Starters make the best bullpen pitchers

[Illini Iceman]

Good to see you starting threads and in general posting with a little more regularity again Jon.

 

I agree, wth is Lou doing?

I've gotten used to the a lot of things this organization does to piss me off, but playing worse veterans over already better younger players is one thing that I just can't get over. This includes starting vs. pitching out of the bullpen.

 

The biggest problem with Marmol throwing 40 pitches in relief is that the Cubs aren't isn't maximizing his production. If he's throwing that many pitches in an outing, certainly not all of them are coming in high leverage situations relative to the typical late inning relief appearance. Of course, if he theoretically throws only in high leverage situations, he's not racking up the innings to maximize his production, either. The simplest solution is to make him a starter.

 

And as for Marshall, I'm mostly speechless:

 

4/18 - 0.1 IP, 2 pitches

4/17 - 0.1 IP, 2 pitches

4/13 - 0.2 IP, 11 pitches

4/11 - 1.0 IP, 13 pitches

4/9 - 1.0 IP, 16 pitches

 

That's 3.1 IP with 44 pitches over 1.5 weeks. Based on his averages from starts last season, if he made two starts over the same 1.5 weeks, we could have expected 10.2 IP with 168.4 pitches.

 

And he's better than Marquis and Dempster.

 

What on earth are they doing to him?

 

Yes, let's make Marmol a starter so every time he has over 100 pitches in the 6th inning you can start crying about Piniella abusing him.

 

The role he has now is good. He works late innings in tight games. Yesterday was fine, he locked down the win in a close game where the Cubs weren't scoring many runs. So maybe he doesn't pitch again this weekend, but you go with the win you have in hand.

 

Having a guy that can come into the game in the 7th or 8th inning and shut down the opponent to get the game to Wood is huge. Hopefully Howry can regain 2007 form so Piniella can go Howry, Marmol, Wood, game over but for now he is going with the guys he can trust.

 

As for Marshall, if Piggy could have done the job he would be getting those opportunities. Marshall has been moved into the role and is doing fine. Pitching in relief for a few weeks until Eyre comes back won't hurt him. He will then go back to AAA to be stretched back out and be ready if Marquis or Dempster can't hold their spot in the rotation.

[WilcoFan]

I consider myself pretty statistically inclined, but I just can't see the value in having any young, successful reliever moved into the starting role. Aggregate innings statistics don't really do it for me. Sure, theoretically he'd be pitching more innings, but the necessity of those innings has to be weighed against the necessity of the innings he pitches in relief. Sometimes, aggregate statistics cannot tell you the significance of a player to an individual game, as baseball games are chunked into 9 inning statistical segments.

 

 

I think a number of factors need to be considered:

 

1. What would be lost in terms of GAME-TO-GAME production if Marmol was removed from the bullpen? For example, if he were to start (and be as productive a starter as he is a reliever) and win every fifth day, but on 2 out of the 4 other days each day the bullpen (consisting of, theoretically, less quality relievers) blows games, isn't that a net loss in terms of games won for the team? Are all innings of equal relevance to the overall W-L record? (NOTE: This is all theoretical, of course, but it is what needs to be considered)

 

2. Is the dropoff from Howry (7th)-Marmol (8th) to, for example Wuertz (7th)-Howry (8th) bigger than the dropoff from Marmol-Marquis as a starter? Is Marquis starting over Marmol costing the team more games than if Marmol had to be replaced in the pen?

 

 

Once again, I think the aggregate needs to be avoided here. Sure, Marmol over more innings would probably have better aggregate stats than Marquis. But the significance of those innings to the nine-inning statistical segments known as games is what concerns me.

[Guest]

1. What would be lost in terms of GAME-TO-GAME production if Marmol was removed from the bullpen? For example, if he were to start (and be as productive a starter as he is a reliever) and win every fifth day, but on 2 out of the 4 other days each day the bullpen (consisting of, theoretically, less quality relievers) blows games, isn't that a net loss in terms of games won for the team? Are all innings of equal relevance to the overall W-L record? (NOTE: This is all theoretical, of course, but it is what needs to be considered)

So you're saying that Johan Santana and Jake Peavy should pitch relief because they'd be more valuable there.

