PDA

View Full Version : Lineup Protection



WiffleWOOD
04-01-04, 01:04 PM
These articles are fairly long. I'm running to class right now, and don't have time to read them.

However, I thought it would be interesting to get a conversation going about the idea of "protection" in the line-up.

Here are two studies on the topic:

http://www.baseball1.com/bb-data/grabiner/protstudy.html

http://www.baseballprimer.com/articles/wright_2003-06-30_0.shtml


Let's get a dialogue going, perhaps between "statheads" and "old-schoolers," on the topic, in response to the articles or thoughts in general.

WiffleWOOD
04-01-04, 01:04 PM
and as always, please play nice :)

i'll be back in a bit to throw my own two cents in

Dave in MD
04-01-04, 01:12 PM
I don't believe in lineup protection. Good hitters hit in any lineup. I will read those articles.

AngelAstro
04-01-04, 02:18 PM
This is how think about protection.

I assume that the pitcher does not want to pitch to the "protector". Thus, the pitcher will try to pitch to the "protectee". This makes sense. I would much rather not walk JT Snow if he is batting in front of Barry Bonds, because then Bonds will definitely come up with someone on base. A situation no pitcher would want because Bonds would have the chance to knock in two runs rather than one run. So, in this case, the pitcher will try to not throw balls to Snow.

Now the pitcher is trying to throw more pitches to Snow near the strikezone. If the pitcher is successful at throwing more pitches near the strikezone (a big if), chances are Snow's OBP will decrease because the probability of Snow walking is low - because none of the pitches are balls.

However, I don't think that lowering OBP will raise batting average or slugging would necessarily change because Snow is still the same hitter that he always has been. Snow's inherent hitting ability isn't going to increase just because he can't walk anymore. For example, his power isn't going to increase. Why would Snow's ability to hit a homerun increase? It would not. I think we fail to realize that hitting a baseball, especially at the major league level is incredibly difficult. Even if Snow can now put the bat on the ball more often, because walking is not an option, this does not mean all these balls will be hit out of the park. Snow will still end up hitting grounders to the infield or pop ups to the outfield, all of which can and/or will be turned into outs.

So, if you ask me, the only thing a protector can possibly do for you is to give the protectee more strikes. What the protectee will do with those strikes will come down the ability of the protectee.

The data given in the reports linked above don't support my hypothesis. There are many players whose OBP increase while protected and players whose OBP decreases. In fact, I don't think the data given can support any hypothesis. There doesn't seem to be any consistant data. For any given player any stat seems to fluctuate given or not given protection. Thus it appears that protection really is an assumption rather than a rule.

nyybleachercreature
04-01-04, 02:29 PM
When a batter ahead of a big hitter comes up, I don't think the pitcher is saying to himself, "I better throw this guy strikes, or Im in toruble." But, I do think that a batter ahead of a big hitter is much more likely to see a fastball when there is in the count of 3-2 or 3-1. Now, I don't know how many pitchers would be comfortable throwing breaking stuff on 3-1, and Im guessing 3-2 counts on hitters ahead of a big hitter don't happen enough to pad stats all that much.
None of this applies when a batter is hitting in front of someone like Bonds over the last few seasons. I really do feel that Bonds makes pitchers think about what they are throwing to batters in front of him. I don't think any pitcher wants a man on 1st when Bonds comes up, nevermind 1st and 2nd or bases loaded. But a hitter that is feared as much as Bonds has been doesn't come around too often.

AngelAstro
04-01-04, 02:32 PM
Originally posted by nyybleachercreature
... But, I do think that a batter ahead of a big hitter is much more likely to see a fastball when there is in the count of 3-2 or 3-1. Now, I don't know how many pitchers would be comfortable throwing breaking stuff on 3-1, and Im guessing 3-2 counts on hitters ahead of a big hitter don't happen enough to pad stats all that much.

That's just it, maybe the protectee will get that grooved fastball on a 3-1 or 3-2 count. But to expect that grooved fastball to turn into a hit is folly. There is no reason that the batter would hit that ball any better in that situation than he would in any other situation. That's why to me, protection is irrelevant.

incarnadine
04-01-04, 03:46 PM
Originally posted by AngelAstro
That's just it, maybe the protectee will get that grooved fastball on a 3-1 or 3-2 count. But to expect that grooved fastball to turn into a hit is folly. There is no reason that the batter would hit that ball any better in that situation than he would in any other situation. That's why to me, protection is irrelevant.


