Lineup Analysis

There is a great tool available online via Baseball Musings called the Lineup Analysis Tool.  Basically, you input the names, OBP and SLG for the nine guys in the starting lineup and through a series of calculations, the software tells you how many runs that lineup will score (on average).  Sounds cool, huh?  Well, it goes a step further.  It also spits out the best and worst alternative lineups using the same names and stats.   

So, if you are curious if batting Izturis third is a problem, you can test your hypothesis with this tool.  If you think that Napoli, no matter how low his BA gets, is the better offensive force behind the plate, you can test that with this tool as well.  So, this is the first post in a short series of work I will publish on my recently resurrected blog Haloville - Land of Angels and on Halos Heaven dealing with the 2008 L.A. Angels, their offense and their lineup.

For starters, here is the lineup (w/ rough 2008 stats) that I used as a baseline.  This is the lineup that Scioscia is using right now.  It is a lineup without Matthews Jr. and with Jeff Mathis catching.  Similarly, I named Izturis the starting shortstop because he is getting the lion's share of starts right now.

• Figgy (.372/.319)
  
• Kotchman (.330/.445)
  
• Izturis (.325/.350)
  
• Vlad (.347/.482)
  
• Hunter (.342/.479)
  
• Anderson (.308/.398)
  
• Kendrick (.355/.502)
  
• Rivera (.308/.422)
  
• Mathis (.302/.376)

Now, these numbers will shift from moment to moment and day to day.  I used the numbers from Sunday night, July 27.  Hopefully, you are screaming at your computer right now.  Because these are not good numbers.  Rivera is raking, and most of these numbers are below the career norms for these players - especially Vlad who sports a career line of .390/.575.  But, I am going to run many numbers - career numbers, three-year average numbers, 2007 numbers, 2006 numbers, etc.  So, I should cover a wide array of possibilities.  

So, the above lineup - with those stats - will average 4.795 runs per game.

The optimal ordering of those players yields an average of 4.870 runs per game.  A modest increase.  But how is the lineup different to account for an extra tenth of a run per game.  Here is how:

• Figgins
  
• Kendrick
  
• Kotchman
  
• Guerrero
  
• Hunter
  
• Anderson
  
• Mathis
  
• Rivera
  
• Izturis

That lineup LOOKS better intuitively.  Why?  Well, Kendrick should not be hitting below 5th in the order.  He leads the team in OPS and is just a better hitter than most everyone on the team right now and the fact that he is routinely hitting behind Izzy or Anderson - not to mention Aybar! - is silly.  He needs AB's.  A lot of them.  But, that is the only glaring change from what Scioscia is putting out there on a daily basis.

What if we replace Anderson with a favorite of mine - OBP machine Reggie Willits?  Well - using his 2008 numbers seems unfair.  He hasn't played much at all and has not posted good numbers.  What if we use his career line - something bolstered by his great season last year and slightly diminished by his sporadic play and performance this year.  That line is: .380/.325.  

Here is the lineup I used with Willits in place of GA:

• Figgy
  
• Kendrick
  
• Kotchman
  
• Guerrero
  
• Hunter
  
• Rivera
  
• Mathis
  
• Izturis
  
• Willits (.380/.325)

This lineup generated 4.932 runs per game.  That is an increase of .14 runs per game.  Which means every week the Halos score 1 more run.  That means that the team scores about 23 more runs per year and that equates to roughly 2 extra wins a year if all else remains equal.  That isn't a small increase in offense over the course of the year.

Also, the tool changed the lineup to create even more runs!

To get to 4.96 runs with this crew, the tool spit out the following lineup:

• Willits
  
• Kendrick
  
• Kotchman
  
• Guerrero
  
• Hunter
  
• Mathis
  
• Izturis
  
• Rivera
  
• Figgins
  

This lineup would score about 27 more runs per year.  Now we are approaching 3 extra wins by putting the right guys in the lineup.

What if we start making adjustments based on past performance.  Like - using Rivera's career numbers instead of his weak 2008 numbers.  This is fine because at 30, Rivera isn't in decline and as we have seen - he seems healthy and productive.  If we used 2006 numbers that would be a bit optimistic.  Using his career numbers seems fair.  And, how bout we replace Mathis with Napoli to see what that does.  I won't change the order, just plug in new guys w/ their new numbers.

Here is the lineup I'll look at now - this is a fairly safe bet for what Scioscia would put out there...

• Figgy
  
• Kotchman
  
• Izturis
  
• Guerrero
  
• Hunter
  
• Anderson
  
• Kendrick
  
• Rivera (.338/.470) career
  
• Napoli (.308/.472) 2008
  

This lineup generated 4.964 runs per game.  Now we're getting somewhere....

When the software optimizes the lineup we move to 5.086 runs per game with a lineup of:

• Figgy
  
• Kendrick
  
• Rivera (!)
  
• Guerrero
  
• Hunter
  
• Kotchman
  
• Anderson
  
• Napoli
  
• Izturis

So, comparing this lineup with the first one I listed above we get real change.  The Halos will score 47.1 more runs by replacing Mathis w/ Napoli (when Nipples is healthy, of course) and moving Rivera and Kendrick up in the lineup and Kotchman and Izturis down.  

That is 4 to 5 wins over the course of the year.   I'll continue to work on this and post another entry in the coming days...