[Hairyducked Idiot]

I really don't see what's wrong with Marmol staying in relief. A truly dominant relief pitcher used in high-leverage situations (which Piniella is mostly doing, although I'd like to see Marmol a bit less in 2-run leads) can be just as valuable as a high-innings starter.

[Guest]

I really don't see what's wrong with Marmol staying in relief. A truly dominant relief pitcher used in high-leverage situations (which Piniella is mostly doing, although I'd like to see Marmol a bit less in 2-run leads) can be just as valuable as a high-innings starter.

So you're saying that Johan Santana and Jake Peavy should pitch relief because they'd be more valuable there. :D

[Hairyducked Idiot]

http://baseballprospectus.com/statistics/sortable/index.php?cid=204021

 

Marmol has been the most valuable relief pitcher in all of baseball to this point, worth a full win over replacement already. His innings have been 58% more important than a basic, start-of-the-game inning.

 

http://baseballprospectus.com/statistics/sortable/index.php?cid=204017

 

Compare the Support-Nuetral Value Added, adjusted for level (SNVAL) for the best starters in the league versus Marmol's Wins Expected Added adjusted for level (WXL), which are basically the same stat. Marmol's 0.865 wins added above average is right up there with the best starters in the league.

[Guest]

http://baseballprospectus.com/statistics/sortable/index.php?cid=204021

 

Marmol has been the most valuable relief pitcher in all of baseball to this point, worth a full win over replacement already. His innings have been 58% more important than a basic, start-of-the-game inning.

 

http://baseballprospectus.com/statistics/sortable/index.php?cid=204017

 

Compare the Support-Nuetral Value Added, adjusted for level (SNVAL) for the best starters in the league versus Marmol's Wins Expected Added adjusted for level (WXL), which are basically the same stat. Marmol's 0.865 wins added above average is right up there with the best starters in the league.

Granted, I didn't phrase it as a question, but if you believe that pitching in relief is just as valuable as starting, would you move Santana & Peavy into relief roles so you could "leverage" their innings?

[Hairyducked Idiot]

Peavy's best ERA+ in a season was 171, Santana's was 182. Marmol put up a 326 last season.

 

IF you knew for a fact that Peavy and Santana could become even more dominant as relievers, and you had a manager with the discipline to use him only in truly high-leverage situations, then their value could be comparable in either situation, yes.

[Hairyducked Idiot]

In terms of wins above replacement, using the stats I listed above, the best starting pitchers each season are usually worth 7-9 wins and the best relievers 6-8.

 

Given that there's no guarantee of dominance in one translating to dominance in another, I don't think there's much incentive to be moving the best relievers to starting or vice-versa.

[Jon]

Through 17 games, Marmol has had 9 appearances with a total of 11 IP and 188 pitches. He's on pace for 86 appearances, 105 IP, and 1,792 pitches. In 2007, he threw 69.1 IP and 1,158 total pitches.

 

Do you really think he's going to keep going at this rate? Looking at wins this early in the season doesn't mean a whole lot.

 

And as I stated before, Marmol is not only used for high leverage situations. He pitched the 7th and 8th innings after entering with a 3 run lead and kept going with a 4 run lead on 4/15. In his other 40+ pitch appearance, he came into the game to pitch the 7th and 8th after entering with a 4 run lead.

[WilcoFan]

1. What would be lost in terms of GAME-TO-GAME production if Marmol was removed from the bullpen? For example, if he were to start (and be as productive a starter as he is a reliever) and win every fifth day, but on 2 out of the 4 other days each day the bullpen (consisting of, theoretically, less quality relievers) blows games, isn't that a net loss in terms of games won for the team? Are all innings of equal relevance to the overall W-L record? (NOTE: This is all theoretical, of course, but it is what needs to be considered)

So you're saying that Johan Santana and Jake Peavy should pitch relief because they'd be more valuable there.

 

Not necessarily, but I'm just not sure what kind of upgrade Marmol provides to this team relative to overall team wins for the season if he were to be a starter versus a reliever.

