Lineup formulas/tools

[Dontworry]

A writer over at hardballtimes.com recommended this tool to create your team's ideal batting order -- http://home.comcast.net/~stein.stephen/cardcomp/lineup.html I couldn't actually get the tool to work myself (anyone?), but the assumptions on which it's based seem reasonable to me. That is,

1. Put the best OPS in 3rd
2. Put the best remaining Slg in 4th
3. Put the best remaining OBP's in 1st and 2nd (with the better Slg in 2nd)
4. Arrange the remaining players in order of descending Slg

What do you think? Sounds pretty logical, right? So I looked back to last year's stats and applied the methodology to my own team. Not exactly the results I was expecting.

Try it with your own team and see what you get.

[RuthMayBond]

A writer over at hardballtimes.com recommended this tool to create your team's ideal batting order -- http://home.comcast.net/~stein.stephen/cardcomp/lineup.html I couldn't actually get the tool to work myself (anyone?), but the assumptions on which it's based seem reasonable to me. That is,

1. Put the best OPS in 3rdWhat if it's a guy like Billy Hamilton or John McGraw who gets most of his OPS from the "O"?

<3. Put the best remaining OBP's in 1st and 2nd (with the better Slg in 2nd)>

What if one guy has a barely better SLG but WAY better OBP, wouldn't you want him in 1st?

<4. Arrange the remaining players in order of descending Slg>

But shouldn't you consider OBP?

[digglahhh]

You do get some werid results here's what the best prospective line-up for '06 would look like:

1. Lo Duca
2. Floyd
3. Delgado
4. Wright
5. Diaz/Nady (either fits the formula)
6 Beltran
7. Reyes
8. Matsui

Of course part of this is because of Beltran's poor season last year. If you used his career numbers he'd be plugged in the 2 slot, I think, depending on the exact specfications of SLG vs. OBP ( he'd have the highest remaining in both categories).

One potential problem with this formula is it doesn't take into account "protection," assuming it really exists. Who gets walks is sometimes determined by where they hit. There's no guarantee the walk rate for certain guys replicates when you move the order around.

There are other tangential questions as well, for example Reyes' OBP is terrible, but what to make of the fact that he's a virtual lock to go first to third on a single? Does 60 SBs mitigate some of his poor SLG? A single and stolen base on the next pitch isn't really all that different from a double. There are ton of things to consider here, a simple formula probably doesn't do the process justice- especially when you have guys with unique talents and deficiencies like Reyes.

[BoSox Rule]

Best 2006 Red Sox lineup

1. Kevin Youkilis
2. Jason Varitek
3. David Ortiz
4. Manny Ramirez
5. Coco Crisp
6. Trot Nixon
7. Alex Gonzalez
8. Mike Lowell
9. Mark Loretta

[Dontworry]

I think it's pretty logical, but like a purely statistical approach it is too superficial. It doesn't take into account, example, how good of a situational hitter your 2man is, or how good your cleanup guy is w/ runners on, or even the simple fact that some guys are not very good at certain order slots. Overall though, it isn't a horrible way to go, its better than just looking at BAs and RBIs and going from there.

[It's Over The Wall!]

Crazy Order:
1. Bob Abreu
2. Chipper Jones
3. Derrek Lee
4. Ryan Howard
5. Matt Murton
6. Matsui
7. Carlos Lee
8. Kotchmen
9. Morneau

This is crazy! Abreu would not hit 1 and murton would not hit 5
and moneau would not hit 9

[Ubiquitous]

The Book just shipped this week. It has a section on lineups and what it found was a little different.
Your three best hitters should bat somewhere in the #1, #2 and #4 slots. Your fourth- and fifth-best hitters should occupy the #3 and #5 slots. The #1 and #2 slots will have players with more walks than those in the #4 and #5 slots. From slot #6 through #9, put the players in descending order of quality

And your pitcher if he is not a good hitting pitcher should bat 8th, not 9th. The increase in plate appearances is not offset by the gain of having a better hitter before your best hitters.

[Gino Guacamoolie]

A writer over at hardballtimes.com recommended this tool to create your team's ideal batting order -- http://home.comcast.net/~stein.stephen/cardcomp/lineup.html I couldn't actually get the tool to work myself (anyone?), but the assumptions on which it's based seem reasonable to me. That is,

1. Put the best OPS in 3rd
2. Put the best remaining Slg in 4th
3. Put the best remaining OBP's in 1st and 2nd (with the better Slg in 2nd)
4. Arrange the remaining players in order of descending Slg

What do you think? Sounds pretty logical, right? So I looked back to last year's stats and applied the methodology to my own team. Not exactly the results I was expecting.

Try it with your own team and see what you get.


It's a pretty logical approach, but it doesn't take handedness into account, nor which way a player bats.

[Steve_Stein]

a "toy". A way to answer all the questions that arise on simulation game bulletin boards - "what lineup should I use with these guys"?