One could make the case that if a certain pitch is expected, the batter has a better chance at putting it in play. To take it to an extreme, batters certainly have a better chance to get hits during BP, where everything is grooved.

Since not every pitcher can throw all his pitches for strikes, a protectee can mentally discount the probability that the ball he sees leaving the pitcher's hand is a curve (for instance). Therefore, he might be better able to predict it's position in the strike zone and make contact more often. This confidence could also lead to a harder swing, raising SLG.

Maybe I'm misunderstanding, but it would seem that removing some of the pitcher's "degrees of freedom" for deception would aid the batter, both in putting the ball in play and in power.

<i>(N.B. I'm not saying I believe that protection exists. However, that'd have to be decided empirically, since you can certainly make an</i> a priori <i>armchair explanation for why a protectee would be more likely to see certain pitches (those that the pitcher is most confident he can throw for strikes) and thus be able to hit them more often and with more authority.)</i>

AngelAstro
04-01-04, 03:54 PM
Originally posted by incarnadine



One could make the case that if a certain pitch is expected, the batter has a better chance at putting it in play. To take it to an extreme, batters certainly have a better chance to get hits during BP, where everything is grooved.

Since not every pitcher can throw all his pitches for strikes, a protectee can mentally discount the probability that the ball he sees leaving the pitcher's hand is a curve (for instance). Therefore, he might be better able to predict it's position in the strike zone and make contact more often. This confidence could also lead to a harder swing, raising SLG.

Maybe I'm misunderstanding, but it would seem that removing some of the pitcher's "degrees of freedom" for deception would aid the batter, both in putting the ball in play and in power.



This may be the case; however, the point I am trying to make is that even if the protectee can put the ball into play more often, that doesn't necessarily mean his AVG or SLG will increase because the defense will still be able to field the ball. True, there may be a slight spike in AVG or SLG, but I bet it would be negligible.

nyybleachercreature
04-01-04, 03:59 PM
Originally posted by AngelAstro


That's just it, maybe the protectee will get that grooved fastball on a 3-1 or 3-2 count. But to expect that grooved fastball to turn into a hit is folly. There is no reason that the batter would hit that ball any better in that situation than he would in any other situation. That's why to me, protection is irrelevant.
There is a better chance of a grooved fastball becoming a hit than a curveball, in my opinion. The more balls put into play, in general, would mean a higher BA.

AngelAstro
04-01-04, 04:06 PM
Originally posted by nyybleachercreature

There is a better chance of a grooved fastball becoming a hit than a curveball, in my opinion. The more balls put into play, in general, would mean a higher BA.

Not to be facetious, but why? Just because someone hits the ball harder doesn't necessarily mean it won't be caught. Perhaps that harder hit ball would be lined straight to the leftfielder. Any ball hit into play, except a homerun, can be turned into an out. If the protectee does not have a great ability to hit homeruns in the first place, I see no reason that putting the ball into play will result in a higher AVG or SLG.

incarnadine
04-01-04, 04:24 PM
Originally posted by AngelAstro
Not to be facetious, but why? Just because someone hits the ball harder doesn't necessarily mean it won't be caught. Perhaps that harder hit ball would be lined straight to the leftfielder. Any ball hit into play, except a homerun, can be turned into an out. If the protectee does not have a great ability to hit homeruns in the first place, I see no reason that putting the ball into play will result in a higher AVG or SLG.

Well, that's an assumption, that all balls hit into play are equal. A harder hit ball travels faster: grounders get into the hole and up the middle quicker (possibly) than infielders can get to them. A scalded liner might be tougher to get to than a soft one for outfielders.

Mike Cameron has a certain subset of the outfield he can get to in a given amount of time. If the ball gets to a given point in <i>x</i> seconds, it might be within his range. If it gets to the same point in <i>x</i> -- .01 seconds, it might not be catchable.

AngelAstro
04-01-04, 04:31 PM
Originally posted by incarnadine


Well, that's an assumption, that all balls hit into play are equal. A harder hit ball travels faster: grounders get into the hole and up the middle quicker (possibly) than infielders can get to them. A scalded liner might be tougher to get to than a soft one for outfielders.

