THE BOOK--Playing The Percentages In Baseball

Ah, a Cubs discussion.

Does this make any sense as to why Soriano does better with bases empty? Soriano is a free swinger. He is going to swing at a higher proportion of pitches in, and out of, the strike zone, no matter the situation. With bases empty, especially leading off an inning, the pitchers follow their usual strategy of throwing more strikes, so as not to walk the batter. So, a higher proportion of the ‘marginal’ pitches Sori swings at are at least strikes, and so he does better. And vice versa for men on base.

If true, then it’s the pitchers’ fault for not adapting their pitching scheme better to a batter who is not really typical of the slot he occupies.

DS, there are many, many hitters you can say that (and similar things) about. I think you have to give pitchers more credit than that.

I don’t buy that it is anything more than a statistical fluke.  That is, until I see evidence that says otherwise.

One standard deviation is roughly a 15 point difference, so we see here a difference of around 2 standard deviations.

While that doesn’t necessarily mean that Soriano definitely prefers to bat with bases empty, it points very strongly toward that.

That is dead wrong. Sort of.  We don’t really care how many SD’s someone is off. By the time someone points something out to us, OF COURSE it is going to be unusual.  That is the classic example of selective sampling or cherry picking.  Take any distribution that is completely random (no skill whatsoever).  Within that distribution, 5% will be off the mean by more than 2 SD’s.  Well, someone can and will point those 5% (or 1%, or .1% if the sample is large enough) players out, and say, “Well, these guys are way off the mean - something must be going on!”

Until we figure out how much, if any, skill, there is, those SD’s mean NOTHING (because they are cherry picked, as is Soriano’s)!

If there is little or no skill for players batting with runners and and without, which I suspect is the case, as with most ‘splits’, then the # of SD from the mean means nothing, since everyone gets regressed 100% (or near 100%).  Even if there is a little skill, most of that 2 SD will get regressed toward the mean.

So what’s the point of even talking about it until and unless you determine how much hitting with runners on is a skill?

How about we do the same thing for clutch hitting, batter/pitcher matchups, day/night performance, first half/second half, home/road, odd/even days of the week, and we’ll print on a piece of paper the players who are more than 2 SD’s from the mean for any number of PA’s, in each of those categories - about 5% of all players. 

Can I then say (about them), “While that doesn’t necessarily mean that they definitely prefer to bat with ‘X’, it points very strongly toward that.”

Of course not!  So why does Tango say that about this and Soriano?  I don’t know.

Understood MGL. But I merely meant that, IF his split is due to real causes, then this is perhaps why it happens (although I didn’t quite state it that way).

This (where to bat Sori) gets a lot of discussion here in Chicago. The sportswriters always bring up this split. Most of them assume it’s meaningful; a few don’t.

Me, I’m with those that don’t, and wouldn’t hesitate to move Soriano, especially if the Cubs acquire B Roberts.

You are right, if I had the whole list of players, and then just picked off the extremes.

But, that’s not what I did.  I only randomly selected a player, since Soriano was being discussed, and noticed his split.  Unless I was prescient and happened to fall on one of the most extreme splits in baseball, then what I said was true. 

To get to the truth, we’d need to do like we always do, look at the whole distribution of players, and compare to what the random distribution would look like.

I agree that the difference between his splits has no meaning at all and in the long run these would even out if given enough plate appearances.  However, since we don’t know for sure, I’d lean toward being cautious and keeping Soriano in the spot he has done much better in.  In 2006 he hit much better with runners on base and in 2007 it was the opposite.  Overall he’s been better win nobody on base, but I don’t think it means it will continue to be that way.  I just wouldn’t take a chance.  Soriano is important to the Cubs offense and until someone can say for certain that he’s going to perform the same with runners on base and that these historical numbers are an aberration, I wouldn’t take the risk.

Especially since the batting order itself doesn’t matter to begin with.  Given the choice, I’d lean toward the human aspect.

Of course, I’d only do this with good hitters.  If Adam Everett feels more comfortable as a #2 hitter, I’d still put him at the bottom of the order.

If they don’t get B Roberts, I’d leave Sori batting first, because they don’t have a good leadoff man anyway. I certainly would’t want to give Theriot or Pie 730 PAs. But If they get Roberts--a really nice classic leadoff type-- I make a change.

David, if you can post your Cubs forecasts exactly following this template (see post 1):
http://www.insidethebook.com/ee/index.php/site/comments/optimizing_the_batting_order/

I can try out a few examples tomorrow, presuming Soriano doesn’t change his talent level based on men on base or not.

I’ll do that.  I’ll take some time to think it over and post it a little bit later on today.  Thanks a lot.

Things that we don’t know how much is skill or luck, or we are not sure exactly, can and should always be used as tiebreakers.

Some things have so little gain (such as where you bat players in a reasonable order) that things like where a batter is more comfortable is a better requisite.  For example, batting your pitcher 8th gains 2-3 runs a year, or so.  If there is any chance that the hitter you bat 9th is going to feel slighted, don’t do it!

Tango, I am not sure “coming across Soriano’s splits by chance” necessarily is a whole lot different than cherry picking it/him.  It depends on a lot of things.  For one thing, how many splits do you generally look at?

But it is not worth arguing.  The main reason I brought it up is to set the record straight on the conventional argument (Bill James even uses it all the time), “But look at so-and-so.  His splits are off the chart!  Surely THAT must mean something.”

I ran the numbers once, and if I recall correctly batting Soriano lower in the order is worth something like seven or eight runs a year - I used the lineup simulator at Baseball Musings. That’s presuming that you don’t do something stupid and bat someone like Ryan Theriot at the top of the order, which pretty much wipes out the gains anyways.

