Lineup help

[alorrigan]

I need some help with the lineup for the team I help coach. Our offense has really struggled all year and I want to really shake up the lineup for our next game on Saturday. For the next game I'm thinking of setting the lineup completely SABER-based. I only have very basic stats (singles, doubles, walks, hbp, etc.) nothing too deep. Is there anywhere I can find info on what numbers should be put where in the lineup or even what is your own personal opinion on what to bat where.

 

EDIT: Sorry just saw the other post about a similar subject.

[1992casey]

It looks like you already have enough to compute slugging percentages. If you have sac flies, you'd have everything you need for OBP. If you lack those, maybe you can just compute OBP without them.

 

Someone else would have to fill you in on the construction of a SABER-based lineup (which spots emphasize OBP, slugging or total OPS). rluzinski can probably also fill you in on some new stuff that came out in a book this year. You might want to try one of those lineups, too.

[rluzinski]

Don't worry about sac flies if you don't have them, just figure OBP without them.

 

First, remember that batting order isn't nearly important as some might think. The difference between the best and worst lineup isn't all that much. Second, since run environment dictates lineup optimization significantly, I'll just give my advice in generalities. This is just my opinion but it's based on some decent studies on the topic.

 

Put good OBP guys 1 and 2. If they have good speed as well, good for them but it's not too important. That's probably the single largest flaw of many lineups in the majors; putting some guy who's fast and/or "good with the bat" at #2. Screw that. You want a guy who will get on base.

 

At 3, put a guy who has a lot of power but my not get on as much as your #4 guy. Your #4 guy should be the stud, with good power and OBP skills. The next best power hitter should go #5. Order the rest in terms of their overall ability of getting on and power.

[1992casey]

By the way, here's the formula for OBP:

 

(H+BB+HBP) / (AB+BB+HBP+Sacrifice Flies)

 

And here's the formula for slugging pct.:

 

Hits + Doubles + (2*Triples) + (3*Homers)/AB

 

or you could do:

 

Singles + (2*Doubles) + (3*Triples) + (4*Homers)/AB

 

The second formula is better at showing how slg. pct. actually works. The first formula is a better fit for how most stat lines are presented (it saves a step or so). Chose the one you like, though; you'll get the same thing in the end.

[Wally Whitetail]

The difference between the best and worst lineup isn't all that much.

 

How can you say that?

 

Yeah, if you put straight up statistics into a simulator that would work. These are human beings not robots. Some players do not hit well in certain slots in the lineup and hitters are handled differently depending on where they are hitting. A guy like Ryan Howard batting in the 8 hole isn't going to put up the power numbers he did in the heart of the lineup because he will be pitched around much more often. Also, that person getting almost 100 less atbats per year is going to kill your offensive production. Not to mention that the people getting the most at bats in your lineup won't even be the guys who have the best obp, that will have a HUGE effect on how many runs you score in a season.

 

Here is a good resource setting your batting order:

 

Setting a Batting Order

[rluzinski]

Quote:

---

Yeah, if you put straight up statistics into a simulator that would work. These are human beings not robots. Some players do not hit well in certain slots in the lineup and hitters are handled differently depending on where they are hitting.

---

 

Many studies have done on this topic and there just isn't very much statistical evidence to support that claim. I don't mean to suggest that there's no interaction among batters but it's most certainly significantly less than traditionalists assume. If you are unswayed by statistical evidence than you are in the wrong part of this forum.

 

Also, that person getting almost 100 less atbats per year is going to kill your offensive production. Not to mention that the people getting the most at bats in your lineup won't even be the guys who have the best obp, that will have a HUGE effect on how many runs you score in a season.

 

Again, you are over estimating the effects. Lineup optimization has been studied to death and many go through the step of making the worst lineup (pitcher batting first) to establish a scale. The author always seems suprised by the relatively small effect it has on runs scored. Batting lineup optimization is important, but it appears to be not as important as some think.

 

LINK

[Wally Whitetail]

If you are unswayed by statistical evidence than you are in the wrong part of this forum.

 

I didn't know that I have to be a stat geek to have anything meaningful to say in this forum I guess. I would have thought you would want views from both sides of the coin.

 

My question is how can you do lineup optimization without human interaction? And since there isn't human interaction how can you meansure it's importance? You can't just automatically say there isn't much coorelation because there isn't anything out there that says otherwise. Isn't it pretty much just a guess that it doesn't?

 

I guess I just don't see how giving Chad Moeller and the pitcher's spot 100 more at bats per season and taking 100 at bats away from Carlos Lee and Prince Fielder isn't going to have much of an effect.

 

For example if this example leads to 50 less runs (which wouldn't be significant persay) that could mean the difference in wins in potentially every 1 run game, most 2 run games and even some 3+ run games we won.