 

However, if you could theoretically bring in Johan Santana or Jake Peavy to any game that is close throughout a season and had competent starting pitching, I could see how an argument could be made that they would be more valuable to overall team wins using that strategy. I'm also not sure if that strategy would even be physically possible.

 

I guess the rule is that you want your best pitchers pitching in the situations that are most likely to help your team win the most games.

 

You want a really ridiculous, out-of-the-box idea? Eliminate "starters" and "relievers" and just have every pitcher on the team throw until they run into trouble (or, for obvious reasons, run into pitch count issues). Inevitably, your best pitchers would throw more innings, and the remainder of your pitchers could (theoretically) throw the innings that involve the least risk. You can throw the Kevin Harts and Jason Marquises of the staff against the least successful offenses, and be able to bring in Carlos Zambranos when games are on the line.

 

I'm sure there are arm issues raised by this strategy, but hey, its out there.

[Hairyducked Idiot]

Through 17 games, Marmol has had 9 appearances with a total of 11 IP and 188 pitches. He's on pace for 86 appearances, 105 IP, and 1,792 pitches. In 2007, he threw 69.1 IP and 1,158 total pitches.

 

Do you really think he's going to keep going at this rate? Looking at wins this early in the season doesn't mean a whole lot.

 

And as I stated before, Marmol is not only used for high leverage situations. He pitched the 7th and 8th innings after entering with a 3 run lead and kept going with a 4 run lead on 4/15. In his other 40+ pitch appearance, he came into the game to pitch the 7th and 8th after entering with a 4 run lead.

 

Therein lies the problem. Presumably, as the season goes on, his innings will go down and his leverage index will go up.

[RynoRules]

I usually get it when people pick the team apart here, but this one befuddles me.

 

Tim, are we really comparing Marmol to today's Peavy and Santana? Haven't statistics acknowledged that the most important relievers are those that come in when the game is on the line in the 6th, 7th and sometimes the 8th? Shouldn't we therefore utilize one of our best pitchers in that situation?

 

I get that we all think Lily sucks this season, but aside from him the rotation has been fine. We are going to knee-jerk and send our best reliever into the rotation b/c one guy has stunk through 4 or so starts?

[Jon]

Therein lies the problem. Presumably, as the season goes on, his innings will go down and his leverage index will go up.

And along with it the wins. Even if his leverage index increases, I can't imagine that he'll pitch enough innings to make his value as a relief pitcher comparable to what it could be as a starter.

[KingCubsFan]

I really don't see what's wrong with Marmol staying in relief. A truly dominant relief pitcher used in high-leverage situations (which Piniella is mostly doing, although I'd like to see Marmol a bit less in 2-run leads) can be just as valuable as a high-innings starter.

So you're saying that Johan Santana and Jake Peavy should pitch relief because they'd be more valuable there. :D

That's not a good comparison. Santana and Peavy are two of the best pitchers in baseball. As a starter, Marmol has never done anything to indicate he could be at their level. In fact, he's never really shown he could be more than a middle of the rotation guy. However, he's shown to be one of the best relievers in the game.

[Hairyducked Idiot]

Therein lies the problem. Presumably, as the season goes on, his innings will go down and his leverage index will go up.

And along with it the wins. Even if his leverage index increases, I can't imagine that he'll pitch enough innings to make his value as a relief pitcher comparable to what it could be as a starter.

 

You are still operating under the assumption that starters are inherently more valuable than high-leverage relievers, which I don't think the numbers bear out.

 

Last season, he was worth 3.7 wins above replacement, primarily because he was used mostly in low-leverage situations (leverage index of 1.08 is pretty bad for a reliever). Zambrano was worth 5.6, Hill 4.8, Lilly 4.7, Marquis 3.4. 2007 was sort of a down year for relievers, but there were still six worth more than five wins above replacement, and there is no reason Marmol can't join that list this season.

 

If you start looking back at the leaders throughout the years on the links I posted, the standard for a truly great (tops or nearly tops in the league) season is about seven wins above replacement for both starters and relievers.