In real life, or even with a better stat set (L/R splits and stuff) you can certainly do better. The toy is just a quick and dirty way to get a good starting point. Then you can swap guys around here and there and get a reasonable lineup.

I'm intrigued by what I'm hearing about the method in "The Book". I had a preferred lineup analysis method about 15 years ago (http://home.comcast.net/~stein.stephen/lineup.html) which did actual analysis instead of using simplistic rules. But some of the things "The Book" says are similar to what my older method said - the best hitter is often the 2 hitter, not the 3. I also found that good OBP is very useful in the 9 slot, which means optimal NL lineups should usually not have the pitcher bat last. Does The Book say anything about that?

[Steve_Stein]

In real life, or even with a better stat set (L/R splits and stuff) it's a quick and dirty way to get a good starting point.

I meant to say: "In real life, or even with a better stat set (L/R splits and stuff) you can certainly do better. The toy is just a quick and dirty way to get a good starting point."

[Tango Tiger]

I also found that good OBP is very useful in the 9 slot, which means optimal NL lineups should usually not have the pitcher bat last. Does The Book say anything about that?

Yes, it does. And my research concurs with your assertion.

[Chisox]

Yes, it does. And my research concurs with your assertion.
When I played Video Games and was an NL team, I'd always bat my pitcher in the seventh spot. That way, I could have more speed at the bottom of the line-up so the lead-off batter could utilize his speed and advance better when the order turned over. Is there any statistical advantage to that? (BTW, I've never been one for the pitcher batting ninth, anyway, due to the above logic.)

[Tango Tiger]

#7 and #9 are equivalent in terms of where to put the pitcher. #8 is best, and #4 is worst. It's all laid out in the book.

Thanks...

[astromaker]

Astros "perfect" lineup
1) Brad Ausmus
2) Preston Wilson
3) Morgan Ensberg
4) Lance Berkman
5) Jason Lane
6) Craig Biggio
7) Adam Everett
8) Willy Tavarez
9) Pitcher

Makes you wonder why you don't see more catchers leading off.:D

[Tango Tiger]

At a minimum, you need to have a lineup against lefties and against righties. Preferably (and especially if you have a severe GB or FB hitter), you need to break those down against FB and GB pitchers.

If you repost your players with their batting hand, and with their expected overall BA, OBP, SLG, then I'll give you my take on it.

[redbuck]

Bill James did some analysis and concluded that lineup makes almost no difference to a team's performance.

I guess if you wanted to make one you would have the highest OBP guys the highest because they would get the most plate appearances.

and you probably want to space right handers and left handers

[Tango Tiger]

The difference in optimization amounts to 5 to 15 runs over a season. Some people think that's "no difference", and others think that's "alot". Whatever adjective you want to use is fine.

[digglahhh]

The difference in optimization amounts to 5 to 15 runs over a season. Some people think that's "no difference", and others think that's "alot". Whatever adjective you want to use is fine.

Right, whether or not those 5 or 15 runs are particularly meaningful is determined by luck.

What most people consider "luck" is really random statistical variation. The real "luck" is if that statistical variation happens to occur at the intervals when it helps you the most.

That's my take.

[Chisox]

#7 and #9 are equivalent in terms of where to put the pitcher. #8 is best, and #4 is worst. It's all laid out in the book.

Thanks...

Thank you.

[Reed Johnson]

2006 Blue Jays Lineup:

1. L. Overbay
2. F. Catalanotto
3. T. Glaus
4. V. Wells
5. S. Hillenbrand
6. B. Molina
7. A. Rios
8. A. Hill
9. R. Adams

[tybor]

Can anyone help me find a simulator that will allow me to plug in different lineups for my co-ed softball team to see which generates the most runs? I found Steve Stein's but it looks like you can only plug in MLB players.

Here is my question - how would you order this 12-person lineup in order to maximize runs scored?

--Four big hitters with OBP of .700 who homer in about 1/4 of their at-bats and double in another 1/4 of their at-bats (SLG ~ 1.300)
--Four contact hitters with OBP of .500 who hit mostly singles and doubles (SLG ~ .800)
--Four women with OBP of about .300 who hit almost exclusively singles

We are having a debate about which is the better lineup philosophy:
A. Bat in an approximately best-to-worst progression.
B. Spread out the power guys evenly throughout the lineup (ie, don’t bunch them).

From what I have read, it seems that the former is better. I would like some real data in order to convince the other guys on the team.

Thoughts?

[Ubiquitous]

This site lets you plug in (http://www.baseballmusings.com/cgi-bin/LineupAnalysis.py?Player0=Alfonso+Soriano&OBA0=+0.351&Slug0=+0.560&Player1=%232&OBA1=+0.362&Slug1=+0.381&Player2=%233&OBA2=+0.346&Slug2=+0.443&Player3=%234&OBA3=+0.390&Slug3=+0.467&Player4=%235&OBA4=+0.339&Slug4=+0.450&Player5=%236&OBA5=+0.339&Slug5=+0.429&Player6=%237&OBA6=+0.328&Slug6=+0.432&Player7=%238&OBA7=+0.318&Slug7=+0.326&Player8=%239&OBA8=+0.243&Slug8=+0.269&Model=1) whatever numbers you like.

