It's time to talk about the Sox young pitching

[Wanne]

1 hour ago, ChiSox59 said:

Between Rodon, Giolito, Lopez, Fulmer, Hansen, Cease, Dunning, Adams, Clarkin, Stephens, Guerrero, Puckett, I am pretty certain they can find 5 very good rotation options.

The ones that don't stick in the rotation join Burdi, Hamilton, Johnson, Vieria, Fry, and Bummer in the pen.

There is a world of depth and talent there.  I just listed 18 guys that the Sox control long term and are all legit major league prospects.  The pitching is the least of my concerns. 

This!!!!!   I'm pretty excited for Hansen and Dunning to get up here at some point.  There's ton's of talent.

On the pen side...don't forget about Ryan Burr either. 

[turnin' two]

30 minutes ago, Eminor3rd said:

That would depend on your definition of "lots." There may (or may not) be hundreds over the course of baseball history, but it isn't "most," which is the whole point. 

They're not superfluous. If I asked you, "does pitcher X's ERA seem sustainable?" you wouldn't be able to answer accurately by telling me whether or not his K's were high. The reason it exists in the from it does is so that it can combine predictive factors and scale it to be on par with ERA. A guy could have high strikeouts but still suck because he walks too many batters. How could you tell if the walks were high enough to cancel out the strikeouts? You use FIP (or something similar). 

Also, flyball rate has nothing to do with FIP or xFIP. You might be thinking of HR/FB%, which does.

Well, the thing that makes it superfluous is that if you look at a guys stats, you not only see Ks, but also BBs.  So the relationship is right there for you.

https://www.fangraphs.com/library/pitching/fip/

FIP uses homeruns.

 

xFIP uses flyballs

 

https://www.fangraphs.com/library/pitching/xfip/

 

I guess I just don't see the use for it.  It seems to me like having a stat for the sake of having a stat to talk about.  It isn't anything groundbreaking or, at least in my opinion, revealing.

Edited May 14, 2018 by turnin' two

[Dick Allen]

21 minutes ago, turnin' two said:

There are plenty of outliers.

Buehrle picked off 100 guys in his career.  That equates to about 6 per year.  That seems like a stretch to say it is a major factor.  His defense was great.  He made on average 11 putouts per season.  So he equated for 17 puts per year with his defense and pickoffs.   I don't know.  Seems like kind of a stretch to me.   I guess I would like to see the article though.

While I agree with you that his move and defense weren’t gigantic factors, a good move even not picking a guy off can slow advancement on hits.

i think His success was mostly due to his command, and his controlling the tempo of the game. When his command was off a bit, he usually got tattooed.

 

as for FIP and xFIP, I wouldn’t be too concerned about it yet. Lopez ever gets a little command, his k rate will skyrocket, and his walk rate will drop. I think the concern with Fulmer and Giolito is it seems more often than not, they cannot come close to hitting their spots, and now it looks like the are going to start running like crazy on Gio.

Edited May 14, 2018 by Dick Allen

[Eminor3rd]

57 minutes ago, turnin' two said:

Well, the thing that makes it superfluous is that if you look at a guys stats, you not only see Ks, but also BBs.  So the relationship is right there for you.

https://www.fangraphs.com/library/pitching/fip/

FIP uses homeruns.

 

xFIP uses flyballs

 

https://www.fangraphs.com/library/pitching/xfip/

 

I guess I just don't see the use for it.  It seems to me like having a stat for the sake of having a stat to talk about.  It isn't anything groundbreaking or, at least in my opinion, revealing.

Pretend there's a pitcher with the following stats: 2.81 ERA, 9.5 K/9, 4.1 BB/9, 0.9 HR/9, 9.5% HR/FB

If I asked you, "How likely is it that that pitcher's ERA is sustainable?" How would you answer? How confident would you be in your answer?

According to DIPS, there IS an answer that can be derived from those numbers; you have enough information, technically. You might think, ok, that ERA is pretty low so the bar for sustainability is probably kind of high. That K rate seems alright. Would have been above average a few years ago but with the way strikeouts are up now, I guess it isn't super high. Not terrible. I wonder what average is, maybe around 8? Probably a good sign, then, if I have it right. Okay that definitely seems like more walks than you want, but maybe not egregious -- do the strikeouts cancel out the walks? Are the walks a bigger problem than the strikeouts are a benefit? The homerun rate is a bit low but not crazy low. I imagine that will probably go up and affect his ERA a bit. But how much? Is it a run here or there?

Or, you could look at FIP, which takes all of those factors, weighs them mathematically based on how they've been related to rates historically, adjusts for offensive environment, adjusts for league, and then gives you a number than can be compared directly to ERA. Then, it's simply a matter of saying either "Okay, that's relatively close," or "Whoa, that's WAY off. There's reason to believe that that performance will produce a different result."

The purpose of FIP is to put those underlying (K/BB/HR) numbers into context so that they can most precisely (and as objectively as possible) answer the question "is this pitcher's performance likely to continue?"

This doesn't make the counting stats obsolete or useless -- this isn't a matter of whether or not a stat is "better" or "worse" -- it's just a matter of a stat being derived to take the guesswork out of a specific question in which those stats are applied.

[Eminor3rd]

1 hour ago, ptatc said:

This is what I've seen as well and is part of the issue. People look at FIP because it is better than looking at ERA. However, the only thing FIP predictsis ERA.

So you are correct in saying that that FIP doesn't really predict future performance. FIP only states that if the strikeout/walk ratio doesn't change he is likely to have a higher ERA.  However, it seems fairly common sense that the more batter you walk (and don't negate with strikeouts) the more likely you are to have a higher ERA.

Thus a higher FIP doesn't truly predict much of anything. Thus it is accurate to say he has given up fewer runs than the K/BB ratio says he should, whether that is due to good defense, weather conditions or just plain luck.

Yes -- that's a bit of an oversimplification, but not much, you basically have it. And that's extremely useful, because you assume that over the long run, those things all even out.

So now imagine you are trying to acquire a pitcher. You have two options, both of whom have produced good ERAs. Do you want the one whose peripherals are in line with his ERA, or the one whose peripherals suggest that his ERA is a result of external factors?

A pitcher can't control the defense, weather, or luck -- and all those factors might ultimately screw him in the end -- but you'd rather start with a guy who is controlling the things he CAN control well.

[ptatc]

5 minutes ago, Eminor3rd said:

Yes -- that's a bit of an oversimplification, but not much, you basically have it. And that's extremely useful, because you assume that over the long run, those things all even out.

So now imagine you are trying to acquire a pitcher. You have two options, both of whom have produced good ERAs. Do you want the one whose peripherals are in line with his ERA, or the one whose peripherals suggest that his ERA is a result of external factors?

A pitcher can't control the defense, weather, or luck -- and all those factors might ultimately screw him in the end -- but you'd rather start with a guy who is controlling the things he CAN control well.

Sure but personally I would look at the K/BB ratio or WHIP instead. FIP does that as well but it also factos in HR. If you look at HR you need to look at the park factor and down the line.

[Eminor3rd]

1 minute ago, ptatc said:

Sure but personally I would look at the K/BB ratio or WHIP instead. FIP does that as well but it also factos in HR. If you look at HR you need to look at the park factor and down the line.

Park, league, and era factors are all baked into FIP. The difference between FIP and xFIP is that the latter regresses homerun rate to current league average -- making it more of a "predictor" than normal FIP, which is , as you said, descriptive.

Like I said to turnin' two -- FIP is useful specifically because it combines those things in a way that can be compared apples to apples with ERA. It isn't more or less valuable than its components, it's a specialized expression of those components.

[turnin' two]

6 minutes ago, Eminor3rd said:

Park, league, and era factors are all baked into FIP. The difference between FIP and xFIP is that the latter regresses homerun rate to current league average -- making it more of a "predictor" than normal FIP, which is , as you said, descriptive.

Like I said to turnin' two -- FIP is useful specifically because it combines those things in a way that can be compared apples to apples with ERA. It isn't more or less valuable than its components, it's a specialized expression of those components.

Where is park factor baked in?

 

https://www.fangraphs.com/library/pitching/fip/

 

I don't see it anywhere.

