Fun with line ups

This week Lynn Henning wrote an article speculating on the Tigers lineup for next season.

Henning in addition to penning his own lineup, concluded the article with this:

Much of the manager’s disposition there is a matter of what you think about the computer-generated wisdom that it’s better for a team to load the front end of its order with its best hitters.

That, of course, can collide with a manager’s intuition, and even more, his experience.

“I know a guy who did that one game,” Leyland said of one manager who followed the best-hitters-at-the-top formula. “And his team got shut out on two hits.”

First of all, Jim Leyland citing that sample size of 1 is kind of nonsensical. On a day when your team manages 2 hits the order is irrelevant so it does nothing to refute the premise. Second, it almost makes it sound like Henning is on board with the intuition-over-science angle. But when you look at Henning’s line-up…

  1. Granderson
  2. Guillen
  3. Sheffield
  4. Ordonez
  5. Casey
  6. Monroe
  7. Rodriguez
  8. Inge
  9. Polanco

I took the Detroit Tigers ZiPS projections and ran them through the lineup analyzer using Henning’s line-up. Henning’s lineup generates 5.310 runs per game.

The optimal line-up, which would never be implemented for political and ego based reasons produces 5.342 runs per game and stacks up like this:

  1. Sheffield
  2. Guillen
  3. Casey
  4. Ordonez
  5. Granderson
  6. Monroe
  7. Inge
  8. Rodriguez
  9. Polanco

Now the second most optimal lineup is very similar to Henning’s, but modified for generally-accepted-baseball-practices and ego reasons:

  1. Sheffield
  2. Guillen
  3. Granderson
  4. Ordonez
  5. Casey
  6. Monroe
  7. Inge
  8. Rodriguez
  9. Polanco

Now what we see with the Tigers lineup options is that like last year, they look to have a pretty balanced lineup. With the same cast of characters the least optimal arrangement produces 5.051 runs per game.

So what happens if you change the mix of people? Let’s swap in Thames for Monroe and Shelton for Casey. This mix produces lineups that range from 5.110 to 5.404

Does it really matter?

Because the Tigers lineup is balanced, the sequencing becomes less important. The simulations show what happen over the long haul and it doesn’t do much to predict what will happen in tomorrow’s game. Plus, the fact of the matter is a standard lineup barely exists anyways. Between rest days, platoons, and injuries teams typically use more than 100 different lineups during the season (Jim Leyland used 120 different lineups in 2006).

Henning’s lineup is very reasonable, and it tests well also. I still feel it is more important to have the right guys playing than the order in which they hit.