Saturday, February 25, 2012

Best Dodgers Lineup

Every season there is much talk about lineups. Which lineup will score the most runs and win the most games. There are many theories about how the batting order should be constructed. Speed and on base guy(s) near the top followed by your best power hitters. I have put this theory to the test using my simulator, which plays 100K games using different lineups to see which one produces the most runs and wins. In this exercise I made the Dodgers the away team, always giving them 9 at bats in every regulation length game. I kept the home team constant, and in this exercise I had the Dodgers facing the Washington Nationals. I had the Nationals throw (LHP) Gio Gonzalez and (RHP) Jordan Zimmerman in an attempt to find the Dodgers best lineup to use against both left handed and right handed pitchers. Of course I did not run through every single permuation of lineup possibilities, but after a while you can easily hone in on the best ones. And without any further ado, here is the Dodgers best lineups vs LHP and RHP. Note that in this exercise, I used the 2012 Zips Projections for all Dodgers hitters.

vs LHP Lineup vs RHP Lineup
A.J. Ellis A.J. Ellis
Andre Ethier Andre Ethier
Matt Kemp Matt Kemp
Juan Rivera James Loney
James Loney Juan Rivera
Juan Uribe Juan Uribe
Mark Ellis Mark Ellis
Dee Gordon Dee Gordon
Pitchers Spot Pitchers Spot

Notes: Zips loves A.J. Ellis' OBP and hates Dee Gordon's OBP. The simulator is pretty consistent on which lineup it likes the best facing both a LHP or a RHP pitcher with the only difference being the swap of Loney/Rivera in the number four and five holes. Facing a RHP the Dodgers did score slightly more runs with Adam Kennedy playing 2B instead of Mark Ellis, but the runs were more than lost on the defensive side. How much does it hurt the Dodgers batting Gordon and M.Ellis at the top of the lineup as opposed to what the simulator says is the best lineup? Against the LHP around 1 win and against the RHP lineup around 0.15 wins. Probably not enough to get red in the face about, especially against RHP.


Anonymous said...

Does the simulator account for runs scored in part due to stolen bases? For example, does it account for Gordon getting on base and stealing 2nd and/or 3rd and scoring on a sac fly or scoring from 2nd on a single?

Table said...

Did you compare the run scoring of a lineup with Kemp 4th and Kemp 3rd? If not then your result are faulty

Table said...

It also does not make sense that you have a lineup for left handers and one for right handers, not if you used Zips which does not project platoon splits. Could you explain this?

Xeifrank said...

I'll try to answer the questions in order.

Yes, the simulator takes in to account speed. Each player is given a speed score which is used in the simulations for such things as advancing bases, likelihood of stealing and how often they get caught.

Yes, I compared a lineup with Matt Kemp hitting 3rd and 4th. The simulation wants Matt Kemp hitting 3rd and not missing out on too many plate appearances but not hitting too close to the pitchers spot in the lineup (2nd).

I used Zips projections and then adjusted for splits (vs LHP, RHP etc). I use a one size fits all formula for split adjustment for the Zips projections. I normally run this with my own set of proprietary projections which account for actual "real" splits a little bit better. But if used my own set of projections then people would question that black box too, so you can't win. :)

Thanks for all the great questions.

Anonymous said...

Have you updated the mock draft software for 2012? It is greatly appreciated.

I won a title last year, anchored by Kemp and Kershaw.

Thank you very much.

Xeifrank said...

2012 Mock Draft software will get updated by March 7th.

Anonymous said...

I was looking for the software, but Im not seeing it?
Soccer Drills
Can you let us know the location of the software?