THE BOOK--Playing The Percentages In Baseball

MGL, as a Marlins fan, I can tell you that Fredi Gonzalez (the Marlins manager) indeed has no idea how to fill out a lineup. Even beyond lineup optimization by methods found in The Book, he doesn’t even use OBP to order the lineup. I’m fairly certain he’s using the cliche’s of lineup order (first guy’s fast, second guy is a “move runners over” type, third guy’s the best, cleanup is power, and so on) to do this.

Personally, it’s annoying me that he’s placed Uggla and Ross, two of his best hitters, 6th and 8th respectively, especially with Uggla being one of his better on-base percentage guys.

My apologies, at #1, it should say 7th, not 8th in the lineup.

I’m curious, how would you optimize this lineup MGL? I’m not entirely familiar with The Book’s suggestions, though I’ll be checking it out right now.

I find it interesting following Don Wakamatsu moving Gutierrez all over the lineup.  And when he does it, he does it for stretches at a time.  So, for a couple of weeks he’s 9th, and then 5th, then 2nd, etc, etc.

http://www.baseball-reference.com/players/split.cgi?n1=gutiefr01&year=2009&t=b

As it turns out, the one spot that he’s performed by far the worst was as #2.  I suppose with a young player, it makes more sense to be careful, because, while batting order makes no difference to us, a player himself MAY think it means something.

Tango, you do work for the Mariners. Tell Waka what he supposed to do!

Gonzalez has got his best hitter batting 8th and his worst batting 2nd and Helms *the #2 hitter) is not even the typical fast, bunting type of hitter, so I have no idea what he is even thinking.  Can you get a whole lot worse than batting your best hitter 8th and your worst hitter 2nd?  Also, his second best hitter is batting 6th.

This lineup is just atrocious. It looks like he threw everyone’s name in a hat and starting randomly choosing them.  Even worse than that, it actually looks like he just looked at their 2009 BA with no regard to anything else, like walks or power.

I ran his lineup in my sim and got 4.789 rpg against a RHP who was a little worse than average.  I then ran this lineup:

C COGHLAN LF .293 9 36
D UGGLA 2B .242 23 67
H RAMIREZ SS .364 19 85
J CANTU 1B .276 13 71
J HERMIDA RF .267 13 47
C ROSS CF .269 20 69
W HELMS 3B .274 2 26
J BAKER C .268 8 39

That scored 4.872 or 13 runs per 162 more.  I could probably tweak that lineup a little and squeeze another few runs (per season) out of it, but 13 runs per season is a lot when it comes to lineups, as Tango can verify, and as I said, I didn’t really try and optimize it - I just put the good guys at the top and the bad guys at the bottom.  Of course I am “cheating” as I am using my projections to order the lineup and the sim is using those same projections to generate the runs. If somehow Freddy knows something about these player’s projections that I don’t know and it is reflected in his lineup (which I doubt) then my sim is going to be worthless (or at least of less value than I think) in terms of optimizing a lineup.

MGL,

To be fair to Gonzalez, I’m not sure how you’re getting Hermida as the best projected hitter, as I saw most projections have Ramirez as that guy. Did this take into account pitcher handedness? That would make more sense to me.

Just using ZiPS updated wOBA projections for the season and the suggestions The Book has on p. 132, I got a lineup like this (note: not adjusted for pitcher handedness, this was a quick-and-dirty attempt to optimize):

1) Coghlan
2) Uggla
3) Hermida
4) Ramirez
5) Ross
6) Cantu
7) Baker
8) Helms

It probably wouldn’t be as optimal as the one you used, because as you mentioned, your estimator uses your projections to calculate the rpg, and I’m sure you and ZiPS differ in some places. However, would you mind running that lineup against the same pitcher type? I promise, no more requests, I’ll do my own research work from here.

Right of course Hanley is projected better than Hermida.  My bad.  I have Hanley at 27 runs per 150 and Jeremy at 15.

Of course versus a RHP, Hermida might be better, which is another reason to hate Freddy’s lineup.

Keep in mind I am running 100,000 games on my sim, so the rpg have the appropriate confidence intervals.

Your lineup, Michael, generated 4.910 rpg!  That is 18 runs per year better than Freddy’s.  Again, giving up that many runs, even for one game, is a lot, assuming that that is correct of course.

Not to excuse, but maybe explain. Is Helms “hot” right now? A LOT of managers put far more stock into how “hot” a player is than they should (I could understand dropping a guy who in in a three week slump if him batting .150 is killing the team; I mean a guy goes 5 for 8 and managers move him to the 2 hole).

ALL managers put stock in hot and cold streaks.

“I could understand dropping a guy who in in a three week slump if him batting .150 is killing the team; I mean a guy goes 5 for 8 and managers move him to the 2 hole.”

3 week slumps have ZERO predictive value, so while you may understand it, it is still wrong.

I have no idea whether he has been hot or not, but the fact that he is hitting .274 is likely one of the reasons - I guess.

Thanks for the simulator help, MGL. And regarding Helms, I think Fredi likes a “veteran hitter” at the #2 slot, a guy who can give you “productive outs.” I don’t like it either.