA guys? If you’re going to get sub-league average performance, pay for it so that it doesn’t hurt your future. Don’t toss your rookie in a swing role on the MLB roster to be stretched out, then back to the bullpen, then stretched out. That’s what the Yanks did with Joba and it gave him no shot to get comfortable or getting his arm right or anything. He should have started the whole year and maybe skipped a turn or 3 when his spots came up in July/August.

    Aside: Why do teams limit innings only in August/September? Wouldn’t it be better to prepare the kid to pitch in the post season by having him make a full 5 starts in September and miss a few in June/July or something?

    You have better mental preparation and feel more confident when you are given that prominent starting role every 5 days. You’ll know where your arm strength is at, what you can do to improve, etc etc. You don’t learn how to pitch fatigued when you’re in the pen, or make any adjustments since you only run through the lineup once usually. Nor do you actually really learn how to pitch I imagine. Bullpens seem more and more about throwing than actual pitching these days.

    Vote -1 Vote +1

    • Kevin S. says:

      I agree. I kind of felt that somebody like Hughes (or Porcello, Anderson, etc. last year) should just be started up a month late, then let fly without restrictions. If he’s your fifth starter, he won’t be needed until May anyway.

      Vote -1 Vote +1

  22. DrBGiantsfan says:

    I get what you are saying, and I do think that teams need to think more outside the box when it comes to #5 starters, such as converting relievers to starters, but I think you also need to at least try to have 5 quality starters on the team. Ideally, you also have one decent option in the bullpen and 1 or 2 more stashed at AAA, but as others have pointed out, you need them in case ANY starter goes down, not just #5.

    Look at the difference between the 2008 and 2009 Giants. In 2008, they basically took your approach and between Correia, Hennessey and Matt Palmer all of whom had ERA’s over 6. The Giants’ #5 starters won a total of 3 games between them that year.

    They signed RJ in 2009, and though he wasn’t great, and got injured halfway through the season, he had an ERA<5 and won 8 games, 5 more than the #5 starters won in all of 2008. Assign RJ any WAR you want, but he gave the Giants 5 more wins than what he was replacing! $8 M is pretty good price for 5 wins, no?

    Vote -1 Vote +1

  23. Padman Jones says:

    The big problem I have with this idea – and I do like the idea – is that once the season gets underway, is there really any difference in ‘pressure’ between any of the rotation spots? Certain guys have more pressure because of higher expectations – see e.g. Zack Greinke, Roy Halladay, Felix Hernandez, Tim Lincecum, et al – but it’s not like every game is a #1 starter vs. another #1 starter, or a 2 v 2, and so on. In the middle of July, a guy making his first start isn’t going to care whether he’s the nominal #5 guy or not – he’s going to have to deal with the pressure of pitching in an MLB game.

    Vote -1 Vote +1

  24. smurfy says:

    Good thinking, Marc. Flexibility in constructing the best starting staff your resources can support, and providing opportunities for development.

    All this flack these guys are giving you, forget it. Let them crawl back in the box they crawled out of.

    Vote -1 Vote +1

  25. Tyler says:

    Rob Neyer made an interesting comment regarding the OP by Marc. Basically, he said major league managers are too stubborn to change their basic approach to the construction of their rotation/’pen but that it’s probably true that a different system could be more efficient. He seemed to think that the idea of applying this idea to ALL clubs was poor,but that it could certainly work with specific teams who had the right kind of talent on their roster.

    Interesting stuff, might want to have a look over at Sweet Spot, his blog:


    Vote -1 Vote +1

  26. okin15 says:

    Very simply, a team that expects to win it all needs at least 5 legitimate starters, plus these three guys, because when all is said and done, there’s always a starter or two injured, and it’s not always the “5th” guy”, so you need legitimate people to slide up.

    Vote -1 Vote +1

  27. Tyler says:

    okin, who’s the “legitimate” 5th starter on the Yankees or Red Sox from last year?

    Vote -1 Vote +1

  28. Chili says:

    Saying “don’t overspend on anyone” is correct, but not useful.

