Batting Order: Does It Matter?

[GAC]

I was talking with a fellow RZer this morning, and we got on the topic of lineup construction/batting order. And we start talking mainly about the sabermetrics position on this subject, and the conversation also included "The Book: Playing The Percentages in Baseball" by Tango.

I've heard this book referenced at times on here in various discussions; but personally have not read it. I'm correcting that and ordering a copy today.

But I'm going to post an insert from the book on batting order below.

And correct me if I am wrong; but it seems the sabermetrics position is that batting order, because it is a continuous loop, does not matter. You can't hide a batter.

So you can bat them wherever? It just doesn't matter?

Now if most who are into statistical analysis - and I'm asking this from an educational position, not confrontational - agree with this premise that batting order does not matter, then why, over these last many years on this forum, do we continuously hear many lament over where various players are batting, and that guys with a high OB% should be at the top of the order if lineups are a continuous loop, each player gets his turn, and you're never going to have, for the most part, the same guy leading off every inning?

Isn't that then a contradiction? Doesn't even Bill James state that batting order does not matter, the number of runs a team scored would be roughly the same over a course of the season.

So what about the thinking that involes the idea of alternating lefties and righties, batters offering "protection", speed at the top of the lineup with bat control in the number two spot, etc., etc., etc.?

One thing we do know is that the higher a player hits in the order, the more times they will come to the plate

Batting #1

750 PA

Batting #2

725 PA

Batting #3

710 PA

Batting #4

700 PA

Batting #5

680 PA

Batting #6

660 PA

Batting #7

640 PA

Batting #8

625 PA

Batting #9

580 PA

Chapter 5: Batting (Dis)Order


One of the running themes throughout this book is context. To understand the impact of your possible choices, you have to understand the environment in which you are working. Context. Whenever you are trying to figure out what to do, take a step back, and ask yourself, “What's the context?” Context, context, context. We can't repeat it enough.
In this chapter, we'll turn our attention to the batting order and how to construct an optimal lineup. What does conventional wisdom say about this? Some managers or fans think you want a fast guy at the top of the order, without too much regard for how often he gets on base; others think that OBP is the most important attribute for a leadoff hitter, and that speed is secondary. You want the #2 hitter to move the runner over into scoring position, even if it means getting an out; therefore he should be a proficient bunter with excellent bat control. Your #3 batter should be a very good hitter, maybe your best—unless you want your best hitter batting cleanup. Then, maybe set up a lefty-righty situation down the order, trying to end with your worst batters at the bottom of the lineup. There, easy, end of chapter.
But why. Why? Why must the order follow such a pattern? Let's take a step back and understand the context of the batting order.
Loop
The most important context of the batting order is that it is a continuous loop. You get to set your batting order 1 through 9, and each player takes his turn. Once the inning completes, the next batter leads off the next inning. So the term leadoff batter is a bit of a misnomer. He's the leadoff batter of the game, but he won't often lead off an inning.
What if the rules were set so that the batting order is restored at the start of each inning. Tim Raines would lead off the first, second, third, and all the way through the ninth, innings. If those were the rules—if that were the context—you'd need a new strategy in place to determine the optimal lineup. One-quarter of the time, you'd end up having a 1–2–3 inning, leaving your cleanup batter on deck. In the current rules of baseball, the benefit of that huge cost is that he gets to bat in the following inning. In this fictitious league, it's the #1 hitter who will lead off the second inning. You will realize rather quickly that you can't have your best hitter in the cleanup spot if he will have 25% fewer PA than your top three hitters.
If we go back to the real world of baseball, we realize that you can't hide a batter. The best you can do is to defer him to the bottom of the batting order. But even batting ninth, he will eventually come up. How often will he come up?

[KittyDuran]

OK MISTER... how did you get GAC's password?

[oneupper]

For one, I think the order is VERY important. Not only do you want your better hitters getting more PAs (which is obvious), but you want to generate a favorable game "dynamic" (momentum...or "script") if you will.

I subscribe to the "hitting is contagious" theory (very hard to prove).
Basically, good hitters make their teammates better hitters.

A few reasons:
Tired pitchers are usually easier to hit.
Opposing managers will use their "B" bullpen guys when behind.
Three hits in a inning will usually net at least a run. Two...maybe not.


Give your best hitters lots of PAs...and bunch them together in the lineup rather than spreading them out.

[Rounding Third]

I have had that book on order for a couple weeks now and they ran out or something. My parents bought it for me for my birthday but it still hasn't arrived.

I think its more important who is in the lineup rathere than where. But I think where also has importance. Protecting a players and how many more at bats somebody gets being the main reasons why.

