PDA

View Full Version : Sean Marshall's bad luck



RedsManRick
05-29-2012, 12:37 PM
Sean Marshall is striking out more than 12 batters per 9 IP, walking a hair over 2 and allowing 1 homer. Combined, this figures suggest an ERA exactly in line with the 2.26 he put up last year.

Unfortunately, his ERA is 4.24. Maybe he's just making a lot mistakes? Well, his GB/FB rate is up and his LD% is down. What gives? A .435 BABIP.

There's an article on Baseball Prospectus right now that looks at a very similar reliever from a results standpoint who's going through the same problem, Johnny Venters. He's got a .449 BABIP that is a big reason why he looks like he's scuffling despite pitching pretty well.

http://www.baseballprospectus.com/article.php?articleid=17163#114372

The article actually looks at his hits allowed and judges them based on luck and finds, not surprisingly, that he suffered from a rash of bloops and seeing-eye singles.

We can't automatically assume the same for Marshall, but I'd be willing to bet on a similar finding. If somebody had the time & energy, it would certainly be interesting to take a look.

RedlegJake
05-29-2012, 01:55 PM
I read a piece on Marshall that said except for an unlucky hit or two he really is pitching as well as he ever has. I'm not worried a bit.

wheels
05-29-2012, 10:38 PM
Marshall is as lockdown as he was last year. There is zero doubt in my mind.

I believe Dusty is terribly overreacting. To the detriment of both Chapman and Marshall.

The Voice of IH
05-29-2012, 10:43 PM
Marshall is as lockdown as he was last year. There is zero doubt in my mind.

I believe Dusty is terribly overreacting. To the detriment of both Chapman and Marshall.

I don't think it is to any detriment of anyone. Chapman gets to pitch in high pressure situations, grow as a ball player and limit his innings. Marshall already has a deal, and is back in the role that he feels most comfortable in.

dougdirt
05-29-2012, 11:32 PM
I don't think it is to any detriment of anyone. Chapman gets to pitch in high pressure situations, grow as a ball player and limit his innings. Marshall already has a deal, and is back in the role that he feels most comfortable in.

Has Marshall said he is more comfortable as a set up man?

reds44
05-29-2012, 11:36 PM
I don't understand how it is to the detriment of anyone. Marshall may be as lockdown as he ever was, but Chapman is about as lockdown as anyone has ever been right now.

dougdirt
05-29-2012, 11:42 PM
I don't understand how it is to the detriment of anyone. Marshall may be as lockdown as he ever was, but Chapman is about as lockdown as anyone has ever been right now.

Well, if you planned to see Chapman start at all this year, it is a detriment to that because a team isn't going to remove their closer mid-year to put him in the rotation.

Johnny Footstool
05-29-2012, 11:55 PM
I don't think it is to any detriment of anyone. Chapman gets to pitch in high pressure situations, grow as a ball player and limit his innings.

Which of those things wasn't Chapman doing before Dusty made him the closer?

757690
05-30-2012, 02:03 AM
I donít have a subscription to BP, so I donít know how they are backing up their claim that Ventersí BABIP was mostly due to luck. Iíll just assume it accurate.

But I understand BABIP differently than they do, and donít think a high BABIP is a result of bad luck, nor do I think that a pitcher canít control his BABIP. I think BABIP is just like any other stat, something that pitchers can control with skill and talent. However, for some reason, most pitchers in the majors that last long enough for there to be enough data on the matter, possess very similar skill and talent in regards to controlling their BABIP.

Itís just like any player's other stats, when just looking at that player. Letís take Arroyo for example. Since 2004, he has had a BB rate between 2-3 per 9 innings and a K/BB rate between 2-3. That is his skill and talent level for those stats. We can guess pretty accurately that his current walk level of 1.3 will likely rise and his K/BB rate of 5 will likely lower. We can guess this because we can assume based on his past that his BB rate will be between 2-3 and his K/BB rate will be between 2-3 as well. And if they donít regress back to his career levels this season, we can assume they will next season.

The same is true for his BABIP. It will likely be in the .280-.300 range, just like most other pitchers, and when it did drop to a significantly lower rate of .239 in 2010, it went back up to .278 in 2011. It currently is .324 this season, so we can guess that it will lower over the course of the season, probably near his career average of .283.

Now would we say that Arroyoís current 5.0 K/BB rate is due mostly to luck? No, at least I donít think we should. It most likely is due to the fact that Arroyo is in a grove; that in these first 10 starts of 2013, has been able to be exceptional at controlling his K/BB rate. And, as usually happens, we can assume that this grove will end, and he will have a few starts during this season in which he is not very good at controlling his K/BB rate, which will raise his seasonís rate back to his career norms. When this happens, is this bad luck or a slump? Is this a result of bad luck or of Arroyo not making quality pitches? I would say itís the latter on both counts. In fact, that is what we have said about players when their stats decrease. We donít say itís bad luck, we say they arenít doing something right and need to fix it, and most good players usually do fix it.

I think the same thing happens with BABIP. Currently Marshall has a very high BABIP. Some of that is due to luck, but I would imagine that most of it is due to him throwing too many hittable pitches. Heís in a BABIP slump, and just like if he were in a K slump or HR slump, we would expect him to get out of it and eventually, by seasonís end, have his K or HR or BABIP rates in range with his career rates. Not because he will no longer have bad luck, but because he will no longer throw as many hittable pitches.

The only difference between K, HR, BB rates and BABIP rates is that, for some reason, most pitchers have similar BABIP rates while they do not for most other stats. At least, that is how I understand BABIP.

dougdirt
05-30-2012, 06:40 AM
If Marshall were throwing so many hittable strikes, then his strikeout rate wouldn't be elite.

757690
05-30-2012, 07:12 AM
If Marshall were throwing so many hittable strikes, then his strikeout rate wouldn't be elite.

Why are they mutually exclusive?

Zach Grienke this year has two games in which he pitched a combined 6 innings, struck out a combined 8 batters, yet gave up a combined 19 hits and 15 runs.

I constantly see relievers give up multiple runs and strike out the side in the same inning.

Cooper
05-30-2012, 07:20 AM
That is one convoluted post there 757690.

dougdirt
05-30-2012, 07:22 AM
Why are they mutually exclusive?
Well, generally because you can't be both highly hittable and highly unhittable. It simply defies logic.



Zach Grienke this year has two games in which he pitched a combined 6 innings, struck out a combined 8 batters, yet gave up a combined 19 hits and 15 runs.
That is a situation where K/9 is misleading because he wasn't facing your typical number of hitters in his innings. In those 6 innings, he faced 39 batters. That is about 40% more than he typically faces in 6 innings. So his K/9 was 'high', but 8 out of 39 is 21%, which again is good, but not the same elite number that his 12/9ip would suggest.



I constantly see relievers give up multiple runs and strike out the side in the same inning.

Again, in a small sample size, that works. In the larger sample size though, you aren't going to see that happening and contribute it to some sort of skillset.

PuffyPig
05-30-2012, 07:35 AM
But I understand BABIP differently than they do, and donít think a high BABIP is a result of bad luck, nor do I think that a pitcher canít control his BABIP. I think BABIP is just like any other stat, something that pitchers can control with skill and talent. However, for some reason, most pitchers in the majors that last long enough for there to be enough data on the matter, possess very similar skill and talent in regards to controlling their BABIP.



So, even though BABIP is a controllable stat, for some reason no one ever does, or at least controls it in the exact same manner.

The only logical explanation for that statement is that pitchers have little or no control over their BABIP.

I've never heard the explanation that all pitchers can control BABIP, but they all control it virtually the same. Even if true, it means that BABIP is not being controlled.

757690
05-30-2012, 08:29 AM
Well, generally because you can't be both highly hittable and highly unhittable. It simply defies logic.

That is a situation where K/9 is misleading because he wasn't facing your typical number of hitters in his innings. In those 6 innings, he faced 39 batters. That is about 40% more than he typically faces in 6 innings. So his K/9 was 'high', but 8 out of 39 is 21%, which again is good, but not the same elite number that his 12/9ip would suggest.

Again, in a small sample size, that works. In the larger sample size though, you aren't going to see that happening and contribute it to some sort of skillset.

1) It is not illogical that a pitcher can throw many unhittable pitches in an inning and a few very hittable ones. Happens all the time.

2) Greinke for his career strikes out 21% of all batters he faces. So for those two games, he was exactly on pace in terms of strike outs.

3) We see hitters strike out three times in a game and go 3-6 with two doubles and a homer all the time too. Are you saying there is no skill set attributed to his hitting? Was it all luck? Was it all out of his control?

757690
05-30-2012, 08:38 AM
So, even though BABIP is a controllable stat, for some reason no one ever does, or at least controls it in the exact same manner.

The only logical explanation for that statement is that pitchers have little or no control over their BABIP.

I've never heard the explanation that all pitchers can control BABIP, but they all control it virtually the same. Even if true, it means that BABIP is not being controlled.

I am saying the opposite. That all pitchers control their BABIP in the same manner and get similar results.

Look, there is nothing that I am saying that go against the basic notions behind DIPS, or FIP or other pitching advanced metrics. I am just giving an alternative explanation as to why it's happening. If my theory is true, FIP is still superior to ERA, and one can use BABIP to evaluate pitchers. If a pitcher has a higher then normal BABIP, it will likely regress to the pitcher's career average rate, which will likely lower his ERA, and the reverse is true.

We assume a pitcher can control his K/BB rate, so that we can guess that when it is lower than his career rate, it will regress back to his career rate. We don't say that his lower than normal K/BB rate is the result of bad luck, we say it's because he's not pitching very well at this time. I think the same logic applies to BAPIP.

PuffyPig
05-30-2012, 08:38 AM
We see hitters strike out three times in a game and go 3-6 with two doubles and a homer all the time too. Are you saying there is no skill set attributed to his hitting? Was it all luck? Was it all out of his control?

No one says that hitting is out of the control of the hitter.

Most everyone says that pitchers cannot control balls hit into play.

Apples and oranges.

757690
05-30-2012, 08:44 AM
No one says that hitting is out of the control of the hitter.

Most everyone says that pitchers cannot control balls hit into play.

Apples and oranges.

I am questioning the very notion. Why do people think that? I think they are wrong in that assumption. I gave a very detailed, rarional argument for why I think this notion is wrong.

BCubb2003
05-30-2012, 08:54 AM
BABIP is probably the most counter-intuitive stat of all. I'm sure the numbers back it up, but even if you slice off the balls that are stung but caught, and the dribblers that get through, it just seems reasonable that if a pitcher can control whether a ball is hit at all, the pitcher can control whether a ball is well struck, and a well-struck ball is more likely to be a hit.

Boss-Hog
05-30-2012, 08:59 AM
BABIP is probably the most counter-intuitive stat of all. I'm sure the numbers back it up, but even if you slice off the balls that are stung but caught, and the dribblers that get through, it just seems reasonable that if a pitcher can control whether a ball is hit at all, the pitcher can control whether a ball is well struck, and a well-struck ball is more likely to be a hit.

I agree...I have never understood the stance of those that disagree with this (and it's not for a lack of reading plenty of material here on the subject).

elfmanvt07
05-30-2012, 09:19 AM
I agree that it's not exclusively luck. I think anything from around .250 to .350 could very well be a function of a pitcher hitting/missing his spots, or simply hanging a few curves here and there. Making a pitch that's difficult to hit is certainly something at which pitchers work very diligently.

However, something to keep in mind is that HR are actually factored out of BABIP. If a pitcher throws an EXTREMELY hittable pitch, it could very well leave the ball yard. These aren't counted in BABIP.

Also, once a pitcher's BABIP approaches .400, I'm willing to bet a great deal of that can be attributable to luck. Marshall is at .435 as we speak. Having seen him pitch this year, I saw several sharp grounders be hit almost PERFECTLY through the holes in the infield. I personally think that he HAS been unlucky this year, but that's not the exclusive source of his problems.

traderumor
05-30-2012, 09:31 AM
BABIP is probably the most counter-intuitive stat of all. I'm sure the numbers back it up, but even if you slice off the balls that are stung but caught, and the dribblers that get through, it just seems reasonable that if a pitcher can control whether a ball is hit at all, the pitcher can control whether a ball is well struck, and a well-struck ball is more likely to be a hit.I'm not sure I'd consider it counter-intuitive. Once the ball leaves the pitcher's hand, it is out of his control and a LOT of variables kick in as to whether the ball is a hit or out. That makes plenty of sense to me. I'd say that there are so many variables, that a lot of other things lead to BABIP that make it random and normalize over time.

Also, you beg the question on "well-struck." That term would need an agreed definition to even consider your argument.

IslandRed
05-30-2012, 09:33 AM
I haven't seen any research lately that suggests pitchers have no control over BABIP. Several articles have pointed out that pitchers with different profiles (strikeout rates, groundball rates, etc.) have different baseline BABIPs.

Mariano Rivera's career BABIP is .262 over 1200 innings. This is not an accident of history. While he's an extreme outlier, it proves the point, because his HR/FB rate is also way below the so-called normal level. That cutter of his didn't always miss bats (not that a 8.26 K/9 is shabby) but it sure as heck was hard to square up.

What we can say is, for the vast majority of pitchers, the differences in BABIP are negligible enough that if you ignore them, you're not giving away much of anything in terms of predictive value. Particularly when it comes to young pitchers, since it takes years for a pitcher's true level to be known anyway.

REDREAD
05-30-2012, 10:30 AM
IBut I understand BABIP differently than they do, and donít think a high BABIP is a result of bad luck, nor do I think that a pitcher canít control his BABIP. I think BABIP is just like any other stat, something that pitchers can control with skill and talent. .

Great post. I agree with you.

Marshall has a high BABIP (and ERA) this year because he's hung some curveballs in a small sample size. Overall, he's pitched better than a casual look at his ERA would indicate. Sure, every pitcher has spells where hits tend to just drop in, but I think the effect of bad luck has been exaggerated.

Pitchers and hitters are constantly making adjustments. Welsh was talking last night that Homer has been working on some things with Price. Homer had a great start last night. While we can't conclude cause-and-effect, logic would seem to indicate that Homer's adjustments this year had more to do with his success than luck did last night.

You also make a great point that bad pitchers will have a high BABIP, and they generally don't last long. For example, if I pitched a game, I can guarantee my BABIP would be very high, because I don't have the skill set of a ML pitcher.

REDREAD
05-30-2012, 10:34 AM
Well, generally because you can't be both highly hittable and highly unhittable. It simply defies logic.
.

Each pitch thrown is a discrete event.

It is logical that a pitcher could throw 100 pitches, and have 80 of them be good, and 20 of them be hanging curveballs.

Most starting pitchers have 2-4 pitches they throw. They aren't all equally effective. Also, it's difficult to repeatably throw a pitch "unhittable" and never make mistakes.

REDREAD
05-30-2012, 10:40 AM
The only logical explanation for that statement is that pitchers have little or no control over their BABIP.
.

Again, I disagree with this.

What if a pitcher wanted to increase their BABIP. Of course, they would never want to do this in a real game.

Don't you think Arroyo could go out, throw up meatballs and have a higher BABIP?

If pitchers have NO control over their BABIP, it wouldn't matter what Arroyo chose to throw, the BABIP would be constant. That doesn't make sense.

I have given the example of Maraino Rivera in the past. Without even looking up numbers, I guessed he had a lower than average BABIP. The numbers agreed.
He has a low BABIP because his cutter generates lots of popups.
Someone dismissed this as an "outlier", but this clearly illustrates that a pitcher
can consistently have a lower BABIP than league average.
It's possible.
It's difficult, but possible.
It's so difficult that many pitchers have difficulty doing it every year over their entire careers. IMO, that's why a pitcher with a lower than average BABIP in his career year tends to have a higher one the next year. You could make the same arguement for several other metrics as well.

traderumor
05-30-2012, 10:46 AM
Again, I disagree with this.

What if a pitcher wanted to increase their BABIP. Of course, they would never want to do this in a real game.

Don't you think Arroyo could go out, throw up meatballs and have a higher BABIP?

If pitchers have NO control over their BABIP, it wouldn't matter what Arroyo chose to throw, the BABIP would be constant. That doesn't make sense.

I have given the example of Maraino Rivera in the past. Without even looking up numbers, I guessed he had a lower than average BABIP. The numbers agreed.
He has a low BABIP because his cutter generates lots of popups.
Someone dismissed this as an "outlier", but this clearly illustrates that a pitcher
can consistently have a lower BABIP than league average.
It's possible.
It's difficult, but possible.
It's so difficult that many pitchers have difficulty doing it every year over their entire careers. IMO, that's why a pitcher with a lower than average BABIP in his career year tends to have a higher one the next year. You could make the same arguement for several other metrics as well.It is a poor analysis that uses the exception to attempt to prove the rule. That is why it is an "outlier." Just like sometimes Ohio has mild winters, like this one. Should we now conclude its due to an overall climate change?

jojo
05-30-2012, 10:56 AM
The primary take home with BABIP is that assuming a pitcher is a major league quality pitcher, his BABIP will tend to fall between .290 and .310 with extreme values regressing back to this range given enough innings (with some caveats).

Historically only a small group of outliers have been able to maintain a career BABIP significantly lower than the league average (and generally it's taken like 3000 innings to find such arms). So there are some pitchers who have exhibited measureable "control" of their aggregate BABIP but again, they appear to be extremely rare and their influence on BABIP has been fairly small. On the other hand, injured, bad, or "done" pitchers can clearly maintain a BABIP that is significantly above "normal". They are "hittable" in the sense that their LD rate increases and clearly hitters achieve "better" contact. But this class of arm quickly gets selected against as such pitchers tend to hit the DL or get their walking papers and matriculate out of the majors. Given the weight of historical data it just isn't likely a pitcher can maintain a BABIP that deviates significantly from normal as a "true skill" (i.e. consistently over time) as there just aren't many examples of chronic outperformers and underperformers get culled fairly quickly.

That is why BABIP is useful for pitchers (just like LOB% and HR/FB%). Large swings in BABIP despite little change in peripherals can dramatically impact counting stats but generally are indicative of the influence of randomness and thus offer insight (context) into evaluating performance.

bucksfan2
05-30-2012, 11:01 AM
Great post. I agree with you.

Marshall has a high BABIP (and ERA) this year because he's hung some curveballs in a small sample size. Overall, he's pitched better than a casual look at his ERA would indicate. Sure, every pitcher has spells where hits tend to just drop in, but I think the effect of bad luck has been exaggerated.

Pitchers and hitters are constantly making adjustments. Welsh was talking last night that Homer has been working on some things with Price. Homer had a great start last night. While we can't conclude cause-and-effect, logic would seem to indicate that Homer's adjustments this year had more to do with his success than luck did last night.

You also make a great point that bad pitchers will have a high BABIP, and they generally don't last long. For example, if I pitched a game, I can guarantee my BABIP would be very high, because I don't have the skill set of a ML pitcher.

Im not a huge believer in the normalization of BABIP. I tend to think there is a reason why a pitcher has a high BABIP or a low one. I think they can control it on a game to game basis. Couple of things that lead to that.

1. Major league pitchers are good.
2. Major league hitters will punish bad pitches.

If you have made it to the majors in all likelihood you are a good pitcher. There may be outliers, but the vast majority of pitchers have had success at every level of minor league baseball. When they get to the bigs they have the ability to get major league hitters out. The problem is when you are missing you spots, hanging your breaking balls, getting little movement on your fastball your going to get hit, and hit hard. Sure a ball that a batter centers up has a chance to get caught, but its much less of a chance than if the batter had not hit it well.

If you look at Homer's performance last night you would understand why the Reds were able to turn most of the balls put in play into outs. You have a pitcher who is giving up weak ground balls and fly balls your going to turn the vast majority of those into outs. A weak hit ground ball gives the defender time to move and make a play on the ball. A hard hit ball gives the defender less time to get to it.

I do think an issue with a normalizing babip is pitchers having good games and pitchers having bad games. They will have their good stuff one day and have their bad stuff the next game. I think that leads to normalization more than luck.

elfmanvt07
05-30-2012, 11:09 AM
Something worth mentioning is that even though Marshall's BABIP is .435, his LD% is actually DOWN from last year, from 20% to 19%. This contradicts certain parts of the hard hit balls theory.

jojo
05-30-2012, 11:12 AM
Something worth mentioning is that even though Marshall's BABIP is .435, his LD% is actually DOWN from last year, from 20% to 19%. This contradicts certain parts of the hard hit balls theory.

We're looking at 17 innnings (less than 80 batters faced). As samples go, his counting stats are completely vulnerable to randomness.

elfmanvt07
05-30-2012, 11:14 AM
We're looking at 17 innnings (less than 80 batters faced). As samples go, his counting stats are completely vulnerable to randomness.

I very much agree, but if someone is arguing that his BABIP is due to more hard hit balls, I believe that his LD% contradicts that a bit.

jojo
05-30-2012, 11:15 AM
I very much agree, but if someone is arguing that his BABIP is due to more hard hit balls, I believe that his LD% contradicts that a bit.

Absolutely.

757690
05-30-2012, 11:24 AM
I very much agree, but if someone is arguing that his BABIP is due to more hard hit balls, I believe that his LD% contradicts that a bit.

Not necessarily. Totally different subject, but the separation of hits into three categories is a bit inaccurate, imo.

There are plenty of hard hit ground balls and fly balls. Clearly, line drives are more likely to become hits; ahowever, a deep drive off the wall, or in the gap, is counted as a fly ball and a scorching two hopper through the hole between the SS and 3B is counted as a ground bell. Both are likely the result of a hittable pitch.

In such a small sample size as Marshall's this season, a few of those can throw off his BABIP and not effect his LD rate at all.

elfmanvt07
05-30-2012, 11:44 AM
I agree, but I also think that having a large selection of ground balls become hits, even if they're hard hit, is bad luck more than anything.

bucksfan2
05-30-2012, 11:55 AM
Not necessarily. Totally different subject, but the separation of hits into three categories is a bit inaccurate, imo.

There are plenty of hard hit ground balls and fly balls. Clearly, line drives are more likely to become hits; ahowever, a deep drive off the wall, or in the gap, is counted as a fly ball and a scorching two hopper through the hole between the SS and 3B is counted as a ground bell. Both are likely the result of a hittable pitch.

In such a small sample size as Marshall's this season, a few of those can throw off his BABIP and not effect his LD rate at all.

IMO a pitcher has more control over hard hit balls. I think velocity is a better determing factor than LD, FB, and GB. It would be much harder to measure, but a hard hit ball has a better chance of finding a hole than a weak LD or lazy FB.

Patrick Bateman
05-30-2012, 12:06 PM
I get the counter side to BAPIP debates....

But in this case, I think we need to look at things more rationally and in the proper context. We have an extremely successful pitcher in Marshall, has not had any issues in the past with a well above average BAPIP in the past, whom's peripherals are generally well in line with his career norms, and has pitched 17 innings.

Regardless of what fence you are on in the BAPIP debate, we know that this is a stat that is capable of randomly fluctuating in smaller samples. We also know that no pitcher who has been successful over a long period of time has ever had a BAPIP above, say .330 over very long stretches.

I think there is ample evidence there to suggest that Marshall's BAPIP over his first 17 innings this year is nothing but a bad random fluctuation. There simply isn't any evidence out there to suggest that a pitcher, especially one that is performing well in every other area is even capable of controlling his BAPIP badly to this extent.

If Marshall was just dipping his toes into the majors for the first time, this might be a different argument (or if Marshall was generally performing badly in other areas), but at this point, the conclusion is quite clear to me. There could be examples of players controlling their BAPIP negatively, this is not one of them.

traderumor
05-30-2012, 12:29 PM
I do think an issue with a normalizing babip is pitchers having good games and pitchers having bad games. They will have their good stuff one day and have their bad stuff the next game. I think that leads to normalization more than luck.I think "randomness" really eliminates a lot of the debate in this discussion, while "luck" usually just stirs a volatile pot, as in begging the question if "luck" exists. Chance and randomness are better descriptors here and less volatile.

Sorry for the semantical lecture, but I think it is important to be a part of this discussion for clarification purposes.

bucksfan2
05-30-2012, 01:37 PM
I get the counter side to BAPIP debates....

But in this case, I think we need to look at things more rationally and in the proper context. We have an extremely successful pitcher in Marshall, has not had any issues in the past with a well above average BAPIP in the past, whom's peripherals are generally well in line with his career norms, and has pitched 17 innings.

Regardless of what fence you are on in the BAPIP debate, we know that this is a stat that is capable of randomly fluctuating in smaller samples. We also know that no pitcher who has been successful over a long period of time has ever had a BAPIP above, say .330 over very long stretches.

I think there is ample evidence there to suggest that Marshall's BAPIP over his first 17 innings this year is nothing but a bad random fluctuation. There simply isn't any evidence out there to suggest that a pitcher, especially one that is performing well in every other area is even capable of controlling his BAPIP badly to this extent.

If Marshall was just dipping his toes into the majors for the first time, this might be a different argument (or if Marshall was generally performing badly in other areas), but at this point, the conclusion is quite clear to me. There could be examples of players controlling their BAPIP negatively, this is not one of them.

I think more apt questions in asking about Marshall's effectiveness would be:

Has Marshall hung more curveballs this year than last? Has Marshall pitched too many days in a row without rest? Has Marshall been put in situations that aren't as favorable this season as the last?

In the past I would imagine that Marshall had been used in adventageous situations. He would come in later in the game but more than likely faced at least one left handed batter. Early on this year he had come in facing who ever was due up in the 9th regardless of handedness. Its easy to just say "Marshall has been had bad luck" but more importantly looking deeper into the issue may lead you to a different conclusion.

Patrick Bateman
05-30-2012, 01:57 PM
I think more apt questions in asking about Marshall's effectiveness would be:

Has Marshall hung more curveballs this year than last? Has Marshall pitched too many days in a row without rest? Has Marshall been put in situations that aren't as favorable this season as the last?

In the past I would imagine that Marshall had been used in adventageous situations. He would come in later in the game but more than likely faced at least one left handed batter. Early on this year he had come in facing who ever was due up in the 9th regardless of handedness. Its easy to just say "Marshall has been had bad luck" but more importantly looking deeper into the issue may lead you to a different conclusion.

How do those questions at all identify the specific question as to why his BAPIP, and literally nothing else are out of line with career norms.

The questions you are asking are to answer the question "why is Marshall struggling and will he rebound." That's not what this question is about... so i would hardly consider those "apt".

dougdirt
05-30-2012, 02:38 PM
You know what a hanging curveball is? A curveball that a guy hits. I have seen guys refer to hanging curveballs low in the zone, middle of the zone, high in the zone, outer part of the zone, inner part of the zone. I have seen bad curves swung on and missed and good ones absolutely crushed. Most 'hanging' curveballs are simply curveballs where the batter knew it was coming and was able to hit it.

REDREAD
05-30-2012, 03:33 PM
It is a poor analysis that uses the exception to attempt to prove the rule. That is why it is an "outlier." Just like sometimes Ohio has mild winters, like this one. Should we now conclude its due to an overall climate change?

When people are using statistics to make an absolute statement such as
"Pitchers can't control their BABIP" then how many exceptions need to be stated in order to prove the absolute statement false?

Now some people have said "Most pitchers can't control their BABIP", but it's still being pulled out as an absolute.. Such as "Pitcher X has a BABIP of Y, so his performance will decline/improve".. We went through an entire offseason arguing that Matt Belisle would have a better year because he had a bad BABIP the season before.. That did not come to fruition..

hebroncougar
05-30-2012, 03:38 PM
You know what a hanging curveball is? A curveball that a guy hits. I have seen guys refer to hanging curveballs low in the zone, middle of the zone, high in the zone, outer part of the zone, inner part of the zone. I have seen bad curves swung on and missed and good ones absolutely crushed. Most 'hanging' curveballs are simply curveballs where the batter knew it was coming and was able to hit it.

Hearing an announcer say "hanging" curveball drives me nuts. About 90% of the time, they're dead wrong.

traderumor
05-30-2012, 04:09 PM
When people are using statistics to make an absolute statement such as
"Pitchers can't control their BABIP" then how many exceptions need to be stated in order to prove the absolute statement false?

Now some people have said "Most pitchers can't control their BABIP", but it's still being pulled out as an absolute.. Such as "Pitcher X has a BABIP of Y, so his performance will decline/improve".. We went through an entire offseason arguing that Matt Belisle would have a better year because he had a bad BABIP the season before.. That did not come to fruition..I haven't seen any absolute statements in this discussion. Any statistical discussion cannot deal in absolutes, so isn't that understood? And how many "exceptions" does it take to disprove the rule of thumb? 4? 5? 23?

jojo
05-30-2012, 04:19 PM
When people are using statistics to make an absolute statement such as
"Pitchers can't control their BABIP" then how many exceptions need to be stated in order to prove the absolute statement false?

Now some people have said "Most pitchers can't control their BABIP", but it's still being pulled out as an absolute.. Such as "Pitcher X has a BABIP of Y, so his performance will decline/improve".. We went through an entire offseason arguing that Matt Belisle would have a better year because he had a bad BABIP the season before.. That did not come to fruition..

For practical purposes it is accurate to generalize that pitchers can't control their BABIP because pitcher BABIP regresses to .300 (.290-.310) given enough sample size owing a few extreme outliers. That's a fact borne by observation.

Or to put it another way, for every time someone mistakenly (but understandably) argues with certainty that pitchers have absolutely no control over BABIP, there are like 10 people arguing Reds pitcher X has bucked the trend and has a true skill that allows them to maintain a lower than expected BABIP. In other words, someone suggesting BABIP will regress and this will affect a pitcher's counting stats accordingly will be correct much, much more often than someone arguing a pitcher has the true skill ability to maintain a low BABIP.

That said, no one is arguing for a blanket approach. I might be one of the bigger sabermetric proponents and anti-era guys on redszone concerning pitcher evaluation and I've argued unequivically for an approach that considers multiple factors when evaluating pitcher performance.

BTW, it might surprise some to realize that since 2010, only one relief pitcher in baseball has amassed more WAR than Belisle (Sean Marshall) and this season only one relief pitcher in baseball has amassed more WAR than Belisle (Chapman). Belisle's BABIP during that stretch? It's regressed from the .364 of his final year as a Red to .312 during his time as a Rockie.

RedlegJake
05-30-2012, 04:21 PM
Thing is when dealing with stats you almost always have to deal with large numbers to have any validity. There are statistical arguments published that show pitchers CAN control BABIP to a small degree BUT I would expect that to matter over the course of a season or an entire career not a half dozen games or a third of a season. There just aren't enough random events in that small a time span to draw any conclusion but that any variation in that brief span is one of randomness or luck - call it what you want. Marshall's BABIP is insanely high - to me that isn't a factor of whether the ball is hit hard or soft but a factor of whether it randomly lands where they ain't to paraphrase Wee Willie. And he's been unlucky so far in that batted balls have randomly landed where the fielders can't get to them in a percentage that dramatically outweighs both league and his personal averages. All his other pitching peripherals suggest nothing else has changed. Ergo - it is likely nothing else likely has changed except for an unlucky incidence of randomness so far.

757690
05-30-2012, 05:06 PM
For practical purposes it is accurate to generalize that pitchers can't control their BABIP because pitcher BABIP regresses to .300 (.290-.310) given enough sample size owing a few extreme outliers. That's a fact borne by observation.

Or to put it another way, for every time someone mistakenly (but understandably) argues with certainty that pitchers have absolutely no control over BABIP, there are like 10 people arguing Reds pitcher X has bucked the trend and has a true skill that allows them to maintain a lower than expected BABIP. In other words, someone suggesting BABIP will regress and this will affect a pitcher's counting stats accordingly will be correct much, much more often than someone arguing a pitcher has the true skill ability to maintain a low BABIP.

That said, no one is arguing for a blanket approach. I might be one of the bigger sabermetric proponents and anti-era guys on redszone concerning pitcher evaluation and I've argued unequivically for an approach that considers multiple factors when evaluating pitcher performance.

BTW, it might surprise some to realize that since 2010, only one relief pitcher in baseball has amassed more WAR than Belisle (Sean Marshall) and this season only one relief pitcher in baseball has amassed more WAR than Belisle (Chapman). Belisle's BABIP during that stretch? It's regressed from the .364 of his final year as a Red to .312 during his time as a Rockie..

I am proposing that even though nearly all pitcher's BABIP will regress to a narrow range between .280-.310, they still control their BABIP. Just to repeat, it is like every other stat that a players controls with his skill and talent, it fluctuates over the season due to slumps and periods of quality play, but at the end of the season, will regress to that narrow range which is that player's career average.

Brandon Phillips will produce an OPS between the narrow range of .750-.800 every season, as that is what his talent and skill has demonstrated he will OPS since he became a Red seven years ago. His OPS during that time is .778. At different points in the season, his OPS might be well above that range, or well below the range, but based on his career, we can assume that his OPS will by season's end regress to around .775.

When his OPS is well below that range, we do not say its because of bad luck or randomness, we say it's because he's in a slump. He doing something wrong, pulling off the ball swinging at pitches outsode his zone, taking too many pitches, etc. and needs to correct it. He usually does and it regresses to his career norms by season's end. We say that he can control his OPS.

Similarly, pitchers can control their BABIP. When it's high, they are in a slump, when it's low, they are in a grove. Over a season, the slumps and groves even out and pitchers' BABIP regress to their career norms.

The only difference is that the career rates for BABIP are very similar for all pitchers. But that doesn't change the fact that the fluctuations in it throughout the season are due to the same factors as the flucuations of other stats... Slumps and groves.

757690
05-30-2012, 05:23 PM
BTW, it might surprise some to realize that since 2010, only one relief pitcher in baseball has amassed more WAR than Belisle (Sean Marshall) and this season only one relief pitcher in baseball has amassed more WAR than Belisle (Chapman). Belisle's BABIP during that stretch? It's regressed from the .364 of his final year as a Red to .312 during his time as a Rockie.

Not derail this thread, but Matt Belisle is a very interesting case statistically speaking.

While his BABIP was very high as a Red in 2008, his peripherals weren't very good either. His K/9 rate was 4.25 that season, his BB/9 rate was 1.82 giving him a 2.33 K/BB rate. His HR/9 rate was 1.21. His FIP that season was 4.55.

When he became a Rocky, his BABIP did normalize to around .310, but more importantly, his K/9 rate nearly doubled to 7.9, his BB/9 rate was almost cut in half to 1.5, his K/BB rate skyrocketed to 5.24 and his HR/9 rate dropped to 0.8.

