« Pavano: Working Hard Or Hardly Working? | Main | A Touch of Pinstripes »
February 08, 2008
The Great Debate Of '08: What Job For Joba?
This is an attempt to address the debate of "Where should the Yankees use Joba Chamberlain in 2008 - as a starting pitcher or as a relief pitcher?"
First, in this approach, I'm going to assume that Chamberlain will excel as either a starter or a reliever. Granted, that's a bold assumption. But, for the heck of it, in this study, I want to assume that Joba's production would be a constant (of sorts) - meaning great - in either role. And, then, from there, we can see where he gives the Yankees more value.
Using the Complete Baseball Encyclopedia and the stat "Runs Saved Above Average" (RSAA) I came up with the following list of the best starting pitcher and best relief pitcher, in the majors, for each of the last four seasons:
Relief Pitcher Starting Pitcher
RSAA RSAA
2007 Rafael Betancourt 29 Brandon Webb 44
2006 Jonathan Papelbon 28 Brandon Webb 45
2005 Mariano Rivera 27 Roger Clemens 53
2004 Brad Lidge 26 Johan Santana 54
Next, I went to the Hardball Times site to find the "Win Shares Above Bench" (WSAB) for each of these pitchers in their respective seasons:
Relief Pitcher Starting Pitcher
WSAB WSAB
2007 Rafael Betancourt 10 Brandon Webb 17
2006 Jonathan Papelbon 12 Brandon Webb 16
2005 Mariano Rivera 11 Roger Clemens 19
2004 Brad Lidge 11 Johan Santana 20
Anyone else seeing a pattern here? Based on these two charts, it's not a reach to say that (over the last four seasons):
The best relief pitcher in baseball is good for, around, 11 Win Shares above the Win Shares that an average replacement player would have received, given that player's time on the mound.
...and...
The best starting pitcher in baseball is good for, around, 18 Win Shares above the Win Shares that an average replacement player would have received, given that player's time on the mound.
Since a Win Share represents one-third of a team win, it seems to make sense to divide the "11" and "18" WSAB amounts by three - in an attempt to determine how many "wins" the best reliever and best starter would give a team over a replacement pitcher. And, the answer is: The best starting pitcher gives his team 6 more wins than a replacement starter and the best relief pitcher gives his team 3.67 more wins than a replacement relief pitcher.
Therefore, continuing to leap with this, based on what I've shared herein, so far, I going to offer that, assuming Joba Chamberlain is "lights out" (great) as either a starting pitcher or a relief pitcher, he would be worth 2.33 more wins to his team as a starter, than a reliever, compared to a replacement pitcher. (Six minus three-and-two-thirds is two-and-a-third.)
Of course, if Joba were to be just good, and not great, as a starting pitcher this season, but, still great out of the bullpen, that 2.33 advantage would probably get knocked down to null, or less.
Now, it's time to factor in the matter of the "innings limitation" that the Yankees want to have on Chamberlain in 2008. If I had to guess, I would offer that New York does not want Joba to throw more than 145 to 150 innings this season - so that they do not abuse his arm.
Based on all of this, it seems, to me, to be very prudent to use Joba Chamberlain out of the bullpen - for this season. By doing this, you ensure that you keep Joba under 145 innings pitched. And, the "cost" (for lack of a better word) for making this move would be about two wins - and this assumes that Chamberlain would pitch like an elite stud if he was in the starting rotation. Plus, you also address a serious area of need for the Yankees this season - their bullpen.
Don't get me wrong. If "innings" were not a concern for Chamberlain this season, then I would probably say "Start him as those extra two wins might be the difference between making the post-season or not." However, I would not be willing to risk the future of a prospect as great as Joba for what might be two extra wins, at best, this year.
Hey, if the stats said that it was a difference of ten or fifteen wins, then, maybe, just maybe, I would consider noodling the potential frying of a prospect for such a huge impact on a season's outcome. (I'm not saying for sure I would throw caution to the wind - just that I would really give it long thought before being quick to say "no way.")
But, for what might be, again, at best, a two win gain, is it (meaning the abusing of a prospect) worth having Chamberlain start, and not work out of the pen, in '08? Nope. It sure sounds like being penny-wise and pound-foolish to me.
Posted by Steve Lombardi at February 8, 2008 09:35 PM
Comments
I'm curious- why did you list RSAA and then go with Win Shares to make your point? It seems like the disparity in RSAA is much greater between the best starters and the best relievers.
Secondly, I guess I've confused myself on the nature of "average" and "bench" in these cases. Are these defined separately for relievers and starters, or is there some "pitcher" baseline? If it's the former then I suppose the comparison works.
Here's the primary reason why I am wary to "stick Joba in the pen" for all of 2008: he may never get unstuck. As you've admitted, if innings aren't the issue, then starting is more valuable to a team than relieving. If he never starts in 2008 and pitches like 80 innings (a normal reliever workload), no one's going to think he can flip to being a starter.
