[On the evening of October 4, 2005, this blog opened for business. As he was training me over IM, Blez instructed me on a great many things. Posting diaries; uploading files to the Scoop platform; if I were ever to travel back in time, not to step on anything because even the tiniest change could alter the future in ways I couldn't imagine. That last one was a bit odd, but it didn't compare to what came next: "Oh, and mix in some guest authors. I recommend you do so, say, every one hundred and sixty-seven posts." Sure enough, Blez was ultra-prescient, because post No. 168 at this site will be the first-ever guest post. It's my pleasure to hand over the conch of wisdom to Scott M. Collins of Nationals Farm Authority. - Basil.]

Lineup Construction for the Washington Nationals

by Scott M. Collins,
Guest Writer

Alfonso Soriano trudged out to left field for the Washington Nationals for the first time today. If he proceeds to comply with the team's wishes and plays there during the season before leaving whatever team he ends the year with and signing to play 2B with the highest bidder next year, what would the team's lineup look like? Here is the most likely lineup based on the strengths of the players and the proclivities of the team (traditional lineups and a love for L-R-L):

There's not enough L-R-L in there so Soriano or Dutch may switch places with Vidro vs. RHP.

The question remains, are these the most productive lineups? In the past month, there has been an increase in discussion about lineup construction. Here is the original statistical study, a script to implement the numbers, a very fun implementation of said script, some great summarizing notes on these articles and an article on all the above.

Lineup construction is one of my favorite topics even though the perfect lineup may only give a team 20-40 more runs over the course of the season. Seriously, as much as I enjoy tinkering with lineups, analysis shows that as long as the team doesn't do something stupid like give a ton of ABs to guys who have an on-base percentage (OBP) of around 0.300 (ahem... see the 2005 New York Mets), the team's runs per game won't change dramatically. However, when your favorite team only scored 3.94 runs per game (R/G) in 2005, you need everything you can get...

That being said, here is a brief summary of what to look for from each slot in the lineup (adapted from the summary linked above):

1. Best OBP, bar none.
2. Best combination of OBP and slugging (SLG)
3. NOT your best hitter, but say your 5th best hitter so as to spread the easier outs away from being congested in the bottom of the lineup. Whoever doesn't fit into the other slots (wild, huh?!)
4. Best SLG, almost bar none.
5. The second choice for lineup slot #2
6. Second best in SLG, OBP doesn't matter much here.
7. The second choice for lineup slot #6 - usually with a bit more OBP and less SLG
8. The WORST hitter - so as to put space between him and the best hitters (#1-5).
9. The punchless wonder - nearly no SLG, hopefully tolerable OBP

The team would never do this and would be ridiculed by those who don't understand the intent even if they did do it. However, the team has talked about batting Johnson second. If you move everyone down a slot and place the #9 hitter at leadoff (like the team wants), place the pitcher back in the #9 slot (since no one understands putting the pitcher in the #8 slot unless they can bat like Dontrelle or Livan), slide Clayton to #8, and bump Schneider down to #7 (keeping Guillen-Dutch-Soriano as the 4-5-6 all right-handed special), you have a lineup the team might use at some point:

Did we just come full circle? I'm not sure, but I'll leave as an exercise for the student the task of finding out the dramatic increase in R/G if a pre-2005 All-Star Break Ryan Church (AVG/OBP/SLG/OPS = .325/.381/.544/.925) is swapped into CF for Watson/Byrd even though the team would not have a "traditional" lead-off hitter. [ED: click on "Read More" for the answer.] Come to think of it, with Brad Wilkerson, how long had it been since they've had one anyways?

(Answer: R/G with Church in CF = 5.293 for an increase of over 0.3 R/G or 50 runs on the season)