I think this demonstrates that the real reason why Belisle became a great reliever as a Rocky, is that he learned how to pitch. He became a better pitcher. He was unlucky in 2008, but even if he become lucky later, if he didn't improve in other areas, he would have still been a below average pitcher.

jojo
05-30-2012, 05:37 PM
.

I am proposing that even though nearly all pitcher's BABIP will regress to a narrow range between .280-.310, they still control their BABIP. Just to repeat, it is like every other stat that a players controls with his skill and talent, it fluctuates over the season due to slumps and periods of quality play, but at the end of the season, will regress to that narrow range which is that player's career average.

Brandon Phillips will produce an OPS between the narrow range of .750-.800 every season, as that is what his talent and skill has demonstrated he will OPS since he became a Red seven years ago. His OPS during that time is .778. At different points in the season, his OPS might be well above that range, or well below the range, but based on his career, we can assume that his OPS will by season's end regress to around .775.

When his OPS is well below that range, we do not say its because of bad luck or randomness, we say it's because he's in a slump. He doing something wrong, pulling off the ball swinging at pitches outsode his zone, taking too many pitches, etc. and needs to correct it. He usually does and it regresses to his career norms by season's end. We say that he can control his OPS.

Similarly, pitchers can control their BABIP. When it's high, they are in a slump, when it's low, they are in a grove. Over a season, the slumps and groves even out and pitchers' BABIP regress to their career norms.

The only difference is that the career rates for BABIP are very similar for all pitchers. But that doesn't change the fact that the fluctuations in it throughout the season are due to the same factors as the flucuations of other stats... Slumps and groves.

Here's the thing concerning that theory. The major league average bat throughout the history of stats has an OPS of .785 (.262/.329/.456). However, OPS doesn't regress back to that as clearly hitters can maintain OPS levels that are significantly higher or lower than that.

The fact that pitcher BABIP regresses back to .290-.310 suggests there is something intrinsic to the fabric of the game such that once the ball leaves a pitcher's hand 30% of batted balls in play should be expected to end up being a hit.

elfmanvt07
05-30-2012, 06:26 PM
Here's the thing concerning that theory. The major league average bat throughout the history of stats has an OPS of .785 (.262/.329/.456). However, OPS doesn't regress back to that as clearly hitters can maintain OPS levels that are significantly higher or lower than that.

The fact that pitcher BABIP regresses back to .290-.310 suggests there is something intrinsic to the fabric of the game such that once the ball leaves a pitcher's hand 30% of batted balls in play should be expected to end up being a hit.

Well said.

BCubb2003
05-30-2012, 06:34 PM
I'm not sure I'd consider it counter-intuitive. Once the ball leaves the pitcher's hand, it is out of his control and a LOT of variables kick in as to whether the ball is a hit or out. That makes plenty of sense to me. I'd say that there are so many variables, that a lot of other things lead to BABIP that make it random and normalize over time.

Also, you beg the question on "well-struck." That term would need an agreed definition to even consider your argument.

First of all, I salute your proper use of "beg the question." That's rarer than a four-homer game.

The thing is, a baseball doesn't take on a whole new set of characteristics as soon as it leaves the pitcher's hand. The pitcher has imparted direction, velocity and movement that will take place on the way to the plate. I conceded that the numbers seem to back up the idea of BABIP, but I find it more intuitive that if a pitcher can pitch to contact or miss bats, the pitcher can control whether the ball is hit hard or not. It might also be possible for a good pitcher to pitch to his defense, while a lesser pitcher allows more hits to "where they ain't."

757690
05-30-2012, 06:40 PM
Here's the thing concerning that theory. The major league average bat throughout the history of stats has an OPS of .785 (.262/.329/.456). However, OPS doesn't regress back to that as clearly hitters can maintain OPS levels that are significantly higher or lower than that.

The fact that pitcher BABIP regresses back to .290-.310 suggests there is something intrinsic to the fabric of the game such that once the ball leaves a pitcher's hand 30% of batted balls in play should be expected to end up being a hit.

Which is true and which does not negate my theory. Pitchers can control their BABIP, and there could be something in the fabric of the game which causes 30% of balls put into play to end up as hits.

For me, the big question that people smarter than myself should be investigating is, "what exactly is intrinsic in the fabric of the game that results in 30% of all balls put into play to end up as base hits?"

PuffyPig
05-30-2012, 06:47 PM
For me, the big question that people smarter than myself should be investigating is, "what exactly is intrinsic in the fabric of the game that results in 30% of all balls put into play to end up as base hits?"

The fact that "pitchers have no control over balls hit into play" is what is instrinsic in the fabric of the game that results in 30% of all balls hit into play to end up as base hits .

Now the 30% is "what it is". It actually changes over time as the number in the 60's was less.

757690
05-30-2012, 06:51 PM
The fact that "pitchers have no control over balls hit into play" is what is instrinsic in the fabric of the game that results in 30% of all balls hit into play to end up as base hits .

Now the 30% is "what it is". It actually changes over time as the number in the 60's was less.

That just is not necessarily true, I presented a counter argument to that. The fact that it changes over time only re-inforces my theory. That implies that player's skill and talent levels have changed over time, which we know to be true.

jojo
05-30-2012, 07:49 PM
Which is true and which does not negate my theory. Pitchers can control their BABIP, and there could be something in the fabric of the game which causes 30% of balls put into play to end up as hits.

For me, the big question that people smarter than myself should be investigating is, "what exactly is intrinsic in the fabric of the game that results in 30% of all balls put into play to end up as base hits?"

My guess (and this is a bit of hand waving) is that while there might be something interesting in this messiness, it's probably just a kinda boring interaction between talent distribution and environment (both of which can shift with era of the game).

Johnny Footstool
05-30-2012, 11:40 PM
BABIP is volatile (like batting average), but over the course of a pitcher's career, it tends to average out.

