Comments on: 2011 Marlins Season Preview: Optimized Batting Lineups

By: Fish Bites « Marlin Maniac | A Florida Marlins Blog

Fish Bites « Marlin Maniac | A Florida Marlins Blog — Wed, 18 May 2011 14:03:05 +0000

[...] move of bumping Hanley Ramirez up to the second slot in the lineup. Ironically, given the tenets of lineup optimization relayed by The Book, this was actually a great move by Edwin Rodriguez. The Marlins’ lineup [...]

By: Michael Jong

Michael Jong — Tue, 22 Mar 2011 15:34:53 +0000

Isiah,

The hit and run consideration is significantly less impactful for run scoring when compared to simply having a better hitter up at the #2 spot getting more opportunities to hit. In 2010, the #2 slot got 70 more PA on average than the #6 slot in a given season. That’s almost a 2-run difference between batting Morrison and batting Infante at #2.

Consider that the specific situations involving a potential hit and run won’t come up more 18 percent of the time at the #2 spot (less considering it’s the NL and the pitcher bats two spots ahead of #2). Then think about how the hit-and-run not only potentially prevents double plays but also creates outs by inducing bad steal attempts as well, and you can see why it isn’t as serious a consideration. In fact, when considering preventing a potential double play, you might actually want a player with WORSE contact, since a double play generally will not occur on strikeouts.

Managers think about these little plays too often, when the primary consideration should be to get your best players the most opportunities first and foremost. There’s simply no reason to spend 700 PA on Infante when it could be spent on a significantly better hitter like Morrison.

By: Isiah

Isiah — Mon, 21 Mar 2011 19:08:36 +0000

I’ve mentioned this before, but how does the Book take into account the advantage of being able to hit n run? old school coaches have always preferred having a 2 hitter that is great at making contact so that the lead-off guy could run on the pitch, which provides some more safety from double plays (though not so much considering lined outs) and allows runners to advance to third on singles. since infante’s contact numbers are great, and he is faster and a better baserunner than lomo, i’d rather him batting 2nd in general, so it would be a bit easier to get a run early on.

i’d also like to have lomo 3rd and stanton 5th (with Hanley fourth), mainly because lomo takes more pitches (for hanley to see, and it also helps exasperate a pitcher sometimes), but the value of that may be diminished considering that Stanton may get a few more fastballs being before Hanley.

What’s the reason most teams don’t switch over their lineups for lefty-righty splits, is it just egos?

By: Michael Jong

Michael Jong — Sun, 20 Mar 2011 02:32:43 +0000

Xeifrank,

Very interesting stuff, thanks for the sim work on the lineups. I’m not terribly surprised that the vs RHP team hit better; as I mentioned in the blog, it aesthetically looks like the right lineup for the Fish. However, I cannot think of a good reason for that, other than that I did not optimize the lineups properly enough. Those percentages show almost a .02 win difference per game between lineups, translating to close to 3 wins in one season. Seems pretty significant.

By: Xeifrank

Xeifrank — Sat, 19 Mar 2011 17:08:09 +0000

I pitted the two lineups in your post against each other in a head to head matchup. The vs LHP lineup against the vs RHP lineup. I used the same pitcher against both lineups, except I made him left handed for one team and right handed for the other. The pitcher I used was Javier Vazquez, so I made two versions of him (RH and LH). I used the same bullpen (Marlins) and bench for both teams. I played two games, one with each team being the home team. Then I simulated each of the two games 100K times. Here are the results.

FLAvsLHP vs FLAvsRHP
FLAvsRHP win probability: 55.476

FLAvsRHP vs FLAvsLHP
FLAvsLHP win probability: 53.763

So a slight edge to the FLAvsRHP team.

The next step would be to plug in different lineups keeping (WOWY) everything else constant and see which lineup has the highest win probability. Of course there are many permutations, so it could be a science project. Could be reasonably done by selecting 10 or so different lineups.
vr, Xei

vsRHP @ vsLHP