I think, then, that PECOTA's projection is reasonable. To get to 145 innings, that's 50 (one inning) relief appearances and 15 starts (of slightly more 6 innings each). The question for the Yankees is whether to start him in the pen and bring him out for spot-starts with the inevitable injury (looking at you, Mike Mussina). Or, do you have him in the rotation and flip him to the pen late in the season.
One scenario where I would consider the latter: we've seen how dominating Joba/Mo can be in the postseason. If Kennedy and Hughes step it up this season, then there are too many starters for a 5 game division series. Put Joba in the pen there, absolutely. He can make a bigger impact relative to the other starters, coming out of the pen.
I just think if Joba is lights out in spring training as a starter, the Yankees will be hard-pressed to send him to the pen. If there's also an injury to one of the other five rotation candidates, you've gotta start Joba in the rotation for sure.
Posted by: mehmattski ![[TypeKey Profile Page]](http://www.waswatching.com/nav-commenters.gif)
at February 9, 2008 12:24 AM
I don't think you are incorporating the fact that if Chamberlain is a reliever this season, he will likely get 80 innings or less...and then be set back a year in his development as a starter, or he won't be a starter at all.
Wouldn't you agree that the best case scenario for Joba, and this team, this year and going forward, is to mimic Johan Santana's 2003 season, or 2007 Chad Billingsley's season, where he is a man out of the bullpen and a spot starter for the first two thirds of the season, and then is converted to a full-time starter for the last third and playoffs? Johan Santana started Game 1 for the Twins against the Yankees - and that was the only win the Twins got in the playoffs that year.
I really, truly think this is what the Yankees are planning on for Joba. It makes a LOT of sense, for this team now, for the player's development, and for the team going forward. You do NOT want Joba to not progress this year. It is a very, very bad idea to throw out a very valuable year of development just because you feel that Joba is, theoretically, worth a third of a win more in the bullpen than as a starter. He is NOT Jon Papelbon. No, he's better, both in stuff, and the fact that he has the stamina to last more than four innings. Please don't waste his talent in a one-inning role. That would be horrendously stupid.
Posted by: Andrew at February 9, 2008 12:45 AM
Wouldn't you agree that the best case scenario for Joba, and this team, this year and going forward, is to mimic Johan Santana's 2003 season, or 2007 Chad Billingsley's season, where he is a man out of the bullpen and a spot starter for the first two thirds of the season, and then is converted to a full-time starter for the last third and playoffs? Johan Santana started Game 1 for the Twins against the Yankees - and that was the only win the Twins got in the playoffs that year.
___
I agree with this. There is no reason to overcomplicate the analysis. The way to maximize Joba's contribution is use him in the pen early on to reduce his aggregate IP, and then move him into the rotation at the point at which is innings cap won't be overrun.
Posted by: Rich at February 9, 2008 01:03 AM
Yep. Start him in the pen. When Moose falls apart, and he will very likely fall apart, then move Joba to the rotation. Even if Moose holds up, history tells us one or more of the starters will get a tired arm, or a sore shoulder, or a sprained ankle, etc... Hopefully we are in a position to make a mid-year bullpen addition from a team looking for a salary dump or one of the other young guns has shown the goods in the minors and is ready for a mid-year call up.
Posted by: #15 at February 9, 2008 07:57 AM
~~~I'm curious- why did you list RSAA and then go with Win Shares to make your point? It seems like the disparity in RSAA is much greater between the best starters and the best relievers. ~~~
I used RSAA to show me who were the best pitchers in terms of their performance, and then went to WSAB, for them, to show me what that performance meant in terms of wins.
Posted by: Steve Lombardi at February 9, 2008 08:49 AM
~~~Secondly, I guess I've confused myself on the nature of "average" and "bench" in these cases. Are these defined separately for relievers and starters, or is there some "pitcher" baseline? ~~~
As per the Hardball Times:
http://www.baseballgraphs.com/main/index.php/site/blog/?id=A125_0_1_0_C
Posted by: Steve Lombardi at February 9, 2008 08:54 AM
~~~Wouldn't you agree that the best case scenario for Joba, and this team, this year and going forward, is to mimic Johan Santana's 2003 season, or 2007 Chad Billingsley's season, where he is a man out of the bullpen and a spot starter for the first two thirds of the season, and then is converted to a full-time starter for the last third and playoffs?~~~
I would have no problem with this approach. But, again, this means he starts the season, and spends most of it, in the pen - and does not start the season as a starter. And, that's basically what I'm saying here - that he should not be a starter this year...or, at the least, not for most of the season. And, if you do that, it's not costing the team a ton of wins...that many like to say it would...because a starter gives you more innings and prevents more runs (as they often like to mention).
Posted by: Steve Lombardi at February 9, 2008 08:58 AM
If he's in the bullpen, Joba's likely to be the main setup guy to Mo, meaning he'll be pitching innings with higher leverage, making his runs saved more important. Given, 150 IP as a starter or 80 as a reliever, the reliever innings are probably slightly more valuable. I agree a combination would be best.
On a technical note, we really should be comparing performance to replacement level, and the replacement level for starters is a lot worse than for relievers (about one run of ERA difference).
Posted by: Sky at February 9, 2008 09:14 AM