 

In 2003 (Eric Gagne, 9.23) was worth more than any starter, in 2004 Brad Lidge's 8.1 was behind only Johan Santana, in 2005 the starters dominated. In 2006, F-Rod's 7.3 wins were behind only three starters. In 2007, J.J. Putz's 7.4 wins were behind only Peavy and Hudson.

 

I think if you look at the data, you'll find that top-notch, high-leverage relief pitchers are very comparable in value to top-notch starters. It's just that they have more variable careers and there's really only room for one per team.

[Illini Iceman]

Through 17 games, Marmol has had 9 appearances with a total of 11 IP and 188 pitches. He's on pace for 86 appearances, 105 IP, and 1,792 pitches. In 2007, he threw 69.1 IP and 1,158 total pitches.

 

Do you really think he's going to keep going at this rate? Looking at wins this early in the season doesn't mean a whole lot.

 

And as I stated before, Marmol is not only used for high leverage situations. He pitched the 7th and 8th innings after entering with a 3 run lead and kept going with a 4 run lead on 4/15. In his other 40+ pitch appearance, he came into the game to pitch the 7th and 8th after entering with a 4 run lead.

 

He did not spend all of 2007 in the majors. He also pitched 41 innings in AAA so he had about 110 IP total in 2007 between the minors and the big leagues. In 2006 he had 138 IP combined between West Tenn, Iowa and Chicago. Therefore, the 105 IP he is on pace for this year per your calculations would actually be a slight reduction from previous years so I wouldn't worry about his arm falling off just yet.

[Fro]

Yeah, you're assuming that a great reliever is automatically a great starter, which isn't the case. A crappy starter can be a great reliever, and even some great starters might be crappy relievers.

[Jon]

Through 17 games, Marmol has had 9 appearances with a total of 11 IP and 188 pitches. He's on pace for 86 appearances, 105 IP, and 1,792 pitches. In 2007, he threw 69.1 IP and 1,158 total pitches.

 

Do you really think he's going to keep going at this rate? Looking at wins this early in the season doesn't mean a whole lot.

 

And as I stated before, Marmol is not only used for high leverage situations. He pitched the 7th and 8th innings after entering with a 3 run lead and kept going with a 4 run lead on 4/15. In his other 40+ pitch appearance, he came into the game to pitch the 7th and 8th after entering with a 4 run lead.

 

He did not spend all of 2007 in the majors. He also pitched 41 innings in AAA so he had about 110 IP total in 2007 between the minors and the big leagues. In 2006 he had 138 IP combined between West Tenn, Iowa and Chicago. Therefore, the 105 IP he is on pace for this year per your calculations would actually be a slight reduction from previous years so I wouldn't worry about his arm falling off just yet.

I did forget that he spent time in the minors last year, but 7 of those 8 minor league appearances were starts. I'm not talking about abuse, but rather what we can expect of him as a reliever. The abuse is a different discussion.

 

 

So his relief appearances stretched out for a full season would have meant 92.2 IP and 1,550 pitches. He would have led the league in IP. Only two relief pitchers over the last four seasons have totaled 100 IP and five have totaled over 90 IP. If he throws 105 IP this year, though, he's not pitching in a high enough percentage of high leverage situations. Maybe that increase even supports my argument. His LI wasn't very high last year. This year, his gmLI is only 39th among all relievers in baseball and his inLI is way down the list at 62nd. Why should we expect Lou to use him more exclusively in high leverage situations when there has been a clear pattern that Piniella just wants to use or overuse him in the late innings, regardless of leverage. Doesn't that make his IP less valuable?

 

In Marmol's case, not in a theoretical situation in which a truly dominant reliever might produce expected wins in line with such starters in some instances, do the numbers really backup the idea that he should remain a reliever?

[RynoRules]

Through 17 games, Marmol has had 9 appearances with a total of 11 IP and 188 pitches. He's on pace for 86 appearances, 105 IP, and 1,792 pitches. In 2007, he threw 69.1 IP and 1,158 total pitches.

 

Do you really think he's going to keep going at this rate? Looking at wins this early in the season doesn't mean a whole lot.

