Optimizing the Pirates’ Lineup

The Pirates had 134 different batting orders in the 2011 season, or a new batting order each 1.2 games.  So why did Clint Hurdle try so many different lineups?  Some of the 134 can be credited to injuries, considering the Pirates used 26 different players in their starting lineup, more than the 25-man roster an MLB team is allowed at one time.  The other part of the 134 different lineups can be credited to there not being an exact science behind lineup creating, especially the last 3-4 spots.

Baseball Musings has an interesting tool that lets the user enter on-base percentage and slugging average for a team’s lineup and optimizes their lineup based on The Book.  Basically, higher the OBP, the higher in the lineup: 24 plate appearances are gained from moving up one spot in the lineup, approximately.  The lineup the Pirates most often used last season, based on appearances in each spot per player, was:

1. Jose Tabata (57)
2. Chase d’Arnaud (25)
3. Andrew McCutchen (82)
4. Neil Walker (66)
5. Lyle Overbay (36)
6. Ryan Doumit (29)
7. Brandon Wood (38)
8. Ronny Cedeno (78)
9. Pitcher (160)

If those players had their stats subbed in, the lineup would have scored 3.426 runs per game, or about 555 runs in the season. 555 runs per season is horrendous, worse than the last-place Seattle Mariners who scored 556 runs. Assuming the pitching had allowed the same amount of runs, the Pirates’ Pythagorean record, using 1.87 as the exponent instead of 2 (link), would have been approximately 62-100, 10 games worse than what the Pirates actually did and 8 games off their actual Pythagorean record. If they had optimized their lineup, they would have scored 3.733 runs per game, which translates to 605 runs scored, or a Pythagorean record of approximately 69-93. 69 wins is 3 wins off their actual record and 1 win off their actual Pythagorean record.  The Pirates’ optimized lineup is below with each player’s triple-slash line next to their name.  For the pitchers, I used Baseball-Reference’s batting splits.

1. Jose Tabata (.266/.349/.362)
2. Andrew McCutchen (.259/.364/.456)
3. Lyle Overbay (.227/.300/.349)
4. Ryan Doumit (.303/.353/.477)
5. Neil Walker (.273/.334/.408)
6. Chase d’Arnaud (.217/.242/.287)
7. Brandon Wood (.220/.277/.347)
8. Pitcher (.094/.136/.133)
9. Ronny Cedeno (.249/.297/.339)

The lineup is sorted by on-base percentage as well as slugging average.  If a batter hits second, third, or fourth, he will have a greater opportunity to drive in the runner.  An example of this is Neil Walker’s runs batted in.  If Walker had hit leadoff, he wold have had a significantly smaller number of RBIs as opposed to hitting cleanup.

With the additions of Barajas, replacing Doumit at catcher, and Barmes, replacing Cedeno at short, and Jones, replacing Overhaul at first, and Alvarez, replacing Wood at third, and Presley instead of d’Arnaud, the Pirates’ lineup would look like this, using Bill James’ 2012 projections:

1. Jose Tabata (.283/.349/.387)
2. Alex Presley (.301/.346/.445)
3. Andrew McCutchen (.277/.368/.455)
4. Neil Walker (.273/.332/.426)
5. Garrett Jones (.261/.325/.451)
6. Rod Barajas (.224/.274/.403)
7. Pedro Alvarez (.252/.332/.429)
8. Clint Barmes (.248/.303/.393)
9. Pitcher ( .094/.136/.133 )

According to Baseball Musings, that lineup will produce about 668 runs over a 162-game season.  That would be, assuming the runs allowed is the same, a Pythagorean record of 76-86. The four-win improvement is right on par with the 3.6 wins that the Pirates added this offseason.  If the lineup was optimized, it would appear like this:

1.  Andrew McCutchen (.277/.368/.455)
2.  Alex Presley (.301/.346/.445)
3.  Neil Walker (.273/.332/.426)
4.  Garrett Jones (.261/.325/.451)
5.  Pedro Alvarez (.252/.332/.429)
6.  Rod Barajas (.224/.274/.403)
7.  Clint Barmes (.248/.303/.393)
8.  Pitcher ( .094/.136/.133
9.  Jose Tabata (.283/.349/.387)

That lineup would score 4.335 runs per game, or 702 runs over a 162-game season. The Pythagorean record would be 80-82.

I highly doubt that Hurdle would ever bat someone like A.J. Burnett eighth, as that would take one bunt situation away per game, but the calculator says that it will be a good idea to do so. While this is all hypothetical, it’s interesting to see hard it is to be an MLB manager to figure out a lineup.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s