Acquiring Another Starter

Written by Josh W. | 20 July 2010

As you surely already know, Andy Pettitte suffered a groin injury and will be out for 4-5 weeks. Mike Axisa over at RAB already did a pretty good job of describing the Yankees' potential courses of action, so I won't repeat what he said. I would however like to discuss the idea of trading for a replacement.

Before delving too deep into this discussion, it's important to know that the Yankees seem to have already made their decision, starting Mitre on three day's rest yesterday to prepare him to take Pettitte's slot in the rotation.*

Because we know that Mitre is the in-house choice for the Yankees, we can properly evaluate any potential replacement for Pettitte. This is because the value of whoever the Yankees would trade for is primarily equal to the difference in performance of the traded for player over Mitre.

ZiPS pegs Mitre's current talent level at a ERA of 5.08 (also projecting both time in the rotation and the bullpen). For the purposes of this discussion, Let's assume Mitre would perform as a 5.00 ERA pitcher, with an average of 5 innings per start over 6 starts. This would mean that Mitre would allow 16.66666 runs.

How does a player like Ted Lilly compare?

ZiPS puts Lilly's current talent level at an ERA of 3.91. This seems overly generous when looking at Lilly's 4.62 xFIP, but for again for the purposes of simplicity we will assume that Lilly would perform as a 4.00 ERA pitcher. If we assume he would average 6 innings a start over 6 starts, then he would allow 16 runs. So Lilly would allow about 1 run less than Mitre in 6 more innings. Those 6 innings are obviously going to be filled by the bullpen. Yankee relievers this year have an ERA of 4.17. This means that in those 6 innings, the Yankees relievers would allow 2.78 runs, which we will round up to three. Subtract these three runs from Lilly's 16 run total and you get that Lilly would save about 3.44 runs more than Mitre.

The immediate value of trading for a Lilly type is equal to about a third of a win above replacement. If we assume that one win above replacment is equal to about 4.5 million on the free agent market, than a third of a win is equal to 1.5 million. Thanks to Victor Wang, we know that this is about equal to a grade C pitching prospect (as graded by John Sickels) 23 years or older. Without spending too much time looking at Sickel's previous rankings, this would seem like a Romulo Sanchez type of prospect. In other words, the immediate value of trading for a Lilly-type pitcher is not worth very much.

From a short term perspective, trading for a pitcher (that reasonably can be acquired) doesn't make a lot of sense. The Yankees simply don't need the help. They are on the (good) side of the marginal win curve where adding more talent doesn't do a whole lot, at least in the regular season (adding a Cliff Lee type for the postseason is another issue).

(more after the break)

Marginal Win Curve

What I mean by being on the good or bad side of the marginal win curve is that adding 2 WAR to a 70 win team doesn't do a whole lot for that team. There's simply not much difference between a 70 and 72 win team. However, if a 92 win team were adding 2 WAR, it would make a significant difference, because a 92 win team can use all the help it can get to acheive playoff entry, and it is already very close to being good enough to get into the playoffs, so wins have more value. There is a pretty significant difference between a 94 and a 92 win team.

This can be seen in the following graph (made by Nate Silver):

Coolstandings.com projects the Yankees to finish this season with about 101 wins. When you are a 101 win team, you're basically guaranteed to get into the playoffs, so adding more talent simply doesn't make much difference in a regular season perspective.

This discussion gets more complicated when discussing the value of a marginal win to a team in the playoffs, but that doesn't play into this discussion because a Ted Lilly type of player would not make a whole lot of difference in the postseason. Any starting pitcher the Yankees would acquire now would likely be a 5th starter, who would just get moved into the pen in the postseason anyway.

Other Considerations

Does acquiring another starter have other benefits besides simply performing better than Mitre?

Mitre is not a very good pitcher, and consequently does not throw many innings. Presumably, whomever the Yankees would acquire would be a better innings eater. Lilly certainly is. Some may posit these additional innings could help keep the bullpen in better shape. Yet, this is a bit of a shallow argument. Bullpens are built through chaining; pitchers generally have designated roles. This likely means that if Mitre were in the rotation, his spot would still be occupied by a Dustin Moseley or Chad Guadin (his spot has been occupied by these two since he has be on the DL). When another pitcher would be acquired, this would simply push Mitre into the spot of Moseley/Gaudin, and that's simply not a big change (if we consider Mitre, Moseley, and Gaudin to all be replacement level pitchers). Other relievers out of this role would not be affected that much, and relievers this year haven't been overused, so that's not an issue.

The other potential benefit is that acquiring a starter would allow the Yankees to better manage Phil Hughes. As we all know, Phil has some innings constraints on him this year, and in all likelihood, will not throw more than 180 innings. Earlier today Joseph Pawlikowski at RAB talked about planning Phil's future starts. I'll post the relevant (for this discussion) part here: 

Using the 6.1 innings per start guideline, that would bring Hughes to 158 innings heading into September. That’s just 22 innings below the arbitrary 180-inning ceiling they mentioned earlier in the season, so maybe that’s four more starts. He clearly wouldn’t make it through September at that pace, so the Yankees have to hope they have enough of a cushion to skip Hughes a few times in September. Even so, given the tough schedule I can’t see them skipping him more than once. Chances are that if the Yankees don’t have a comfortable lead in the East by mid-September, Hughes could actually hit the 200-inning mark.

It seems that it might be tough for the Yankees to properly limit Phil's innings. Acquiring another starter would obviously help a lot in this situation, but only if the Yankees aren't comfortable using Mitre (or any other in-house option) longer than Pettitte's DL trip.

Conclusion

The Yankees do not need another starting pitcher to suceed in this regular season. They are an excellent team already, and whoever they would acquire would not make a lot of difference in the postseason or in the bullpen.

The importance of adding another starter seems to hinge on the Yankees need to manage Phil's innings. Skipping enough starts could be difficult, and the long-term benefits of having Phil under a certain innings threshold (this year) are very important (to avoid injuries). Yet, it is possible the Yankees could just stick with Mitre to manage Hughes' innings.

It seems to me that adding another starter is unnecessary, as whoever they would add would in all likelihood be not that much better than Mitre or other in-house options. Managing Phil's innings will be difficult, but that can be done with in-house options.