Mike Cameron has a certain subset of the outfield he can get to in a given amount of time. If the ball gets to a given point in <i>x</i> seconds, it might be within his range. If it gets to the same point in <i>x</i> -- .01 seconds, it might not be catchable.

Absolutely true, but you are also assuming that everytime a hitter gets a grooved fastball that he is going to smoke it. I don't know if this is the case or not, but I rather doubt it. There is just too many things that have to be perfect for a batter to correctly hit any pitch that even if the batter knew with exact certainty the upcoming pitch, he could probably not pound the ball. For instance, if in batting practice every pitch can't be hit for a homerun, how can we know for certain that in game the batter will be able to pound any given pitch?

incarnadine
04-01-04, 04:42 PM
Originally posted by AngelAstro
Absolutely true, but you are also assuming that everytime a hitter gets a grooved fastball that he is going to smoke it.... how can we know for certain that in game the batter will be able to pound any given pitch?

It's not necessary for it to be a certainty, it's just a question of whether the pitcher's strategy of attempting to avoid the walk (by throwing a somewhat more predictable subset of his pitches which he has more confidence of throwing in the strike zone) make it <i>more likely</i> that a batter can hit and hit for power.

The pitcher's strategy always <i>ought</i> to be to pitch in a manner most likely to keep a guy off base. So he shouldn't change strategies just because Babe Ruth, Barry Bonds, or Christ Jesus is waiting on one knee in the on-deck circle. But he might anyways, to his detriment.

AngelAstro
04-01-04, 04:53 PM
Originally posted by incarnadine


It's not necessary for it to be a certainty, it's just a question of whether the pitcher's strategy of attempting to avoid the walk (by throwing a somewhat more predictable subset of his pitches which he has more confidence of throwing in the strike zone) make it <i>more likely</i> that a batter can hit and hit for power.

But this doesn't mean that you will see an increase in AVG or SLG.


]Originally posted by incarnadine


The pitcher's strategy always <i>ought</i> to be to pitch in a manner most likely to keep a guy off base. So he shouldn't change strategies just because Babe Ruth, Barry Bonds, or Christ Jesus is waiting on one knee in the on-deck circle. But he might anyways, to his detriment.

I agree with you here. So, in other words protection doesn't, or at least shouldn't, matter.

nyybleachercreature
04-01-04, 06:22 PM
I know some sabermetric people have said that any ball hit into play has the same chance as becoming a hit as any other. To this, I always point out Mariano Rivera. A lot of contact is made off of Rivera, but much of this contact is made on cut fastball right on the fists. I really don't think contact made on pitches like this have the same chance as becoming a hit as a guy who gives up just as many balls in play, but off the good part of the bat.
I reall don't think a curve on a 3-2 pitch to the 5 batter (saying the pitcher wouldnt throw this curve to a 2 or 3 hitter due to protection) has as much chance of being hit into play for a hit as a 3-2 fastball to a 2 or 3 hitter (even if adjusted for the difference between a 2 or 3 batter and a 5 batter). And all of that is assuming a 3-2 curve has as much chance of contact being made as a 3-2 fastball, which I dont think it does.

AngelAstro
04-01-04, 06:36 PM
Originally posted by nyybleachercreature
I know some sabermetric people have said that any ball hit into play has the same chance as becoming a hit as any other. To this, I always point out Mariano Rivera. A lot of contact is made off of Rivera, but much of this contact is made on cut fastball right on the fists. I really don't think contact made on pitches like this have the same chance as becoming a hit as a guy who gives up just as many balls in play, but off the good part of the bat.

I wouldn't tell that to Luis Gonzalez. :(


Originally posted by nyybleachercreature
I reall don't think a curve on a 3-2 pitch to the 5 batter (saying the pitcher wouldnt throw this curve to a 2 or 3 hitter due to protection) has as much chance of being hit into play for a hit as a 3-2 fastball to a 2 or 3 hitter (even if adjusted for the difference between a 2 or 3 batter and a 5 batter). And all of that is assuming a 3-2 curve has as much chance of contact being made as a 3-2 fastball, which I dont think it does.

