Keeping Things in Order
Everyone seems to have one these days. No, I'm not referring to yellow wristbands, a negative view of Bode Miller, or a crush on Evangeline Lilly. I'm referring to ideas for the optimal lineup for their favorite baseball team.
Way, way back in November, I doodled out some lineups for the Phillies versus right and lefthanded pitching [somewhat obsolete now that Jason Michaels is no longer with the team]. When some Phloggers convened recently, much of the discussion centered around how to balance out the Phillies lineup in terms of right and lefthanded batters. Several weeks ago, Daily News scribe emeritus Bill Conlin was touting a lineup that featured David Bell batting second in order to breakup the Black Hole and have someone to play small ball behind Jimmy Rollins. More recently, Cyril Morong devised a mathematical model to determine the most optimal lineup for any team. Since it's a bit complicated to replicate for your team, David Pinto has created a tool on Baseball Musings to make the process much more user-friendly.
Using projections published in the 2006 edition of The Bill James Handbook, and Pinto's tool, I've revisited the issue for the Phillies. According to the model and tool, the best Phillies lineup would produce 5.446 runs per game [using the 1998-2002 data set; for the pitchers, I used my own "projections" for the staff as a whole]:
Optimal Handbook Proj. Lineup OBP SLG Abreu .415 .498 Clutchly .360 .505 Jimmy .338 .424 Howard .397 .628 Pat .368 .480 Rowand .334 .438 Lieby .342 .437 Pitcher .150 .175 Bell .324 .386
No, that's not a mistake on my part, the lineup tool does indeed place the pitcher in the eighth slot. From what I've read, it is not an uncommon result. After having examined the lineups other websites have posted, I think what tends to be an optimal lineup is one that separates the "easy outs" as much as possible in the lineup. And if you recall, last year's Phillies' lineup of Bell, Lieberthal, the pitcher, and then Jimmy Rollins once the lineup turned over featured -- at least for most of the season -- four relatively easy outs in a row. As a matter of fact, according to the lineup tool, that lineup should theoretically produce 4.971 runs per game, given last season's statistics.
Here's another lineup, one that Charlie Manuel has often alluded to as a strong possibility for 2006: Rollins, Rowand, Abreu, Burrell, Utley, Howard, Lieberthal, Bell, and finally the pitcher. Again, using The Bill James Handbook projections, that lineup should theoretically produce 5.120 runs per game. By switching Bell to second and Rowand to seventh while bumping Lieberthal down to eighth, similar to what Conlin suggests, the lineup would theoretically yield 5.084 runs per game.
It is a fun model to play with, however I think it places too much independence on the players in each slot in the lineup with no regard for how two players -- one on base, one at the plate -- might work together as a team to advance the runner.
What the model does point out is how lineup order matters less than we tend to think. Pinto has added a feature that displays not only the best lineups, but the worst as well. For the Phillies, the difference between the best and worst is only .751 runs per game.
That's a pretty small difference when you consider the predicted worst lineup has the pitcher and David Bell batting first and second.
-----------------
A reminder on the BS&S.com Reading Group. If I am the only one in it, it won't be a group will it?