Since we don’t know for sure that those splits are based on luck and not something Soriano - or the pitchers - are doing differently, I don’t think it’s worthwhile to move him in the lineup.

I have not seen many men on base splits.  I’ve probably looked at maybe 10, ever, like Abreu, Ichiro, Bonds, Pujols.  I’m sure a few others in the past, like Hawk, Raines, Rice, Brett, Murray.  Not many.  Not enough to get a handle on a distribution.

I think it makes a big difference if you pick a player at random, or pick a player from the extreme.  When you pick a player at the extreme, you preclude the possibility that someone can be more extreme.  Choosing someone at random means that someone could still be more extreme.

***

The Book, p.116:

Pitchers show inherent abilities to perform slightly differently with the bases empty than when throwing from the stretch. Any similar variation in batter performance is small, if it exists at all.

So, Andy’s looked at it, and while he sees the men on base changing enough for pitchers, he doesn’t really see it as much for hitters, though there is some likelihood that it exists.

tangotiger--

Here are my forecasts for the 8 Cubs in the starting lineup.  Let me know if you need any other information.  And thanks for checking this out.  I’m interested to see the results.

Soriano,R,.285,.335,.535,30,yes,good,not much,alot
Fukudome,L,.290,.365,.450,15,normal,good,not much,alot
Lee,R,.305,.400,.510,25,no,normal,alot,normal, alot
Ramirez,R,.300,.360,.550,30,no,slow,normal,normal
DeRosa,R,.280,.355,.425,12,no,slow,alot,alot
Soto,R,.275,.340,.430,.15,no,slow,alot,alot
Pie,L,.270,.320,.420,12,yes,good,not much,alot
Theriot,R,.275,.330,.370,3,yes,good,normal,not much

About the all relief rotation, a poster at Batter’s Box kept suggesting that you take two young starting pitchers and have one pitch the first five innings. Then, the second starter would essentially “start” another game at the sixth inning (unless they were required to come into the game earlier because of a blow out). Thoughts on this? I believe the pitchers suggested were Marcum or Listch and Janssen (IIRC).

The Book suggests a similar approach, since there’s a noticeable dropoff in performance each time through the order.

David,

Just doing it real quick, against LHP, without looking at all the things (speed, handedness, etc).  Just the player’s hitting profile:

1. DeRosa
2. Lee
3. Soriano
4. Ramirez
5. Soto
6. Fukodome
7. Theriot
8. Pitcher
9. Pie

Putting Soriano as leadoff hitter will cost you 2 runs.  But, if he truly is better with bases empty and/or feels much more comfortable there, you can leave him there.

The big standouts is Ramirez is best suited at cleanup, and Lee in the top 2 spots.

Against RHP
1-4. same
5. Fukodome
6. Soto
7. Pie
8. Pitcher
9. Theriot

I’ll look at the DP, baserunning tomorrow, to see if that plays a role.

Btw, what’s the lineup being bandied about in Chicago?  And any thoughts as to differences with LHP and RHP?

Thanks a lot.  I’ve always actually thought DeRosa would be a good player to have in the leadoff spot.

The lineup we keep hearing is this:

Soriano
Fukudome
Lee
Ramirez
DeRosa
Soto
Theriot
Pie

The Cubs still want to get a right handed hitting outfielder to play some CF against tough lefties and maybe a strict platoon as well though I think they’ll give Pie a chance against lefties (some of them).  With their interest in Brian Roberts this long, one has to think a deal will get done and Jay Payton will probably be that 4th outfielder.

As of right now, the Cubs don’t have anyone else that can play CF so there wouldn’t be any platooning going on. 

Interesting Fukudome is 5th, but then again, I’m not as optimistic about him as a lot of people are or the projection systems are.  I see a player who will be a good one, but far from elite and probably closer to average than above average. 

If you have time, what would the best lineup be using Marcels? 

So 2 runs obviously isn’t anything for people to get that worked up about especially when you consider his wide splits and there just might be a reason for it.  I was surprised to read Sheehan’s article when BP often talks about how lineup construction matters so little.

Marcels won’t work really for Pie or Fukodome.  I doubt your forecast on the other players would make much difference. 

In any case, I prefer that others tell me the forecasts, as I only want to be responsible for arranging the pieces.  There’s two parts, and the critical part is the forecast.  It’s important that the reader realizes that he’s the one who’s driving this, not me.

I think it’s a very safe bet that the Opening Day lineup will be:  Soriano, Fukudome, Lee, Ramirez, DeRosa, Soto, Theriot, Pie

I get a 3-run optimization with my lineup against that one, against LHP and RHP.

Certainly, not that big a deal, and I’d defer to managerial insights on it.

"That’s presuming that you don’t do something stupid and bat someone like Ryan Theriot at the top of the order, which pretty much wipes out the gains anyways.”

Oy vey - Lou’s now talking about doing just that. And then batting Soriano second. So the two players with the lowest projected OBP on the team would be the one-two punch to start the game.

Oy is right. The reason for this is that Lou wants to take the pressure off Soriano to try to steal, given his leg problems. But why not just tell him not to steal often and leave him at #1? Not that he’s an ideal leadoff man, but he’s used to it, he gets lots of PAs, and it keeps Theriot at the bottom of the order.

Reds mainstream media taking a sabermetric view of their batting order:
http://news.enquirer.com/apps/pbcs.dll/article?AID=/AB/20080309/SPT04/803090384/

I don’t think I would ever put a rookie as a cleanup hitter, certainly not if he’s 20 years old.  Junior, ARod, and Bonds didn’t as rookies.  I don’t see why you would.  I’d consider it at the half-way point, but not earlier.

Reds sure have alot of LHH.