clock menu more-arrow no yes

Filed under:

Lineup Construction, the Tango Way

New, 372 comments

For those who have never read a thing Colin has ever posted, Tom Tango is an expert when it comes to sabermetrics and wrote The Book: Playing the Percentages in Baseball back in 2007.  The Book uses methods such as linear regression and Markov chains to help prove and disprove common long-held strategies managers have always used.  The chapter I found the most interesting was the fifth, which took a very in-depth look at the true importance of batting order.

In what is now old news, it turns out that batting order means relatively little since it is essentially set in a loop that only ends with the 27th out.  It is possible to get more runs out of that order through some optimization techniques.  Ideally, you would like your best three hitters (using wOBA as our metric) in the 1, 2, and 4 slots.  The player hitting second should walk more than the clean-up hitter, who should have more power than the man hitting second.  The lead-off hitter should be comparable to the 2 and 4, but with less power as homeruns from the lead spot are less valuable.  The next best hitters should bat 5th and 3rd respectively.  After that, you can just plug the rest of the players into spots 6-9 in order by descending wOBA.  Pretty simple right?  Well, here's what you get when you try this using CHONE and Marcel projections for our 2010 Chicago White Sox.

I decided to use both the CHONE and Marcel projections from Fangraphs, so as to cover for any flaws by the other.  Bill James' are just a bit too optimistic for me.

Marcel
Player wOBA BB% ISO
Pierzynski .314 5.0 .134
Konerko .346 10.9 .202
Beckham .353 9.6 .181
Ramirez .327 6.9 .147
Teahen .324 7.8 .143
Pierre .316 6.0 .076
Rios .338 7.0 .170
Quentin .361 9.4 .223
Jones .307 11.3 .180
Kotsay .305 7.8 .127

 

Using these values, Tango's reasoning, and whatever it is I bring to the table, I constructed the following batting order.  It would never happen in the real world, as you'll realize as soon as you look at it.

  1. Alex Rios-CF
  2. Gordon Beckham-2B
  3. Alexei Ramirez-SS
  4. Carlos Quentin-RF
  5. Paul Konerko-1B
  6. Mark Teahen-3B
  7. Juan Pierre-LF
  8. A.J. Pierzynski-C
  9. Crap Taco-DH

I decided to deviate away from the set plan slightly by putting Rios at the top of the order.  I did this because out of the top four hitters in wOBA, he has the lowest projected ISO which makes hitting him leadoff waste less homeruns than putting Konerko there.  On days when Kotsay is DH, that bottom half of the lineup is a whole bunch of left-handed mediocrity and garbage.  Kotsay and Jones are interchangeable really, as Marcel sees them only sporting a .002 difference in wOBA.

CHONE
Player wOBA BB% ISO
Pierzynski .315 3.7 .136
Konerko .356 10.1 .207
Beckham .350 9.0 .182
Ramirez .329 6.5 .150
Teahen .328 7.5 .161
Pierre .310 5.2 .081
Rios .338 7.0 .169
Quentin .371 9.1 .229
Jones .324 11.3 .203
Kotsay .297 6.9 .109

 

Not too much changes in terms of batting order when you siwth to CHONE.  A.J. and Juan switch spots, and Andruw Jones moves the DH situation up to 7th against southpaws (assuming a strict platoon between him and Kotsay for simplicity's sake).  The name on the left of the slash is the hitter against lefties.

  1. Alex Rios
  2. Gordon Beckham
  3. Alexei Ramirez
  4. Carlos Quentin
  5. Paul Konerko
  6. Mark Teahen
  7. Andruw Jones/A.J.Pierzynski
  8. A.J. Pierzynski/Juan Pierre
  9. Juan Pierre/Mark Kotsay

For those curious, Jim Thome projects in at .360/14.7%/.232 and .351/14.8%/.209 respectively.  This results in the following lineups (left side Marcel, right CHONE):

  1. Beckham
  2. Thome
  3. Rios
  4. Quentin
  5. Konerko
  6. Ramirez
  7. Teahen
  8. Pierre/Pierzynski
  9. Pierzynski/Pierre

I loathe Mark Kotsay more and more every day.