    As Neyer noted, it’s always good to consider new ideas. The best argument is this: If you aren’t contending this year, don’t sign a 5th starter because doing so means paying too much for a lousy performance, and possibly stunting the growth of younger players.

    However, if one is contending, it behooves that team to find and pay for someone who is not of league-average 5th starter caliber, but will, in fact, give his team a chance to win when he starts.

    Vote -1 Vote +1

  29. Deacon Drake says:

    24 starts is too high a bar. Given a healthy rotation all season, and a true #1 starter, you’d expect 35, 34, 34, 33 starts out of 1-4, leaving a maximum (barring rainouts, double headers, etc) of just 24 starts for a fifth pitcher. The bar should be 20 starts.

    Every few years a manager toys with the idea of a four man rotation, but no team really carries the horses to pull it off, and those who do actually have a competent enough fifth guy or a rookie who needs the work and do not assume the risk.

    Vote -1 Vote +1

  30. smurfy says:

    Tyler, thanks, that was a good link.

    I think Rob is right, Marc stated this as a plan, rather than as an example of a plan to cope with current resources, and to plan them. But I think Marc is right to think about it.

    As a Yankee fan, I have been frustrated by the combative stances thrown over axioms, with never a changeup in the “mix.” Hey, here’s someplace that hackneyed keyterm really applies: it’s “who is in the mix for the 5th starter?”

    Vote -1 Vote +1

  31. Tyler says:

    No problem. Can’t much help with the sabermetrics themselves, but this kind of discussion is a little more up my alley, lol.

    I think The Book is a good example of why people in the business need to be more flexible about changing their perception of how to play the game. And there are always times when as fans we see what a manager is doing and berate him for what must be a ridiculous decision, even though we don’t always see what’s going on (but damn, Cito, why do you always have to go with the guy we know is going to screw up the game? He’s no good!!).

    Teams moved to a theoretical 5-man rotation in principle because it helps keep pitchers healthier. It’s of interest to me that the team that first did it for real in the 70s was the Dodgers, who actually had the pitching to use 5 guys in a real rotation, whereas there really aren’t that many pitching staffs with that sort of depth.

    I think, to at least a small degree, managers are lazy. They don’t necessarily want to get into the logistics of figuring out the optimum recovery times specific to each of their pitchers, they don’t necessarily want to be constantly juggling things around and working with schedule blocks and everything to try and really optimize a rotation and they don’t want to leave themselves vulnerable to criticism. And it’s all dependent on the talent you have on your roster, no question, so that should affect things.

    But the Yankess and Joba Rules, the way the Phillies worked Moyer and Martinez… There have to be creative ways to mix up your rotation, especially the back end of your rotation, that are more effective. I know some teams actually use a 4-man rotation earlier in the year when the schedule is sparser; that’s an interesting start.

    We’re already effectively platooning pitchers at the back end of the rotation, so what if we were to take those same pitchers and maybe platoon them in-game? Brett Cecil is an interesting talent but he’s been a closer in the minors mostly, and he blows after the 3rd inning just about every game. What if we start him and use one of the pitchers we rarely use out of the pen but frequently use as a starter in place of Cecil and let the two of them platoon through a single start? The IP of each pitcher doesn’t change much, but you get the best of each pitcher instead of letting him stay out there when his stamina gets low and he gets hammered.

    Stuff like that, I don’t know, I just wish managers would be a little more experimental when the stuff they’ve always tried doesn’t work out so well. *cough*CITOGASTONANDBRIANTALLET*COUGH*

    Vote -1 Vote +1

  32. Jules says:

    This all sounds like a mighty long winded way around: Consider 8 Starters, Pitch the Best 5.

    Vote -1 Vote +1

  33. Tyler says:

    If you have 8 starters who can all go at least 6 innings without collapsing and losing command of the strikezone, then you have almost a luxury rotation. But absolutely, if you have that kind of depth, pitch the best 5.

    I think the point is that teams don’t really seem to have 5 guys who are really reliable as starters for the most part; they can settle on four, but it’s just sort of whoever can eat the innings after that.

    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>