Lineup Analysis

I promised a few days ago to do some lineup math. I'm feeling lazy and obligated simultaneously, so I'll do it, but half-assedly.

The lineup I suggested previously: Mauer, Cuddyer, Kubel, Morneau, Young, Lamb, Harris, Everett, Gomez.

Baseball Musings has a lineup analysis tool. You can see how it works here. Keep in mind that the only things taken into consideration are OBP and SLG (I used 2008 PECOTA projections). You can see the full results here.

The lineup I posted a few days ago would score 4.712 runs/game on average. The best possible lineup (according to this tool) would score 4.730 r/g (Mauer, Kubel, Lamb, Morneau, Cuddyer, Young, Harris, Everett, Gomez), and the worst 4.413 (Everett, Gomez, Kubel, Harris, Mauer, Cuddyer, Morneau, Young). Just to see what would happen, I plugged in cmathewson's Delmon Young prediction (.357 OBP/.507 SLG) instead of PECOTA's (.329/.458). The result? 4.827 r/g using the lineup from a few days ago, and 4.848 using the "optimal" one. As giving Young a .357 OBP immediately vaults him to second place on the team in projected OBP, the optimal lineup with cmath's Young in it is Mauer, Young, Lamb, Morneau, Cuddyer, Kubel, Harris, Everett, Gomez*.

With a more Gardenhire-ish lineup of Gomez, Mauer, Cuddyer, Morneau, Young, Kubel, Lamb, Harris, Everett: 4.605.

Conclusions? I knew you'd ask that... The difference between optimal and least optimal is .32 r/g, or 51.84 runs/season--this is the maximum impact that lineup construction could have on the '08 Twins. The difference between the optimal lineup and the "Gardenhire lineup" is 20.25 runs/season--over half the difference between least and most optimal goes away simply due to Gardenhire not being THAT stupid. The difference between my lineup and the Gardenhire linup is 17.3 runs/season. My lineup is close to what Gardenhire might do if forced to bat Gomez last (although he might just replace Everett with Punto and have Punto lead off and Gomez hit last); batting Gomez last and Kubel in front of Morneau is worth about this many runs.

The difference between the optimal lineups with the PECOTA Young vs. the cmath Young is 21.22 runs/season. Even with these extra 21 runs this lineup still only scores 785 runs this year.

Anyway, play around with this if you're so inclined. There's lots of interesting stuff one could do with this, but like I said, I'm lazy.

*There are some pretty interesting configurations that show up near the top of the optimal list for each set of OBP/SLG entered. For this scenario, consider the following: