Posted by boogerlips on 7/2/2012 9:17:00 PM (view original):

Posted by isack24 on 7/2/2012 3:02:00 PM (view original):
I'm a semi-sabr guy, but there's no way to recreate the variables necessary to prove this by statistical evidence. I'd like to see the parameters any statistician used to make that determination.  (and I say that without judging the results)

Its trivially easy to recreate the variables necessary.

Step 1) Observe hitters when a great hitter is on deck. Record results.
Step 2) Observe hitters when a not so great hitter is on deck. Record results.

Conclusion: If the results from steps 1 and 2 are the same then its safe to conclude that the pitchers pitched the same. 

Oh.  Now I get it.  Let's not worry about pitchers, defense, park effects, etc.  Good point.

I understand the amount of data available, I really do.  But to have a statistically meaningful outcome, the variables have to not only be similar, but the exact same.  And in baseball as much as anything, an elite statistician can prove whatever he/she wants when given similar, not exact variables.

Take, for example, 2011 Ryan Braun vs. 2012 Ryan Braun.  Everyone thought Braun would fall off without Fielder "protecting" him, but he hasn't.  He's been even better.  Is it because "protection" is a myth?  Maybe (and I mean that sincerely, I don't have any dog in this fight).  But it could also be that Braun worked tirelessly in the offseason to improve his plate discipline.  Or maybe it's that the all guys ahead and behind Braun are OBPing at about .130, so no one really cares if Braun gets to swing.  It could be a million things.  Or it could be that protection is a myth.

That's why I'm asking to see these studies and the parameters used, because my guess is that it basically just looked at, using the same example, Braun w/ Fielder and Braun w/o Fielder and then compared stats from a similar time period.

I think saying that protection has been proved to not exist is incorrect. It would be more accurate to say that protection has never been show to exist.

It may seem trivial but I think it is important. It seems quite possible that, in certain situations, there might be some protection effect. Or perhaps some pitchers who are influenced by it.

It was common sabermetric knowledge that there was no such thing as "pitching to the score," but there is now some evidence that some pitchers exhibited statistical trends that look an awful lot like pitching to the score.

If statnerds can't see it on a stat sheet, it must not exist.   I'm having a similar discussion about "clutch" in the MLB forum.

As for HBD, I doubt it's programmed in.

With regards to protection specifically, and why I think nothing like that has ever been shown to exist, is the assumption that "getting pitches to hit" is a good thing. If I get no "pitches to hit," there's a pretty good chance I will get on base.

But, if I can remember the studies on it that I've read, there doesn't appear to be much of an increase in walk rate either.

Of course, one flaw with protection studies is that they tend to assume that a pitcher "fears" the protecting batter in proportion to that batter's effectiveness. Ben Zobrist had like a .950 OPS a few years ago, was he really a "feared" hitter? I honestly don't know, I'm not a major league pitcher. 

If batters have been intentionally walked or "pitched around", it seems to me that it makes sense that pitchers/managers would rather face certain batters.    If that's the case, it makes sense that a guy batting with the "feared" hitter on deck will see more hittable pitches.     Whether he hits them or not is another story. 

Posted by MikeT23 on 7/3/2012 12:00:00 PM (view original):
If statnerds can't see it on a stat sheet, it must not exist.   I'm having a similar discussion about "clutch" in the MLB forum.

As for HBD, I doubt it's programmed in.

A somebody professionally involved in sports statistics (not baseball), I can tell you my three constant sources of frustration are:

(1) People looking at small sample sizes like they mean anything. You're better off just going with a trained coach's eye.

(2) People who scan a whole bunch of data without first forming a hypothesis. If you take a large enough sampling of data, you will always see some truly bizarre things. Just because they correlate, doesn't mean it's a cause.

(3) Assuming the absence of proof is proof of absence. This last one is most relevant to protection- and also clutch hitting. Nobody has proved it is exists. But that doesn't mean it doesn't exist.


