September 9, 2003
Last Friday's chat session, my first with the new Baseball Prospectus chat interface, was a lot of fun for me. I wasn't able to get to even half the questions, but I did enjoy seeing so many BP readers in the queue, and I promise to stay longer next time.
The last question I answered was about the Yankees:
Mike K (NY, NY): Clearly Alfonso Soriano is not an ideal leadoff hitter. Much of his power is being wasted without anyone on base in front of him and his OBP is quite low. I see him as eventually being a 3-4 hitter, though maybe lower in the order this year. But in the meantime, if you were managing the Yanks, who would you lead off?
Joe Sheehan: Nick Johnson, Bernie Williams or Derek Jeter. Not Soriano, basically. My optimal Yankee lineup:
Jeter puts too many balls on the ground to bat second behind a non-stealing leadoff guy, and I don't want to bat him third and move Giambi down.
That response, admittedly off-the-cuff, generated an e-mail from L.M.:
I saw in your recent chat that your ideal lineup includes Nick Johnson as the seventh hitter, ahead of only Aaron Boone and Karim Garcia. But your own numbers show Johnson as one of the elite hitters in the major leagues. I don't get it. Is there some underlying logic behind this lineup contruction?
Before I get into this, let me just say that I love this stuff. I could talk about roster/lineup/role optimization all day, which is just one of the many reasons it's a wonder I'm married. Back in the nascent days of baseballprospectus.com, I wrote a column called Lineupectomy (a couple of which actually show up in the archives), which got its name from something we used to do at Strat tournaments--taking people's teams and creating optimal lineups. It's a geek thing, and as has been pointed out, not remotely the right name for the process, but it's something I spend a lot of time doing.
There's a question as to how much the effort matters. It's something of a stathead truth that the difference between the optimal lineup and a reasonably constructed one is small, less than a win per year. I don't necessarily buy that; as Chris Kahrl pointed out in BP2K1, the simulations on which that idea is based are fairly old, done on ancient technology, and it's possible that we just haven't been able to model it properly yet. I find it hard to believe that doing simple things like getting your OBP guys in front of your SLG guys, making the lineup less vulnerable to attack relievers, and minimizing double plays aren't worthwhile endeavors that can add not just a few runs, but a few wins a year.
My thought process on the Yankees goes something like this: they have seven good hitters, so one of those seven ends up at the back of the line. Ideally, you'd like that to be the worst of the seven, but that's complicated by the fact that players don't change roles as easily as Strat cards, and the collective media and fan base is poised to make a very big deal over any radical changes. Ask Theo Epstein.
With all that in mind, here's a longer answer to the question of the Yankee lineup.
- The three reasonable leadoff choices are Nick Johnson, Bernie Williams and Derek Jeter. All three have most of their offensive value in their OBPs. Johnson is slow, which makes him a nontraditional choice, but not necessarily a bad one. However, batting him leadoff creates a problem of what to do with Jeter.
Given Williams' loss of power this season and his groundball tendencies, I'd be inclined to hit him leadoff, except that he doesn't seem to like the role. Torre hasn't used him there in the last four seasons. That matters, and if hitting him there is going to create problems, it's not worth it. It's not like he's the only option.
Jeter has enough power that batting him leadoff is probably not optimal, but 1) I want him batting in the top three, and 2) as a right-handed batter who hits ground balls, he's a double-play threat in the #2 or #3 spots. With the other two leadoff options unable to steal second--and Jeter a very good percentage base-stealer--Jeter is my choice. The 60 points of OBP he gives the team over Soriano should show up on the scoreboard quickly.
Soriano should not be batting leadoff, I don't care how fast he is or how many bases he steals. He has a .331 OBP, .300 since April. His power is his best feature, and it's largely wasted batting either with the bases empty or behind the two dead spots at the bottom of the Yankee lineup.
- I bat Williams second behind Jeter because he takes a lot of pitches and Jeter is a strong base-stealer. That should serve to negate Williams' double-play balls, which he hits with some frequency (1.60 GB/FB ratio).
Why Williams (.270/.369/.406) instead of Johnson (.297/.438/.496), which is the unspoken question in the above e-mail? I'd rather have the switch-hitter ahead of Jason Giambi, and if I don't bat Williams second, I may have to bat him seventh, and that's a bigger headache that using Johnson so low.
- While Joe Torre has dropped Giambi to fourth of late, I prefer him third, as he's the Yankees' best hitter. I don't see the reason for batting his. 400 OBP any lower in the lineup than is reasonable.
- Soriano's low OBP and right-handed power make him a fairly good cleanup hitter, although you could argue him down to seventh. Almost all of his value at the plate comes from his power, which is exactly what you want in a #4 hitter. His right-handed bat breaks up the Yankee lefties as well, and batting him fourth will create less of a frenzy than dropping him to seventh.
That's important, the same way finding a spot Bernie Williams can accept is important, because lineup optimization isn't worth creating headaches. Dropping Soriano to seventh--a reasonable idea--would create headaches. Scoring as many runs as possible has to be the primary offensive goal, and balancing that with keeping players happy is not an enviable task. Where you draw that line isn't for me to say. Perhaps Soriano would take a move to seventh well, and Joe Torre would be able to keep it from becoming a controversy. From here, it really doesn't matter; Torre has made it clear that he's not moving Soriano out of the leadoff spot.
- After #4, you're left with three hitters who possess similar skills. I'd go Hideki Matsui/Jorge Posada/Johnson to separate the lefties with the switch-hitter, but there are other reasons. The #5 slot leads off more innings than any one other than the #1 slot. Matsui runs better than Johnson and Posada, and hits more ground balls than they do, so if he's on base for them he can move around better on their hits, and this arrangement should minimize double plays.
That's the stathead approach, the one that wouldn't have been using Alfonso Soriano in the leadoff spot for two years. It's nonsensical to take ABs from the three high-OBP players so that the low-OBP can bat with no one on.
If Soriano has to bat leadoff, I suppose the best thing to do is what Torre has settled on the past few days: Johnson/Jeter/Giambi. That breaks up the lefties and righties, although it does have some of the problems (Jeter in DP situations, Giambi fourth) addressed above.
As long as Torre insists on keeping Soriano and his .330/.500 in the leadoff spot, the Yankees will have not just a suboptimal lineup, but one of the least effective ones they can put together.
Joe Sheehan is an author of Baseball Prospectus.
Click here to see Joe's other articles.
You can contact Joe by clicking here