Table of Contents

Monday, March 31, 2008

Can you make a better lineup than Dusty Baker?

As I type this, Opening Day is just over 12 hours away. And with it has come Dusty Baker's first official lineup as a Reds manager, courtsey of John Fay:
Corey Patterson CF
Jeff Keppinger SS
Ken Griffey Jr. RF
Brandon Phillips 2B
Adam Dunn LF
Edwin Encarnacion 3B
Scott Hatteberg 1B
Javier Valentin C
Aaron Harang P
Among the most common (and convincing) critiques of Dusty's managerial tendencies is his selection of leadoff hitters. Statheads "know" that OBP rules all in the leadoff slot, right? Patterson and his PECOTA-projected 0.313 OBP is therefore an awful choice. Right?

I'm pretty sure it is. But I'm not positive, and I'd like to know how big of a deal it is. Fortunately, there's an excellent tool that can answer just this kind of question: John Beamer's Markov model from the Hardball Times 2008 Annual. Because it uses Markov Chains (described here), Beamer's model is a much better approach to understanding the dynamic nature of lineup interactions than simple regression equations (like in David Pinto's oft-cited lineup tool).

So, I'd like to set up a Lineup Showdown! To that end, I'm requesting that each of you reply to this post with two lineups:

1. Your "best" lineup using only Dusty Baker's Opening Day Starting Eight.
2. Your "best" lineup with your pick of players from the Reds' organization.


If you want to submit more or fewer than two lineups (e.g. one vs. righties, one vs. lefties), go ahead. I will take all submitted lineups--along with any others that Baker uses this week--and plug them into Beamer's Markov model, along with 2008 PECOTA projections, and report the results back later this week.

Markov will tell us, for each lineup:
1. Its estimate for how many runs it should produce per game.
2. How many times each batting slot should lead off an inning.
3. Complete run expectancy frequency matrices for every out base & out state.

Should be fun!