Larry Mahnken and SG's

Replacement Level Yankees Weblog

"Hey, it's free!"

The Replacement Level Yankees Weblog has moved!  Our new home is:

Larry Mahnken
Sean McNally
Fabian McNally
John Brattain

This is an awesome FREE site, where you can win money and gift certificates with no skill involved! If you're bored, I HIGHLY recommend checking it out!


Disclaimer: If you think this is the official website of the New York Yankees, you're an idiot. Go away.

September 8, 2006

The Roster Crunch
by SG

With the increasing likelihood of the return of Hideki Matsui, and the potential return of Gary Sheffield down the line, the question that many are asking is, what would give the Yankees their best lineup in the post season? Does bumping Melky Cabrera to the bench and putting Jason Giambi or Gary Sheffield at first base hurt the team more than it helps. There's no way to know for sure, but we can try to estimate using the numbers we have at our disposal. It's still not very likely that Sheffield will be back this year, as he's not even taking batting practice this season, but there's a chance of it.

Since both Sheffield and Matsui have not really played enough this year, I'll assume that they would play to their PECOTA projections coming into the season, although we don't know how the injuries they've suffered may affect that. It's a risk that two guys with pretty significant wrist injuries whose values are almost solely in their bats will play as well as they had projected to entering 2006, especially with all the missed time and the likely lack of meaningful AB they will get(especially Sheffield) before the games really start counting.

Basically, if the Yankees don't collapse, they enter the postseason with the following definites.

CF - Johnny Damon
SS - Derek Jeter
RF - Bobby Abreu
3B - Alex Rodriguez
2B - Robinson Cano
C - Jorge Posada

So, they have LF, 1B, and DH to fill, and they will potentially have the following candidates.

Jason Giambi
Craig Wilson
Melky Cabrera
Hideki Matsui
Gary Sheffield
Bernie Williams

Obviously, offensively, you'd want to get Giambi, Sheffield and Matsui into the lineup if their health allows it. However, we should also consider defense, where Melky is a pretty big upgrade over Matsui. But, it's more complicated if you DH Matsui to keep Melky in there, as you end up putting Giambi at first, where's he's not been particularly good over the last two season.

We don't know how well Sheffield would take to first base. He's an indifferent OF, but he did play 3B in his younger days, although he last saw regular duty there 13 years ago. Sheff's pretty athletic, so I'd imagine he would be okay at first, but it's probably a stretch to assume he'd be even average.

Here are the offensive and defensive run values for each player, using linear weights and zone rating converted to runs on defense. I'm not position-adjusting the offensive values here, as we're trying to compare these players directly to each other. This entry from July has the details on the defensive numbers I'm using, for those who want more details. For everyone but Matsui and Sheffield, it's based on YTD performance, for them I'm using their PECOTA projections coming into the season. Per request, I've added the '06 performances for Sheffield and Matsui now as well.

Since we have small sample sizes for Sheffield and Matsui's defense as well, I'm using their 2003-2005 defensive numbers for defense. For Craig Wilson I'm using his combined offense for Pittsburgh and New York, and his defensive totals for his career as a 1B. I left Aaron Guiel's offensive number alone, since if he plays, he'd probably be platooned, and his career line vs. righties is close enough to what he's hit this season overall. For his defense, I'm just going to assume he's average. So, if we break down those numbers on a per game basis, we get the numbers below.

So now comes the fun part, figuring out which combination is best. Here's the first and most likely combination given Sheffield's uncertain status.

So this combination of players would be worth .624 runs above average on offense, and -.154 runs on defense, on a per game basis, so net value is .469 runs above average per game.

If Sheffield comes back as a DH, here's what we get.

More offense, less defense, and an overall better option than combination one.

But what if Sheffield can play first base? There's no way to know how good he'd be, so I ran a few different options here.

If Sheffield's an average defensive 1B?

If Sheffield's a slightly below average defensive 1B (equivalent to say -6 over 162 games)?

If he's bad, but not quite Giambi bad (-18 over 162 games)?

Obviously, if he's no better than Giambi you're back to Combination 2 above.

As much as it sucks for Melky, if Sheffield comes back, Melky belongs on the bench, even though it hurts the defense. Of course, this assumes that Sheff comes back and hits as he was projected to entering the season. As I said earlier, with a pretty major wrist injury, that's far from a given. Also, it's possible that Melky's steadily increasing SLG is indicative of a change in approach and that his overall value is depressed by his lack of power early on, but there's no way to know that for sure, so I have to go with the numbers we have.

This doesn't consider platoon advantage, which is where Wilson/Guiel come into play. If Sheffield can't come back, then you have Matsui, Melky, Giambi, and a Wilson/Guiel platoon for first base. The question then is if the Yankees are better off with combination 2 above (Melky LF, Giambi 1B, Matsui DH) or some grouping like below. I'm going to give a 10% boost to Wilson based on historical platoon advantages and what I've read about the predictability of platoon splits. Again, I'm assuming Guiel is an average 1B, as we only have 45 innings of data for him there, where he's been perfect (ZR of 1.000).

More bad news for Melky, as this combination would appear to be better as well, although the difference is smaller.

I'm putting up a fifth combination by request from Jody, even though I don't think it has a chance in hell of ever happening.

Lastly, by request, here are all those same combinations if we use Matsui and Sheffield's 2006 performances instead of projections.

There are a bunch of other factors that should probably be considered. It may be better to risk a bad OF defense and get Giambi off the field when Wang is pitching, for example. Conversely, there may be games/matchups where you want to have your strongest defensive outfield. Also, Don raised a good point in yesterday's comments that Giambi needs at least some playing time at first to be able to handle first base passably if the Yankees do somehow reach the World Series. So these numbers do need to be tempered with lots of other considerations.

The most encouraging thing about all these numbers is that whether Sheffield comes back or not, the Yankees have enough quality depth on their bench that they probably can't run a bad lineup out there (unless Bernie sees the OF, which he really should never do). The best thing is obviously if Sheffield can come back healthy and play an average 1B, but even if he can't, they'll be in pretty good shape. They also have the bench to upgrade the defense in the late innings if they have a lead, where the value of a run prevented is probably more important than a run added. A lot of that depends on how many pitchers they feel the need to carry in the postseason. I'll probably play pick the postseason roster in a few weeks, if they don't blow it.

It's on to Baltimore for a three four game set. Here are your matchups.

C. Lidle (3-2, 3.38) vs.E. Bedard (12-9, 3.94)

C. Wang (16-5, 3.69) vs. Cy Loewen (5-4, 5.55)

M. Mussina (13-6, 3.72) vs. H. Penn (0-1, 108.00)

R. Johnson (16-10, 4.76) vs. either R. López (9-15, 5.97) or TBD

Ugh, I was hoping they'd miss Cy Loewen. Maybe they'll finally figure out how to hit that SOB this time. I think 3 of 4 is reasonable, although Baltimore's been pitching pretty well lately. Maybe a split is enough.