 

You could also say that those 50 runs would be distributed even in games where we won by 5+ or so...but how can you determine when those runs would have disappeared and what effect that would have on the rest of the game and that strategy of the team or opposing team at those points? You can't.

 

So if you ask me these kinds of studies are pointless and mean absolutely nothing. It's just someone playing devil's advocate because they have numbers to play with.

[rluzinski]

I would have thought you would want views from both sides of the coin.

 

I did not mean to imply that you had nothing to add to this discussion but if you disregard statistical evidence out of hand, you can see how that might create problems when the title of the forum is "Statistical Analysis".

 

No one ever assumed there was no interaction among batters, quite the opposite. I think many researchers were suprised when they couldn't find any real statistical evidence of protection and other related phenomena. They realized that a batter's performance is relatively consistent, irregardless of his place in the batting order and the players around him. When that realiziation was made, it justified using simulators in order optimization studies. Again, it's not my claim that their are no such effects; only that they are most likely small enough as to be masked by statistical noise.

 

Also, I didn't mean to suggest that the difference between the worst and best lineup was insignificant just suprisingly small. Since the tradational batting lineup (like what you posted) is a fairly reasonable way to contruct a lineup in the first place, further optimization results in only a modest gain in runs scored. It is because of that fact that it's usually not worth all the discussion teh topic usually generates.

[Wally Whitetail]

I think many researchers were suprised when they couldn't find any real statistical evidence of protection and other related phenomena.

 

There is just as much evidence that they do have an effect on the amount of runs scored as there is that they don't have an effect. And that amount of evidence is zero evidence which is exactly why these studies don't hold any weight. They are making guesses and their logical guesses are just as good as mine that say they have a big effect on a particular batter's production. They are trying to prove something that can't be measured and since it can't be measured they just assume that it has little effect. I don't know how they can come to that conclusion, in fact, it's impossible to come to that conclusion without any hard evidence. It's a theory, not fact. I don't agree with that theory and you mentioned to the gentleman who started this thread like it is fact when it isn't.

[rluzinski]

And that amount of evidence is zero evidence which is exactly why these studies don't hold any weight.

 

The evidence is in the form of statistical studies using the scientific method. Again, if you consider them zero evidence than you are in the wrong forum. While their findings are in no way conclusive infallable, they are certainly stronger evidence than the anecdotal evidence usually used to argue the opposing view. You haven't even offered up that.

 

It's a theory, not fact.

 

I doubt you have never read a single word from any of these studies yet you disregard them with a wave of the hand. With a comment like the above, I can't say I'm suprised.

 

I don't agree with that theory and you mentioned to the gentleman who started this thread like it is fact when it isn't.

 

Sheesh. Go back and read the the post. My exact words were:

 

Quote:

---

This is just my opinion but it's based on some decent studies on the topic.

---

 

EDIT: Here's a link to one of the best lineup studies out there:

 

Evaluating Traditional Lineups

[Wally Whitetail]

You haven't even offered up that.

 

I've given plenty of reasons within this thread why I believe my theory is correct. Again, you believe your theory is "right" because there isn't any physical evidence to support mine. That doesn't make it right, it just makes it different. You can't agree to disagree without insulting?

 

I doubt you have never read a single word from any of these studies yet you disregard them with a wave of the hand. With a comment like the above, I can't say I'm suprised.

 

What gives you the right to talk that way to me?!? Because I don't agree with you you are now going to insult my intelligence? Get off your pedestal.

 

I read every word of the link you provided by the way. I just find many holes in the argument. I'm sorry that disturbs you so.

 

Again, if you consider them zero evidence than you are in the wrong forum.

 

Sounds like you want anyone that comes into this forum to roll over and bow to every word you say. Apparently everybody should take the things that go in this forum as gospel. Is this a place for ideas to not be challenged? I didn't think it was supposed to be.

[rluzinski]

I've given plenty of reasons within this thread why I believe my theory is correct.

 

Reasons but zero evidence. Not even anecdotal evidence.

 

You can't agree to disagree without insulting?... What gives you the right to talk that way to me?!?

 

You come to a statistical analysis forum and say:

 

Quote:

---

"So if you ask me these kinds of studies are pointless and mean absolutely nothing."

---

 

Now you want to play the martyr when someone strongly objects? Give me a break.

 

Because I don't agree with you you are now going to insult my intelligence?

 

It has nothing to do with your intelligence. It has to do with your disregarding studies you know nothing about yet think are worthless. You could be genius for all I know.

 

Sounds like you want anyone that comes into this forum to roll over and bow to every word you say.

 

Yes, I post on this forum not to exchange thoughts and ideas about statistical analysis of baseball but simply to boost my ego. Because all the cool kids like statistical analysis. You've got me all figured out. http://forum.brewerfan.net/images/smilies/eyes.gif

 

If you think statistical analysis of baseball has absolutely no merit, start a thread in this forum explaining why you think that. Please don't inject "studies mean nothing" into random threads because that adds nothing to the discussion.