I'm not exactly sure what you are trying to say with this paragraph. I'm guessing that because of lack of protection the #5 batter would get a worse pitch to it and thus would have a less of a chance of putting the ball in play. OK, but there is also a better chance that the curve ball would be a ball, so then the batter would reach base via walk. Assuming that the curve is a strike, there is a better chance that the batter would K, thus lowering all stats (AVG, OBP, SLG). OTOH, the 3-2 fastball to the #2 or #3 batter has a better chance of being put into play. I still believe that once the ball is put in play the defense has the final say as to whether the ball in play will become a hit. Thus, just getting the ball in play does not necessarily mean one should see a significant increase in AVG or SLG.

nyybleachercreature
04-01-04, 07:03 PM
I think its fair to say a ball put in play from a fastball has more chance of becoming a hit than a ball hit in play from a curve. Thats what Im saying

AngelAstro
04-01-04, 07:09 PM
Originally posted by nyybleachercreature
what Im saying is getting the ball in play increases the chances of having a higher BA. It doesnt mean it will, just probably will. I think its fair to say a ball put in play has more chance of becoming a hit than a ball not hit in play.

I would have to agree with you there. :)
All I'm saying is that because a protectee has a great hitter or protection behind, does not mean that one should necessarily expect an increase in BA.

markp
04-02-04, 12:08 AM
I don't understand all of this talk about fastballs vs. curves. Instead of trying to explain why something "should" happen, why not look at whether it did or not. The research done for both articles (and a lot more research not linked) show there's no such thing.
This is just like the "clutch" argument. Because it doesn't seem possible to you that something doesn't exist doesn't mean that you can "prove" it does regardless of reality.
Quite simply, there's no such thing as protection.
A classic example that's been ignored here is (and is often cited as proof that protection does exist) is Maris and Mantle in 1961. The claim is that nobody wanted to walk Maris in 1961 after Houk switched him to 3rd and Mantle to 4th, meanwhile Mantle was walked a lot because he didn't have Maris protecting him any more.
1. Maris walked 70 times in 499 ABs batting behind Mantle 1960. He walked 94 in 590 ABs in 61.
2. Maris OPS+ in 1960 was 161. In 61 it was 167. He had essentially the same season, but 61 looks a lot different because he had 100 more ABs and had fewer 2Bs and 3Bs, a lower BA, and 22 more HRs.
3. Mantle had an OPS+ of 164 in 1960 with Maris protecting him. In 1961 it was 206. Gaining 42 points in OPS+ when unprotected kind of blows the whole protection thing up pretty well.

It doesn't exist.

AngelAstro
04-02-04, 01:18 AM
I'm pretty sure the people that have spoken before don't believe that protection exists. I know I don't buy into the protection argument. I was just trying to reason the idea of protection out. Those numbers of Maris and Mantle are very convincing.

I suppose I should begin to look up more of the numbers, but I like to try to come to my own conclusions first and I'm lazy :)

LoneRedSeat
04-02-04, 09:49 AM
Look at the NL for the flip side of the argument.

The 8 spot in the order in the NL is regarded as the worst spot to hit. Pitchers don't throw the 8th batter much to hit because they know they can go after an easy out in the 9th spot.

It is easy to conclude NL pitchers would approach the 8th batter much differently if there were a true hitter in the 9 hole.

AngelAstro
04-02-04, 10:14 AM
Originally posted by LoneRedSeat
Look at the NL for the flip side of the argument.

The 8 spot in the order in the NL is regarded as the worst spot to hit. Pitchers don't throw the 8th batter much to hit because they know they can go after an easy out in the 9th spot.

It is easy to conclude NL pitchers would approach the 8th batter much differently if there were a true hitter in the 9 hole.

This may be true, but the number 8 batter in most NL lineups is just as bad as the 9th hitter in American league lineups. So, if Enrique Wilson is batting 8th for the Mets, why wouldn't the pitcher go after him? Since the pitcher is batting next, seems to me that it would be an easy 2 outs in a row. Plus, if there are 2 outs in the inning and the 8th batter comes up, I would do my hardest to get that 8th batter out so that the next inning the pitcher bats first so that you have 1 out before the top of the lineup gets at bat. So, in most situations, I would imagine the pitcher would want to pitch to get that 8th batter out.

markp
04-02-04, 10:33 AM
do the same when they aren't batting 8th in the NL. Moving to the AL, where they have the lead-off hitter to "protect" them yields identical results to hitting 8th in the NL.