[GAC]

For one, I think the order is VERY important. Not only do you want your better hitters getting more PAs (which is obvious), but you want to generate a favorable game "dynamic" (momentum...or "script") if you will.

Give your best hitters lots of PAs...and bunch them together in the lineup rather than spreading them out.

This is what I have always subscribed to. And you mention a very key word IMO.... momentum (creating/maintaining it).

I've always been considered on here as "old school", though I consider myself a "hybrid" where I think there are viable and solid positions to be gleaned from both trains of thought.

I've always believed, regardless of the.... "a batting order is a continuous loop in which everyone is going to get a turn a bat regardless of where they are hitting".... you still want to group your solid hitters together, and not break them up, in order to create that momentum and give you that 1-2-3 punch. And I'm referring, within the confines of this discussion, of the 1-5 spots. Some may say 1-6, and that's fine.

For instance - if I'm breaking up my good hitters and inserting a "weak" hitter in there somewhere, then doesn't that give an advantage to that opposing pitcher who then sees he doesn't have to give so-and-so quality pitches because "Juan Castro" is on deck.

I understand those scenarios are still going to present themselves during a game depending on how an inning ends; but don't you try to minimize those situations and come at that opposing pitcher with a succession of your best hitters and not break them up?

It's not some newly discovered thinking, because even as one who is considered old school, I have always believed in the importance of OB%. Where I may differ with saber thought is the idea of what should BUOY (drive) that OB% - primarily walks. I could care less what is driving that player's OB%. Whether it's batting average and/or walks, as long as there is consistency in performance there. Guys like Pete Rose and Tony Gywnn had a career OB% of .375 and .388 respectively. Yet they were primarily BA-driven with a 162 game average when it came to BBs of 71 and 52 respectively.

And in those top 5 slots, it's not ONLY about OB% either IMO. One has to factor in SLG%.

[westofyou]

Deja Vu

http://www.redszone.com/forums/showthread.php?t=57325

[BCubb2003]

I think it matters to the pitcher, who is usually not a sabermetrician.

[Spring~Fields]

I was talking with a fellow RZer this morning, and we got on the topic of lineup construction/batting order. And we start talking mainly about the sabermetrics position on this subject, and the conversation also included "The Book: Playing The Percentages in Baseball" by Tango.

I've heard this book referenced at times on here in various discussions; but personally have not read it. I'm correcting that and ordering a copy today.

But I'm going to post an insert from the book on batting order below.

And correct me if I am wrong; but it seems the sabermetrics position is that batting order, because it is a continuous loop, does not matter. You can't hide a batter.

So you can bat them wherever? It just doesn't matter?

Now if most who are into statistical analysis - and I'm asking this from an educational position, not confrontational - agree with this premise that batting order does not matter, then why, over these last many years on this forum, do we continuously hear many lament over where various players are batting, and that guys with a high OB% should be at the top of the order if lineups are a continuous loop, each player gets his turn, and you're never going to have, for the most part, the same guy leading off every inning?

Isn't that then a contradiction? Doesn't even Bill James state that batting order does not matter, the number of runs a team scored would be roughly the same over a course of the season.

So what about the thinking that involes the idea of alternating lefties and righties, batters offering "protection", speed at the top of the lineup with bat control in the number two spot, etc., etc., etc.?

One thing we do know is that the higher a player hits in the order, the more times they will come to the plate

