Pleading the Fifth—NL Style

The other day over at Fangraphs, R.J. Anderson responded to a quote from Dodgers’ GM Ned Colletti about Colletti’s desire for a legitimate fifth starter. “We’d love to have a bona fide 5th starter,” Colletti stated. Anderson responded by making the case that the Dodgers already have their “bona fide 5th starter” on their roster.

Anderson referenced a quote from Matthew Carruth defining what it meant to be a fifth starter. Essentially, Carruth argues that when fans or GMs argue for “a mythical fifth starter”, they’re usually arguing for someone who produces a league average ERA, pitches 180+ innings and wins 10-12 games. That pitcher is a myth because Carruth says that the true definition of a fifth starter is one who is “the worst starter to have actually pitched.” To illustrate this point, let’s examine the 5th best starters in each NL team’s rotation from 2009. Here, I’m defining the 5th starter as the starter with the 5th best FIP among the group of starters on each team who started the most games for their respective teams.

**Is this guy bona fide, Ned? He was the fifth starter on a division champion last year!** (Icon/SMI)

Among this group of fifth starters, not one pitched as many as 180 innings in 2009. Again, not one team had its fifth best starter (as defined by FIP) meet the innings “requirement” that most people basically see from a “bona fide 5th starter.” In fact, not one even qualified for the ERA title. Aaron Cook led this group by throwing 158 innings, followed closely by Jamie Moyer’s 143.1 innings, Kenshin Kawakami’s 142.2 innings and Rich Harden’s 141 innings. This group of 16 pitchers averaged fewer than 20 starts (19.88) and only 109.1 innings pitched.

Their average record — 5.75 wins and 8.06 losses. Their average ERA and FIP — 5.13 and 5.03, respectively. In other words, fifth starters were — as one would expect — far worse than league average starters in 2009. In fact, they were basically replacement level starters. Cook’s 1.9 WAR led the group, followed closely by Harden’s 1.8 and Kawakami’s 1.7 WAR. As fifth starters go, these guys did a nice job but they’re hardly the innings-eaters Colletti seems to be looking for.

Other members of this illustrious group included Shairon Martis, Josh Geer, Eric Stults, Micah Owings (who’s better known for his hitting, by the way!), Yusmeiro Petit, and Chris Volstad. The group also includes David Bush and his 6.27 ERA, Todd Wellemeyer and his 6.08 ERA (and his team won its division!), and Felipe Paulino’s 5.69 ERA. It also included future Hall of Famer Randy Johnson.

What, then, are the implications for teams’ rosters? The bottom line is that teams do not need to spend valuable resources in an attempt to bring in an innings-eater for the fifth spot in the rotation. While it would be nice to have everyone in the rotation make 28-30 starts and throw 180-plus innings, that’s just not realistic. Why pay a guy $5 or $6 million (or more!) if you’re only going to get 110 innings and a 5.50 era from him? “Fifth starters”, therefore, are very fungible creatures. A team doesn’t need a great one in order to reach the postseason or even to reach the World Series. In fact, only Aaron Cook even made his team’s playoff roster.

Anderson argues that Colletti’s “fifth starter” is probably already on his roster in the form of James McDonald. Even if McDonald’s not the answer, he’s not going to be that much worse than any other team’s fifth starter and it won’t take long to figure out that the Dodgers need to try someone else. Moreover, the likelihood that the pitcher who starts the team’s fifth game — or the first time a fifth starter is needed — also finishes the season in the rotation may not be that great. He probably won’t make 25 starts or pitch the 162 innings needed to qualify for the ERA title anyway. And if he does, it’s likely that one of the “top four” starters in the rotation failed or injured himself sufficiently to push the “fifth starter” up in the rotation.

So, Ned, don’t fret about not being able to pay someone five or six million dollars as your fifth starter. Tell Joe to try McDonald or Stults or one of the other guys you were lamenting in your conversation with Buster Olney. You’ll be just fine (as long as Chad Billingsley pitches the way he did in the first half last year!).

Newest Most Voted
Inline Feedbacks
View all comments
13 years ago

Kawakami threw 156.1 innings.