[Wade8813]

I don't understand how having your pitcher be the #8 hitter be a good hitter helps particularly. If you want someone to get on in front of your leadoff hitter, why not just have that guy bat leadoff, and have the leadoff hitter bat 2nd?

[tybor]

That Baseball Musings site is good. Is there one that will let me play with 12 or 13 or 14 people lineups?

Thanks

[baseballPAP]

With 12 years experience in co-ed leagues, the best teams almost always ended up with lineups that went something like a normal baseball teams top 4 hitters, repeated. Meaning you have a leadoff type, then a singles hitter, then a power/speed combo, then pure power, and repeat. Your best four form the first bunch, next best the second, then the bottom portion is everyone else.

[Tango Tiger]

When I played in my company softball league way back in my early 20s, the very first thing I did with my baseball sim is what Tybor is suggesting: what happens when you have a severely skewed team?

I don't really remember the results, other than you did want your best hitters in the top half. Whether you wanted to have one or two bad hitters in between, I don't remember.

[Bukanier]

I don't understand how having your pitcher be the #8 hitter be a good hitter helps particularly. If you want someone to get on in front of your leadoff hitter, why not just have that guy bat leadoff, and have the leadoff hitter bat 2nd?
That #9 hitter is not supposed to be one of your best hitters, just better than the pitcher. Batting him at the top would be wasting plate appearances for the best hitters.

Similarly, maybe Lou Piniella refers to the book when he bats Koyie Hill 8th and Carlos Zambrano 9th :think:

[Wade8813]

That #9 hitter is not supposed to be one of your best hitters, just better than the pitcher. Batting him at the top would be wasting plate appearances for the best hitters. Alright, I see.

But it still seems just as likely that your worst hitter will kill a rally batting 8th as your slightly better hitter is to start a rally batting 9th.

[BoSox Rule]

Tango I saw you earlier mention that you would do vs. RH and vs. LH. IMO I would go a 3 year split vs RH but would it be something like 4 to 6 for LH since you face them a lot less often?

[Tango Tiger]

Alright, I see.

But it still seems just as likely that your worst hitter will kill a rally batting 8th as your slightly better hitter is to start a rally batting 9th.

This is true, depending on how close the talent level between the 8th and 9th hitter is. That is, in the AL, keep your worst hitter batting last.

Insofar as the NL is concerned, you are wrong. The pitcher in the 9th slot stop more rallies than he kills batting in the 8th slot. The details are in The Book.

[Tango Tiger]

Tango I saw you earlier mention that you would do vs. RH and vs. LH. IMO I would go a 3 year split vs RH but would it be something like 4 to 6 for LH since you face them a lot less often?

You need to regress your sample data because it is... a sample. All data is a sample, and therefore requires you to infer the true based on the sample.

And yes, the smaller the sample, the greater the uncertainty. And, your performance against RP will also inform, to a certain extent, the true rate against LP. I hate to sound like a commercial, but Andy laid it all out in The Book.

[sharrock]

I actually bought The Book specifically for the lineup chapter. I had read Tango's posting online elsewhere and figured I'd check it out. I'm fascinated by lineup constructions and the interactions of the various spots.

One thing I'd note about some of the lineups posted, especially the Red Sox lineup with Tek in the 2 spot... is that GIDP are killers, especially on a slower, high OBP team like the Red Sox. Tek should never hit in front of the best hitters, especially following an OBP machine like Youk.

The Book is well worth checking out for this chapter, as well as others like the section on bullpen.

The most frustrating thing about lineups, to me, is speed. So many people want speed at the top but hate to risk stolen bases or agressive efforts to get from first to third. So what is the point? In the AL, speed is rarely used enough to justify some of the lackluster hitters in the one and two slots.

Regarding pitchers batting 8th, Larussa tried this a few years back and was mocked. He isnt my favorite manager but I give him credit for trying it in the real world context.

[jibjabmraz]

The Book just shipped this week. It has a section on lineups and what it found was a little different.


And your pitcher if he is not a good hitting pitcher should bat 8th, not 9th. The increase in plate appearances is not offset by the gain of having a better hitter before your best hitters.

I've always wondered about this. However, what about the fact that you can (and will) pinch hit for your pitcher more readily than for whoever is typically batting 8th (likely an important defensive player)?

Assuming your best pinch hitter is better than your usual 8 hitter then you would be getting that PH in the lineup in front of your big hitters instead of in front of your worst starter. Also, it would be important to consider that the pinch hitter would, on average, be coming into much higher leverage situations.

I'm not saying it's enough to swing the balance; just wondering if anyone had tried to account for this.

[west coast orange and black]

what can skipper larussa and his cardinals expect by batting his pitcher 8th and a position player 9th to give 3-hole guy pujols more rbi chances?

larussa did that exact thing in 1998 with mcgwire in the 3-hole.