Batting #1

750 PA

Batting #2

725 PA

Batting #3

710 PA

Batting #4

700 PA

Batting #5

680 PA

Batting #6

660 PA

Batting #7

640 PA

Batting #8

625 PA

Batting #9

580 PA

Chapter 5: Batting (Dis)Order


One of the running themes throughout this book is context. To understand the impact of your possible choices, you have to understand the environment in which you are working. Context. Whenever you are trying to figure out what to do, take a step back, and ask yourself, “What's the context?” Context, context, context. We can't repeat it enough.
In this chapter, we'll turn our attention to the batting order and how to construct an optimal lineup. What does conventional wisdom say about this? Some managers or fans think you want a fast guy at the top of the order, without too much regard for how often he gets on base; others think that OBP is the most important attribute for a leadoff hitter, and that speed is secondary. You want the #2 hitter to move the runner over into scoring position, even if it means getting an out; therefore he should be a proficient bunter with excellent bat control. Your #3 batter should be a very good hitter, maybe your best—unless you want your best hitter batting cleanup. Then, maybe set up a lefty-righty situation down the order, trying to end with your worst batters at the bottom of the lineup. There, easy, end of chapter.
But why. Why? Why must the order follow such a pattern? Let's take a step back and understand the context of the batting order.
Loop
The most important context of the batting order is that it is a continuous loop. You get to set your batting order 1 through 9, and each player takes his turn. Once the inning completes, the next batter leads off the next inning. So the term leadoff batter is a bit of a misnomer. He's the leadoff batter of the game, but he won't often lead off an inning.
What if the rules were set so that the batting order is restored at the start of each inning. Tim Raines would lead off the first, second, third, and all the way through the ninth, innings. If those were the rules—if that were the context—you'd need a new strategy in place to determine the optimal lineup. One-quarter of the time, you'd end up having a 1–2–3 inning, leaving your cleanup batter on deck. In the current rules of baseball, the benefit of that huge cost is that he gets to bat in the following inning. In this fictitious league, it's the #1 hitter who will lead off the second inning. You will realize rather quickly that you can't have your best hitter in the cleanup spot if he will have 25% fewer PA than your top three hitters.
If we go back to the real world of baseball, we realize that you can't hide a batter. The best you can do is to defer him to the bottom of the batting order. But even batting ninth, he will eventually come up. How often will he come up?

I am not sure that I have the correct understanding to begin with, what it means, when one says that the lineup order implying that the lineup construction does not matter.

Anymore than I use to understand that strikeouts don’t matter.
Until I grasped that they are speaking statistically, and that in stats one strikeout is just one out in data points, just like any other out is just one data point. So what does it literally mean to a statistician that lineup order don’t matter? I don’t really know. I would have to ask a series of questions and to use examples such as below to understand if it is true or not that lineup orders don’t matter.

Do we have the right understanding when they say that the lineup doesn’t matter ?

The statement that lineup orders do no matter alone would imply that it doesn’t matter because it would not effect a teams chance to score runs and amount of runs scored if the teams worst batters were given the most plate appearances over a given season, which would not seem to be true, but is it true?

Does the statement mean for example that a players on base percentage doesn’t matter and that OBP is to be dismissed? Or that their slugging percentage doesn’t matter or that either or both OBP and SLG do not effect a given teams ability to score more or less runs if the players in a given lineup with the highest batting percentages were given less plate appearances over the season versus the worst batters in the lineup on a constant over a given season?

Can the managers make out the lineup by giving their poorest batters the most PA, plate appearances? would that increase or reduce the teams probability in scorning runs and effect the RS/RA differentiations in a positive or negative manner statistically?

If taken literally to mean that lineup orders do not matter, then it is true that it would not matter if the manager made his order out to be in the following below.

If constant, if the lineup below was the managers lineup every day for the entire season or that batters with comparable stats were placed in such an order it wouldn’t matter, wouldn’t matter meaning that the lineup order cannot effect the teams scoring chances positively or negatively. But can it ?

Cincinnati Team A plays Cincinnati Team B everyday for an entire season.

Left Handed pitcher starting today, and here’s your lineup, just using their season stats.

Cincinnati Team A
Bako C 		.314 .385 .699 	Groundball: 41.5% | Strikeout: 35.6% | GIDP: 1.7%
Griffey RF	.356 .408 .764 	Groundball: 39.4% | Strikeout: 18.9% | GIDP: 2.9%
Phillips 2B 	.330 .506 .836 	Groundball: 44.0% | Strikeout: 20.8% | GIDP: 4.3%
Cueto P         .174 .050 .224 
Dunn LF	        .409 .530 .940 	Groundball: 28.4% | Strikeout: 35.2% | GIDP: 1.2%
Encarncion	.318 .439 .757 	Groundball: 33.0% | Strikeout: 22.0% | GIDP: 2.6%
Votto 1B 	.358 .505 .864 	Groundball: 45.0% | Strikeout: 23.8% | GIDP: 2.0%
Bruce CF	.569 .756 1.325	Groundball: 55.6% | Strikeout: 14.8% | GIDP: 3.7%
Hairston SS	.370 .470 .840	Groundball: 34.9% | Strikeout: 22.9% | GIDP: 0.0%

Vs. 
Left handed pitcher starting today, and here’s your lineup, just using their season stats vs LH 
Cincinnati Team B
Phillips 2B 	.400 .726 1.126	Groundball: 44.0% | Strikeout: 20.8% | GIDP: 4.3%
Hairston SS	.465 .568 1.033 Groundball: 34.9% | Strikeout: 22.9% | GIDP: 0.0%
Dunn LF	        .431 .423 .854	Groundball: 28.4% | Strikeout: 35.2% | GIDP: 1.2%
Bruce CF	.500 .429 .929	Groundball: 55.6% | Strikeout: 14.8% | GIDP: 3.7%
Votto 1B	.405 .578 .983	Groundball: 45.0% | Strikeout: 23.8% | GIDP: 2.0%
Encarncion 3B	.366 .377 .743	Groundball: 33.0% | Strikeout: 22.0% | GIDP: 2.6%
Griffey RF	.282 .309 .591	Groundball: 39.4% | Strikeout: 18.9% | GIDP: 2.9%
Cueto P		.174 .050 .224
Bako C		.226 .373 .599	Groundball: 41.5% | Strikeout: 35.6% | GIDP: 1.7%