 

And as I stated before, Marmol is not only used for high leverage situations. He pitched the 7th and 8th innings after entering with a 3 run lead and kept going with a 4 run lead on 4/15. In his other 40+ pitch appearance, he came into the game to pitch the 7th and 8th after entering with a 4 run lead.

 

He did not spend all of 2007 in the majors. He also pitched 41 innings in AAA so he had about 110 IP total in 2007 between the minors and the big leagues. In 2006 he had 138 IP combined between West Tenn, Iowa and Chicago. Therefore, the 105 IP he is on pace for this year per your calculations would actually be a slight reduction from previous years so I wouldn't worry about his arm falling off just yet.

I did forget that he spent time in the minors last year, but 7 of those 8 minor league appearances were starts. I'm not talking about abuse, but rather what we can expect of him as a reliever. The abuse is a different discussion.

 

 

So his relief appearances stretched out for a full season would have meant 92.2 IP and 1,550 pitches. He would have led the league in IP. Only two relief pitchers over the last four seasons have totaled 100 IP and five have totaled over 90 IP. If he throws 105 IP this year, though, he's not pitching in a high enough percentage of high leverage situations. Maybe that increase even supports my argument. His LI wasn't very high last year. This year, his gmLI is only 39th among all relievers in baseball and his inLI is way down the list at 62nd. Why should we expect Lou to use him more exclusively in high leverage situations when there has been a clear pattern that Piniella just wants to use or overuse him in the late innings, regardless of leverage. Doesn't that make his IP less valuable?

 

In Marmol's case, not in a theoretical situation in which a truly dominant reliever might produce expected wins in line with such starters in some instances, do the numbers really backup the idea that he should remain a reliever?

 

I am still trying to understand how the numbers back your argument that he should be a starter.

[Hairyducked Idiot]

So his relief appearances stretched out for a full season would have meant 92.2 IP and 1,550 pitches. He would have led the league in IP. Only two relief pitchers over the last four seasons have totaled 100 IP and five have totaled over 90 IP. If he throws 105 IP this year, though, he's not pitching in a high enough percentage of high leverage situations. Maybe that increase even supports my argument. His LI wasn't very high last year. This year, his gmLI is only 39th among all relievers in baseball and his inLI is way down the list at 62nd. Why should we expect Lou to use him more exclusively in high leverage situations when there has been a clear pattern that Piniella just wants to use or overuse him in the late innings, regardless of leverage. Doesn't that make his IP less valuable?

 

Yes. I suspect that Marmol's leverage will get higher as the season goes deeper, but I won't use that as an argument.

In Marmol's case, not in a theoretical situation in which a truly dominant reliever might produce expected wins in line with such starters in some instances, do the numbers really backup the idea that he should remain a reliever?

 

If we are getting into the specifics of Marmol rather than the generics of dominant relief pitchers versus dominant starters, then there's a few questions we need to ask ourselves:

 

1) How likely is Marmol to maintain his performance from last season, or that last season represented a true ability level?

 

Pecota is not kind to Marmol at all, projecting a big regression to the mean.

 

Similarity scores on baseball-reference.com come up with three players that I consider good comps for comparable points in his career:

 

http://www.baseball-reference.com/w/wettejo01.shtml

http://www.baseball-reference.com/s/scanlbo01.shtml

http://www.baseball-reference.com/p/plunker01.shtml

 

We're not likely to be getting another 300 ERA+ season from Marmol, to be sure, but I think it's reasonable to expect him to continue to be a top-notch reliever.

 

A top-notch reliever with a good but not great leverage index pitching in the 80-90 IP range is generally worth 4-6 wins above replacement (http://www.baseballprospectus.com/statistics/sortable/index.php?cid=208950 for a list of last season's relievers, Marmol was at 3.7 with fewer innings and worse leverage than we can expect), so I think that's a good base projection for Marmol as we have him now.

 

6.0 wins above replacement last season was Erik Bedard, 5.0 was Carlos Silva, 4.0 was Doug Davis.

 

2) Would "starter Marmol" be in that range of pitchers?

 

PECOTA is pretty harsh on Marmol, projecting him to about a 3.83 ERA. ZIPS has him at about 3.77. I think both of those are a bit high. Some of those comparable players I think point to an ERA in the high 2s, low 3s.