AngelAstro
04-02-04, 10:40 AM
Originally posted by markp
do the same when they aren't batting 8th in the NL. Moving to the AL, where they have the lead-off hitter to "protect" them yields identical results to hitting 8th in the NL.

:link:

Or was it in one of the old Abstracts? I really need to get my hands on one of those.

nyybleachercreature
04-02-04, 12:21 PM
Originally posted by markp
I don't understand all of this talk about fastballs vs. curves. Instead of trying to explain why something "should" happen, why not look at whether it did or not. The research done for both articles (and a lot more research not linked) show there's no such thing.
This is just like the "clutch" argument. Because it doesn't seem possible to you that something doesn't exist doesn't mean that you can "prove" it does regardless of reality.
Quite simply, there's no such thing as protection.
A classic example that's been ignored here is (and is often cited as proof that protection does exist) is Maris and Mantle in 1961. The claim is that nobody wanted to walk Maris in 1961 after Houk switched him to 3rd and Mantle to 4th, meanwhile Mantle was walked a lot because he didn't have Maris protecting him any more.
1. Maris walked 70 times in 499 ABs batting behind Mantle 1960. He walked 94 in 590 ABs in 61.
2. Maris OPS+ in 1960 was 161. In 61 it was 167. He had essentially the same season, but 61 looks a lot different because he had 100 more ABs and had fewer 2Bs and 3Bs, a lower BA, and 22 more HRs.
3. Mantle had an OPS+ of 164 in 1960 with Maris protecting him. In 1961 it was 206. Gaining 42 points in OPS+ when unprotected kind of blows the whole protection thing up pretty well.

It doesn't exist.
the whole curveball-fastball stuff came up once I said that in most cases, I don't think protection exists. I said I think it comes into play only when a hitter like Bonds is behind somone-- a hitter that is repected more than anyone in the game (as far as hitting goes).
Also, the thread started by asking about "old school" and "stat" related ideas toward the subject, so, instead of bringing up the way people are looking at it (stats vs old school), why not just talk about what is being talked about.
I would also be interested to see Maris' number after Mantle was hurt and taken out of the lineup.

nyybleachercreature
04-02-04, 12:23 PM
Also, the way I read the articles was that the idea of protection hadn't been proved or disproved given the numbers used and the results.

markp
04-02-04, 01:18 PM
The sample sizes in just these two articles are pretty substantial. And there has been analysis done on other sets of numbers with the same results.
Protection, like "clutch" has been disproven.

nyybleachercreature
04-02-04, 09:15 PM
Originally posted by markp
The sample sizes in just these two articles are pretty substantial. And there has been analysis done on other sets of numbers with the same results.
Protection, like "clutch" has been disproven.
"Looking at overall performance by totaling the 27 protectee’s stats together, it looks like OPS increase by about 3.5%, (BB+K)/PA increased by about 6.4% and TB/H increased by about 2%."
http://www.baseballprimer.com/articles/wright_2003-06-30_0.shtml

I'd say those numbers are pretty insignificant, but as you said, the sample size was pretty big, so the error would probably be pretty small.

markp
04-03-04, 06:18 AM
Originally posted by nyybleachercreature

"Looking at overall performance by totaling the 27 protectee’s stats together, it looks like OPS increase by about 3.5%, (BB+K)/PA increased by about 6.4% and TB/H increased by about 2%."
http://www.baseballprimer.com/articles/wright_2003-06-30_0.shtml

I'd say those numbers are pretty insignificant, but as you said, the sample size was pretty big, so the error would probably be pretty small.

"Interesting article, but what really needs to be addressed is whether the differences are statistically significant. On the one hand, looking at the totals, all the "unprotected" numbers are lower, which does suggest tht something was probably going on. On the other hand, looking at total OBP and using the binomial distribution, we see that, even in 15,894 PA, the 95% confidence interval around that .358 mean is +/- .0076, putting the protected/unprotected split numbers well within random variation."

Since other studies show a decrease, also well within the boundaries of random variation, it's safe to say that protection has been disproven. When the difference between two sets of numbers is always small enough to fall within random variation, there is no real difference.