[Wally Whitetail]

"So if you ask me these kinds of studies are pointless and mean absolutely nothing."

 

That is my opinion, just because you don't like it that gives you no right to insult me. It wasn't a malicious attack against you or even directed at you in any way. I offered up my opinion.

 

It has to do with your disregarding studies you know nothing about yet think are worthless.

 

Did you not read that I said that I read every word of the first link you provided and now I've read every word of the second "study" you provided. They don't measure human interaction in any way or even attempt to. That leaves a HUGE hole in the study doesn't it? And since they can't measure it they disregard its importance. Why is it so hard to believe that someone would find little value in a study that does that? Doing a study on OPS or VORP, etc. is one thing, combining statistics (which many already have flaws in them in the first place) in a complex interaction such as a batting order to find the "optimal" batting lineup doesn't hold any weight and isn't a good study in my opinion.

 

If you think statistical analysis of baseball has absolutely no merit, start a thread in this forum explaining why you think that.

 

Hold on here. I would like you to quote me as to when exactly I said statistical analysis of baseball has absolultey no merit. You're assuming things out of thin air to make an argument. I said THESE studies are leaving out a huge element. You can't measure the optimal lineup the way they are trying to. Sorry, it's not possible.

 

I find GREAT value in many statistical studies...not this one.

 

You seem frustrated that I don't see it your way and are lowering yoruself to belittling me and my knowledge of statistical studies although having no idea whether I do or not. You might want to think twice when assuming things about people.

[rluzinski]

You seem frustrated that I don't see it your way and are lowering yoruself to belittling me and my knowledge of statistical studies although having no idea whether I do or not.

 

Point me to the evidence (anecdotal or statistical evidence) you used to formulate your opinion that batter interaction (protection, skill set change based on batting position, etc...) is a real and significant phenomena. If you can demonstrate that your opinion is at least based on some kind of hard evidence, I will be happy to retract my claim. As it is, you have still provided zero evidence to support your opinion.

 

That is my opinion, just because you don't like it that gives you no right to insult me. It wasn't a malicious attack against you or even directed at you in any way.

 

Where did I maliciously insult you again? By suggesting you are unfamiliar with the statistical studies pertaining to this subject matter? That fact is obvious and if you don't want that pointed out, don't dismiss studies you have never read.

[Brian the Automator]

Apparently, bjames just can't stay away from you, russ. Wally Whitetail's been banned.

[rluzinski]

I really should have figured that out.

[brewers19]

Man, just when it was getting entertaining.

[Brian the Automator]

[ Man, just when it was getting entertaining. ]

 

Bjames is kinda like Wile E Coyote.

[SoCalBrewfan]

Wow, I missed all the fireworks here.

 

Getting to a point of substance here, though, wally wondered whether a given player might respond differently batting first or fifth in the order. (OK, actually, he asserted that they would and then went non-linear...but that's not what I want to address.)

 

To my knowledge (admittedly very incomplete and out-of-date), that question has not been addressed...rather most studies of lineups have plugged in stat lines like strat-o-matic cards and then run simulations. The link russ posted has the following statement:

 

Quote:

---

As for the assumptions, the first is that we are assuming batters don't care where they hit in the lineup, that if we took (for example) Barry Bonds and hit him leadoff, he would hit as well as he did hitting cleanup, and wouldn't go into a sulk and have his performance suffer as a result. Baseball players are not simply numbers in a transition matrix and a theoretically great batting order is not going to work if it causes a player revolt. This is probably more of a problem when proposing novel relief pitching strategies, but it's also a potential problem here.

---

 

They go on to add possible 'protection' as a factor they are ignoring. Now that's reasonable...any study has to make assumptions, and seek to answer limited questions lest it be completely untenable. But it makes me wonder...does anyone know whether this question has been studied systematically? Even if it's not a matter of sulking, but rather a player feeling pressure to take a different approach when hitting leadoff as opposed to cleanup, that seems like it might well be an effect...perhaps Weeks would put up a different statistical line batting 1st vs. 2nd vs. 6th. Any references to studies of this nature?

[rluzinski]

I've read sveral studies on protection and I recall them all concluded the only thing you might lose is intentional walks. They could not find Bonds' performances being conttroled by who batted after him, for instance. All the old links seem to be dead but I shall continue the search.

 

As for players feeling compelled to change their approach, I suppose they will if they are actively being asked to. Corey Patterson comes to mind. Since the point of these studies is to see how various skill sets can be utilized in various places in the lineup, I think it's fair to assume management could make it clear to the player NOT to change his approach to fit the traditional roles.

 

Finally, I'm sure there are some interaction among batters but what's important is the scale of the interaction. It appears that the connections between consecutive batters is slight enough as to allow simulations of batting orders to simply work. Maybe a study will come along to prove all the above wrong but as far as I know, this seems to be where we are at right now.