 

On average, a pitcher gains about .8 runs in ERA when he goes from relieving to starting and loses it when going vice-versa, but that's a very generic number and varies a lot from case to case. But we don't have anything better to go with on Marmol, so I think that's a fair assumption. Depending on whether you are using the statistical projections or the comp projections, that would put Starter Marmol at anywhere from an ERA in the high 3s to the mid 4s. I don't think he'd be an innings-eater, so even our most optimistic projection as a starter puts him as a 190-200 inning, high 3s ERA type of guy. That's a lot closer to Doug Davis than Erik Bedard, using our above baselines.

 

So I think that lays out a good case that moving Marmol from relief ace to starter would be essentially a wash at best.

 

3) Do the Cubs need to make such a move?

 

The Cubs are really deep in both the rotation and the pen. I'm going to assume he replaces Dempster in the rotation (and not something stupid like replacing Hill), which would make the most sense despite Hendry and Piniella being buddies with him. A pessimistic view of Dempster would be something like what Milwaukee got from Dave Bush in 2008 (186 innings, 5.12 ERA). That was worth 2.2 wins above replacement, so I'd put replacing him with Marmol at about two wins worth of improvement to our starting rotation. We could probably get the same improvement from Marshall, and almost as much from Lieber, but that's a real improvement. If Marmol replaces Marquis, that's a little less improvement.

 

So would taking Marmol out of the pen cost more than the 2 wins we'd gain? I think so. We have a deep pen with good relievers, so everyone right-handed and not a long man behind Wood moves up a notch. Howry replaces Marmol, Wuertz replaces Howry, Gallagher? or someone replaces Wuertz. Hart and Lieber probably stay in their current roles. I'd say that's a net loss at each switch, and it'd only take an average of 2/3rds a win above replacement loss per move to cancel out the gain from Marmol in the rotation.

 

All in all, I don't think the risk of the move is worthwhile for a team that is already playoff-worthy on paper.

[Jon]

It's a good argument, but I spent my morning making this avatar, so I'll have to reply later. I will say, though, that a lot of it still depends on a combination of Lou's usage of the bullpen, which is unlikely to change, and Marmol's ability to improve as a starter in regards to efficiency. His pitching career is still quite young.

 

I'm also not sure how realistic it would be for the move to the rotation to be made in-season, especially since someone like Marshall is being completely wasted and would be a better option than two starters right now. Piniella probably isn't going to replace both Dempster and Marquis as the season goes on.

 

I would like to reinforce that my original point, though, was with how Piniella uses Marmol out of the bullpen. While I'd still like to see him in the rotation, the main problem is that Marmol is being used at times to eat relatively meaningless innings. And since he's not even the closer, his leverage ceiling is limited right now.

[Transmogrified Tiger]

Santana and Peavy are two of the best pitchers in baseball. As a starter, Marmol has never done anything to indicate he could be at their level. In fact, he's never really shown he could be more than a middle of the rotation guy. However, he's shown to be one of the best relievers in the game.

 

I don't really like the Peavy/Santana comparison, but Marmol has shown he can be a starter, quite a good one too. At AA he put up a 1.16 WHIP, 10.39 K/9, and 2.68 K/BB. He followed that up with 7 starts in AAA before going to the Chicago pen: 1.02 WHIP, 10.53 K/9, and 4.00 K/BB. Those are outstanding numbers, and considering his stuff and the fact that he wasn't too old for those levels, they could project him as a very good starter, certainly better than middle of the rotation.

 

Personally, I don't mind Marmol's usage thus far. It's a little more than I'd like at this point pitch count wise, but he's been rested appropriately, and I have confidence that as Howry and Wuertz round into form that he'll be used better.

[Hairyducked Idiot]

Marmol's leverage index has improved nicely of late, up around 1.8 now, and he leads the league in expected wins added.

[TheDude]

I thought the primary reason Marmol was not returned to the rotation was his lack of a third pitch? When you throw in mid-90s and have only two pitches, both easily plus, you're a bullpen pitcher.