Chuck Brownson
13 years ago

So he did.  Thanks, Rob.  Don’t know where I got 142.2 innings from.  That brings the average ERA down to 5.09.

Chuck Brownson
13 years ago

I know where I got that number from.  142.2 innings is the # of innings Kawakami threw in his starts.  he must’ve thrown 13.2 in relief.  I only looked at # of innings from starters and didn’t count relief innings.

13 years ago

It would be interesting to know how often (not just last year) a team’s fifth best starter threw 180 innings with a league average ERA (or FIP).  You can even leave the win criteria out.  I bet it doesn’t happen very often and I bet when it does happen, that’s a darn good team.

Heck, most teams can’t get 180 league average innings out of 3 guys, much less 5.  Basic math will tell you that if every team starts only 5 guys all year long, there will be 75 pitchers better than average and 75 worse or an average of only 2.5/team better than average.  Even worse, only 57 guys in MLB threw 180 innings last year or 1.9/team.  So to have 5 guys on the same team with 180 innings is hard enough.  For all to be league average by any performance metric is near impossible in the modern era.

King Kaufman
13 years ago

It seems to me that Colletti is saying he’s like to have a legitimate starting pitcher in the No. 5 slot, not that he’s looking for some mythical creature called a “legitimate No. 5 starter.”

It sounds like you’re saying that since most teams—all of them in your sample of one league, one year, though I think the “most people define this as 180 IP etc.” is a straw man”—don’t have a legitimate No. 5 starter, then there’s no point in having it as a goal.

“Why pay a guy $5 or $6 million (or more!) if you’re only going to get 110 innings and a 5.50 era from him? “Fifth starters”, therefore, are very fungible creatures.”

Well yeah, but what if you could find someone who can give you 150 IP at 4.55? That’s a whole lot better than 109 IP at 5.13. So that’s a hell of a marginal advantage over all the other teams, who are getting that production on average on Day 5. That’s worth $5 or $6 million. If you’ve got $5 or $6M lying around, why not try to get that guy? Just because it’s “not realistic” to talk about a “legitimate fifth starter”?

There’s nobody walking around as “the best fifth starter in the league.” But if you can get a second “third” or “fourth” starter,  as defined in Carruth’s piece, you should get him.

13 years ago

The article is an interesting exercise, but it misses the point. To state King Kaufman’s point in another way, no team enters the season knowing which starter will be its worst. So you don’t fill that slot with a guy who you assume will be replacement level; you do the best you can, and you hope that none of your supposedly better starters has an off-year.

13 years ago

The problem is that the fifth starter on opening day is rarely the same as the one in your study. If a team forfeit looking for a fifth starter, they will more than likely end up missing the fourth.

Rui Xu
13 years ago

Chuck, great post.  As a fellow Cardinals fan, I love the stuff you do over at VEB as well, when you get the chance.

I wonder if we can relate Colletti’s statement to Dave Cameron’s “Marginal Value of a Win” column he did over at Fangraphs.  The Dodgers, while a favorite in the NL West, are probably sitting in that 85-90 win range where they would be willing to pay more for a marginal win than, say, the Pirates should or that the Yankees should.  I understand that paying a couple million dollars for Jon Garland, in a vacuum, seems like a poor decision when projections say that a random AAA pitcher will likely match that performance, but to me, it seems that the extra few million dollars are simply a risk premium that GMs are willing to pay.  With Garland, you know exactly what you’re getting -you’re going to get 200 IP with a 4.50 FIP.  With an unknown, you’re obviously going to see a lot more variability, and with a team that close to the playoffs, it can’t afford variance of the negative variety.

Of course, that sort of thinking only applies with teams in that playoff “bubble,” and who feel that the fifth starter is the best way to get that marginal win.  Like I mentioned before, I would not expect the Pirates to pay that much for a fifth starter, but I also would not expect a team that could find a marginal win otherwise (and more cheaply) to pay that much for a fifth starter.  I know that your post focused more on the definition of the fifth starter than what teams are likely to need a better fifth starter, but I think that my points are still relevant to the discussion.  Your thoughts?