FWIW, I tend to think most of those sorts of things are dramatically overblown on a macro level but they certainly exist on a micro level. I don't have any personal experience in lineup protection (because I never played baseball), but it seems entirely possible to me that sometimes a guy might get a good batter in a 3-1 count, see Johnny Noodlebat behind him and say "**** it, I'll just put this guy on base and get the guy behind him out." But it also seems entirely possible to me that a pitcher might say,, "no **** this, I'm getting this guy out right now." So while it probably happens from time to time, it will be hard to prove most of these "emotional" factors, because emotions are much more volatile and will tend to even out to a base level over a relevant sample size.

Posted by MikeT23 on 7/3/2012 12:10:00 PM (view original):
If batters have been intentionally walked or "pitched around", it seems to me that it makes sense that pitchers/managers would rather face certain batters.    If that's the case, it makes sense that a guy batting with the "feared" hitter on deck will see more hittable pitches.     Whether he hits them or not is another story. 

Exactly.

However, just as intentionally walking a guy is generally a bad idea (there needs to a really big gap- like a pitcher on deck- to justify an intentional walk from a statistical sense), the best way to get a guy out is generally to throw a decent pitch in the strike zone. So I may in fact do worse when hitting in front of a "feared" hitter.

And just as it is rare to see guys intentionally walked in the first inning, the effects (if they indeed exist) of lineup protection would probably only surface in more emotionally charged situations. Which again make it even more difficult to show, and easier to disregard on a macro level.

Posted by jtrinsey on 7/3/2012 12:17:00 PM (view original):

Posted by MikeT23 on 7/3/2012 12:00:00 PM (view original):
If statnerds can't see it on a stat sheet, it must not exist.   I'm having a similar discussion about "clutch" in the MLB forum.

As for HBD, I doubt it's programmed in.

A somebody professionally involved in sports statistics (not baseball), I can tell you my three constant sources of frustration are:

(1) People looking at small sample sizes like they mean anything. You're better off just going with a trained coach's eye.

(2) People who scan a whole bunch of data without first forming a hypothesis. If you take a large enough sampling of data, you will always see some truly bizarre things. Just because they correlate, doesn't mean it's a cause.

(3) Assuming the absence of proof is proof of absence. This last one is most relevant to protection- and also clutch hitting. Nobody has proved it is exists. But that doesn't mean it doesn't exist.


FWIW, I tend to think most of those sorts of things are dramatically overblown on a macro level but they certainly exist on a micro level. I don't have any personal experience in lineup protection (because I never played baseball), but it seems entirely possible to me that sometimes a guy might get a good batter in a 3-1 count, see Johnny Noodlebat behind him and say "**** it, I'll just put this guy on base and get the guy behind him out." But it also seems entirely possible to me that a pitcher might say,, "no **** this, I'm getting this guy out right now." So while it probably happens from time to time, it will be hard to prove most of these "emotional" factors, because emotions are much more volatile and will tend to even out to a base level over a relevant sample size.

This is what I posted:

"People react differently to stressful situations.    People even have different ideas to what a stressful situation is.   So is it so crazy to believe that some people handle situations better?"


It doesn't have to apply to baseball.  Some people might get frustrated and stressed in a knitting class.    I've heard plenty of students say "I don't test well".    They know the curriculum but can't put it on paper when the time comes.   As for baseball players, I have no doubt that some guys can compartmentalize each AB so they treat that 4th inning AB in June the same way they treat the 9th inning AB in October.   They might not get the result, after all the greats fail 68% of the time, but they can certainly handle the situation better and thus have a better chance for success.

Mike,

Totally agree. To say that "clutch" does not exist is fails. By definition, if I come through with 2 outs in the 9th, I was certainly clutch. Anybody who has ever played sports at a high level has been in times where the pressure negatively affected them. Or lifted them to a higher place.