For example, Pedro Martinez (who was a McCracken BABIP case study) tended to post good BABIPs, with a few bad years in the bunch.

http://www.fangraphs.com/graphs.aspx?playerid=200&position=P&page=7&type=full

jojo
05-31-2012, 01:20 PM
Max Scherzer is essentially in the same place as Marshall concerning an apparent dicotomy between his peripherals and BABIP and he offers a player's perspective:

http://www.fangraphs.com/blogs/index.php/max-scherzer-on-his-high-babip-and-k-rate/


“Have I had some bad luck? I can think of a few times I have and a few times I haven’t. As a pitcher, it’s hard to say that you’ve pitched in bad luck, because you know that you’ve made some mistakes. I can’t sit here and slouch those mistakes off.”

REDREAD
05-31-2012, 02:26 PM
BTW, it might surprise some to realize that since 2010, only one relief pitcher in baseball has amassed more WAR than Belisle (Sean Marshall) and this season only one relief pitcher in baseball has amassed more WAR than Belisle (Chapman). Belisle's BABIP during that stretch? It's regressed from the .364 of his final year as a Red to .312 during his time as a Rockie.

That's true that Belisle has been more successful since leaving the Reds.
However, he was a starter as a Red, and a reliever since he left.
I think it's more likely he's suited to relieving than he had a string of bad luck throughout his entire Reds career.

Marshall (and other relievers) actually have had a similiar parallel of being more succesful relieving.

I'm actually ok with accepting that most pitchers settle in a certain range of BABIP. I guess my main question is this.. Let's say a pitcher has a run of 10 games with either higher or lower BABIP... Is that due to bad luck or is it due to the pitcher pitching better/more poorly than normal? Maybe a combination of both? It's hard to say definitively.

Arroyo had his best season in 2006. That's the season he had the lowest ERA and lowest BABIP (270). Was that luck or skill? I think 270 is below the normal range (290-310) If I have the normal range wrong, I apologize. Interestingly enough, since 2003, Bronson has only 2 years with BABIP above 290. Kind of hard to believe he's been lucky for most of his career. Even in last season's disaster, his BABIP was only 278

jojo
05-31-2012, 03:51 PM
Arroyo had his best season in 2006. That's the season he had the lowest ERA and lowest BABIP (270). Was that luck or skill? I think 270 is below the normal range (290-310) If I have the normal range wrong, I apologize. Interestingly enough, since 2003, Bronson has only 2 years with BABIP above 290. Kind of hard to believe he's been lucky for most of his career. Even in last season's disaster, his BABIP was only 278

Here's a previous post from the archives that addresses you last point:

http://www.redszone.com/forums/showpost.php?p=2511099&postcount=721



My take on Arroyo?

His ERA has been driven by his BABIP both of which have dramatically improved with the significant improvement of the defense behind him. In other words, significant interaction between his performance and the Reds defense is a large reason why he's outperformed his peripherals so dramatically (i.e. his ERA has been better than his FIP would predict) over the last several seasons. Or to say it another way, put him in front of the Reds defense of 2006-2008 and his ERA would be much closer to his FIP.

Why do I argue this?

Here is a breakdown of Arroyo by year showing his ERA, FIP and BABIP. When the FIP-ERA is positive, it means his ERA was lower than his peripherals would predict (i.e. his ERA outperformed his FIP):




Arroyo
Season ERA FIP FIP-ERA BABIP
2006 3.29 4.15 0.86 0.271
2007 4.23 4.57 0.34 0.309
2008 4.77 4.5 -0.27 0.314
2009 3.84 4.78 0.94 0.265
2010 3.88 4.61 0.73 0.239
2011 5.07 5.71 0.64 0.278

It's important to note that Arroyo's ERA has outperformed his FIP in 5 of the 6 seasons he's been a Red. But many would agree his 2006 ERA was an anomaly that was unsustainble. Certainly the magnitude of difference between his FIP and ERA spanning the 2009-2011 seasons would not have been expected based upon his prior performance or legitimately ascribed to a skillset. Realizing some may argue this point, below follow a few tables that hopefully demonstrate why one might make the above statement.

Here's the same breakdown for the Reds' pitching staff over the same years:




Reds
Season ERA FIP FIP-ERA BABIP
2006 4.51 4.63 0.12 0.31
2007 4.94 4.55 -0.39 0.31
2008 4.55 4.53 -0.02 0.312
2009 4.18 4.66 0.48 0.283
2010 4.01 4.18 0.17 0.288
2011 4.16 4.45 0.29 0.282

Here are the BABIP for the Reds staff and Arroyo for the two periods of his tenure as a Red (2006-2008 where the Reds had one of the worst defenses in the league and 2009-2011 where the Reds had one of the best defenses in the league):



BABIP by defensive performance
Reds Arroyo
2006-08 0.311 0.298
2009-11 0.284 0.261
Decrease -0.026 -0.037

The data above indicates that for the period of 2009-2011 when Arroyo's ERA has significantly outperformed his FIP despite declining peripherals, the Red's staff as a whole has consistently outtperformed it's FIP as well. The third table suggests the reason why-the dramatically improved defense has driven a large part of this outcome. So he did not outperform his FIP independent of significant influence by his defense.

So in other words, if the Reds were to pay market value for Arroyo's production over the last several seasons, they'd essentially be "paying double" for the cost associated with building their defense. This also can explain why Arroyo does not have a great deal of trade value despite his ERA's.

All of that said, a look at Arroyo as a Reds does seem to suggest he has outperformed his peripherals to a greater degree than can solely be explained by the impact of the defense behind him. So there may be room to poke at something interesting here, albeit a minor effect. However, when looking at his time as a Pirate and BoSock, he displayed no discernible ability to consistently outperform his peripherals.

The ultimate takehome? If they can get a legit arm for their rotation, they should. At least Arroyo should not be a rationale for preventing such a trade.