Would Team A score the same amount of runs as Team B today?
Would Team A score more runs than Team B today?
Would Team B score more runs than Team A today
Would Team A or Team B score more runs in a four game series?
On the season would Team A score the same amount of runs as Team B
On the season would Team A or Team B score more runs?
Which team wins more one or two runs games, Team A or Team B, or would it matter?

Or is the statement that lineups don’t matter isolated in the use of stats?

Is that similar to saying that strikeouts don’t matter, that they are just another out?
When statistically speaking it is true that a strikeout is just one out, just another out in statistics, it simply doesn’t matter how a player makes their outs, though they are saying it doesn’t matter how a player makes an out, they are not saying that it doesn‘t matter if a player makes outs though.

If one adds in that a given batter walks a greater number of times while striking out a lot, we can find instances where this player still might get on base many more times than another batter, reducing the meaning to a strikeout because the player still maintains a high on base percentage.

While another batter who doesn’t walk a lot and does strikeout a lot the meaning of a strikeout does mean something, but not in statistics, an out is still an out, one numeric data point, yet this batter can have a very poor on base percentage and the strikeouts would appear more significant.

Do strikeouts become more of a factor to the batter who doesn’t walk much versus a batter that does walk a lot? Suppose player A strikeouts 170 times and he walks 128 times = Adam Dunn 2002 OBP of .400 vs batting average of .249 where .400 would indicate that he gets on base many more times than .249 batting average would indicate. Suppose player B, Brandon Phillips who doesn’t walk much, and he strikeout 170 times, would a strikeout have more meaning? Not in stats, an out is still just one out. Seems to mean that it matters how many outs they make, not how they make the out.

[KronoRed]

Your better players should lead your team in appearances at the plate.

[Spring~Fields]

Your better players should lead your team in appearances at the plate.

Then batting order or lineup construction would matter.

[jojo]

I think there are a couple points that can sum the current saber view on batting order:

1. The optimal batting order does make a difference, it's just not as dramatic as one might think (somewhere on the order of 30 runs over the course of the season which translates into roughly 3 wins).

Here's a post summarizing what "The Book" says about the ideal lineup:
http://www.redszone.com/forums/showp...&postcount=224

2. Each drop in the order is associated with roughly 18 to 20 fewer PA's over the course of the season. In other words, batting Dunn 5th instead of 2nd would mean he'd get roughly 60 fewer PA's over a season.

3. Platoon splits are real. Batting four lefties in a row for instance could lead to LOOGY kryptonite late in games though the impact of this is debatable and obviously dependent upon the lefties.

I'm sure others will post anything I've forgotten to bring up or raise objections where appropriate.

[edabbs44]

Your better players should lead your team in appearances at the plate.

If your better hitters lead your team in plate appearances (top of the order), that means that they will probably get less ABs with people on base since your worst hitters will be hitting in front of them.

[RedsManRick]

Jojo, what are your thoughts when it comes to handedness and platoon splits. Does stringing together lefties to beat up on the right starter offset the cost of potential "LOOGY" kryptonite?

And to clarify, the 30 run difference you've cited is against a "reasonable" but not optimal lineup. That is, it's not versus a lineup that starts off: pitcher, catcher, banjo hitting SS...

[RedsManRick]

If your better hitters lead your team in plate appearances (top of the order), that means that they will probably get less ABs with people on base since your worst hitters will be hitting in front of them.

And if they're in the middle or at the bottom of the order, they still have the worst guys ahead of them, they just get fewer plate appearances. I don't follow your point.

[Spring~Fields]

Jojo, what are your thoughts when it comes to handedness and platoon splits. Does stringing together lefties to beat up on the right starter offset the cost of potential "LOOGY" kryptonite?

And to clarify, the 30 run difference you've cited is against a "reasonable" but not optimal lineup. That is, it's not versus a lineup that starts off: pitcher, catcher, banjo hitting SS...

So when they say, “lineups don’t matter”, they don’t mean to just toss out any old lineup arbitrarily as if it will have the same effect regardless, as if you have attempted to optimize your lineup to give your best batters in the lineup order the most chances to get the highest probability of getting on base and potentially scoring runs?
Especially with any given game having the potential to be a one or two run game?