The issue is that "clutch" has never shown to be repeatable, and it is quite logical why it will always be extremely difficult to predict. The situation that will scare me the first time I face it might inspire me the second time. Or maybe I come through the first time and now put pressure on myself the second time.

So while it certainly does exist, and is a valid descriptive commentary of events that have already transpired, it holds relatively little predictive power for the future, which makes it something that should probably (in my opinion) hold less weight than many people attribute to it.

The repetition of the situation is another aspect.   Fighters train so their reactions are reactions and not "mind-controlled" movements.   They "plan" for the eventual situation.    To take that a step further, it would stand to reason that a guy who has "been there, done that" would be more adept at dealing with the situaiton the 2nd time around.  I believe the Heat's championship/Thunder's failure was attributed to that.

While I agree that the situation is fluid, a guy who hasn't been there before is a wild card, whereas a guy who has is more likely to be comfortable, or possibly uncomfortable(choker), and thus have a better "feel" for the situation.

Posted by boogerlips on 7/3/2012 12:22:00 PM (view original):

Posted by isack24 on 7/2/2012 10:14:00 PM (view original):

Posted by boogerlips on 7/2/2012 9:17:00 PM (view original):

Posted by isack24 on 7/2/2012 3:02:00 PM (view original):
I'm a semi-sabr guy, but there's no way to recreate the variables necessary to prove this by statistical evidence. I'd like to see the parameters any statistician used to make that determination.  (and I say that without judging the results)

Its trivially easy to recreate the variables necessary.

Step 1) Observe hitters when a great hitter is on deck. Record results.
Step 2) Observe hitters when a not so great hitter is on deck. Record results.

Conclusion: If the results from steps 1 and 2 are the same then its safe to conclude that the pitchers pitched the same. 

Oh.  Now I get it.  Let's not worry about pitchers, defense, park effects, etc.  Good point.

Given enough sample size, all variables except the variable(s) in focus become obsolete. So its correct to say we don't need to worry about "pitchers, defense, park effects etc", all we need to worry about is collecting a large sample of data and focus on the variables we're interested in.

For instance, when you look at a SLB player's performance history, your goal is to understand how he will perform on average. If that player has been used 175 times and averages a .308.375.450 line, then you can sleep easy at night knowing that is solid information because of the sample size. The parks/defense/pitchers are irrelevant at that point. If the player had only been used 2 times, parks/defense/pitchers would be important.

"Given enough sample size, all variables except the variable(s) in focus become obsolete."

Except that I guarantee the studies that were used were not statistically-relevant sample sizes.  How do I know?  Because they can't be.  A few thousand isn't statistically significant. 

You're also talking about a computer sim, which is obviously different from real life, where people are affected by infinite more things.

FWIW, I agree everything jtrinsey has said in this thread.

Posted by bripat42 on 7/1/2012 2:12:00 PM (view original):

Posted by overeasy on 6/29/2012 5:34:00 PM (view original):
On a team where the top 4 hitters in the lineup are studs and there is a big drop-off after that, I have noticed ~15 IBB for the #4 hitter versus 2-3 for the 1-3 guys. As for any difference in effectiveness, I have not observed anything that would lead me to believe that there is any programming for it.

+1

I also noticed  -- anecdotally -- back when I had a NL team that my regular #8 hitter had walks totals higher than his ratings would suggest. Might have been a coincidence, but I do wonder if the sim "pitches around" guys who are followed by substantially weaker batters.

Would this not relate to the IBB settings in the Management Console? If the owner your playing against has it set to 3, for example, I'd expect a #8 hitter (or #4 if there's a lineup drop off, etc.) to receive more IBBs. It has nothing to do with protection being built into the game.

FWIW, Daniel Murphy yesterday said this:

“For me, looking at it, I finally got to hit in front of David (Wright) today. That’s fun. Everyone should get the chance to do that.”

He went 4 for 5 with 2 doubles and a triple.  He seems to think there's an effect.