[turnin' two]

26 minutes ago, Eminor3rd said:

 

The purpose of FIP is to put those underlying (K/BB/HR) numbers into context so that they can most precisely (and as objectively as possible) answer the question "is this pitcher's performance likely to continue?"

 

This is the purpose of FIP, from fangraphs.

 

"FIP is an attempt to isolate the performance of the pitcher by using only those outcomes we know do not involve luck on balls in play or defense; strikeouts, walks, hit batters, and home runs allowed. Research has shown that pitchers have very little control on the outcome of balls in play, so while we care about how often a pitcher allows a ball to be put into play, whether a ground ball goes for a hit or is turned into an out is almost entirely out of their control."

 

So no, not exactly.

[Eminor3rd]

3 minutes ago, turnin' two said:

This is the purpose of FIP, from fangraphs.

 

"FIP is an attempt to isolate the performance of the pitcher by using only those outcomes we know do not involve luck on balls in play or defense; strikeouts, walks, hit batters, and home runs allowed. Research has shown that pitchers have very little control on the outcome of balls in play, so while we care about how often a pitcher allows a ball to be put into play, whether a ground ball goes for a hit or is turned into an out is almost entirely out of their control."

 

So no, not exactly.

Further down in the same article:

"Essentially, FIP is an attempt to measure how well a pitcher actually performed independent of factors outside of his control that contribute to runs allowed based statistics."

Edited May 15, 2018 by Eminor3rd

[Eminor3rd]

5 minutes ago, turnin' two said:

Where is park factor baked in?

 

https://www.fangraphs.com/library/pitching/fip/

 

I don't see it anywhere.

Park factor is in xFIP

[chitownsportsfan]

You'll like this eminor.  This guy ran a pretty large sample using common DIPS stats to predict future RA/9 and FIP came out the best over a larger sample, over SIERRA, K/BB rate and xFIP, among others.  But it still wasn't very good, so he changed some of the inputs (and this, again, as you linked, goes back to linear weights and Tom Tango) to make what he called "pFIP", which was more predictive over the sample he ran using historical data.

I googled and didn't really find much work on continuing development of pFIP, although there does seem to be some similar efforts underway.  I'm sure the clubs have long had stuff like this, at least the smarter ones.

Anyways, it's a really small sample for Lopez.  Gio I'm not even touching.  The eye test on him is puke, you don't need no saber stats.  Lopez is more nuanced in his "struggles".

[turnin' two]

35 minutes ago, Eminor3rd said:

Park factor is in xFIP

https://www.fangraphs.com/library/pitching/xfip/

 

Again, I don't see it in there.  It uses a league average HR/FB ratio.  Averaged, not park factor.

[Dam8610]

2 hours ago, turnin' two said:

There are plenty of outliers.

Buehrle picked off 100 guys in his career.  That equates to about 6 per year.  That seems like a stretch to say it is a major factor.  His defense was great.  He made on average 11 putouts per season.  So he equated for 17 puts per year with his defense and pickoffs.   I don't know.  Seems like kind of a stretch to me.   I guess I would like to see the article though.

Turns out it's Grantland, and the article isn't about the FIP-ERA difference, but it is brought up as part of the discussion of his defense and holding baserunners making him an outlier.

http://grantland.com/features/mark-buehrle-surprising-success/

[turnin' two]

10 minutes ago, Dam8610 said:

Turns out it's Grantland, and the article isn't about the FIP-ERA difference, but it is brought up as part of the discussion of his defense and holding baserunners making him an outlier.

http://grantland.com/features/mark-buehrle-surprising-success/

Great article.  Thank you for posting it.  Really a lot of great stuff in there and Mark was truly one of a kind.

I have to say that this, though:

" A pitcher who doesn’t throw hard, and doesn’t have a gimmick pitch, and doesn’t strike out a lot of batters, can’t be a successful starter in the major leagues. He might be able to do it once with a lot of luck, but he can’t do it year after year after year."

is a really funny statement in an article that also references Tom Glavine.  

[turnin' two]

2 hours ago, Eminor3rd said:

Pretend there's a pitcher with the following stats: 2.81 ERA, 9.5 K/9, 4.1 BB/9, 0.9 HR/9, 9.5% HR/FB

If I asked you, "How likely is it that that pitcher's ERA is sustainable?" How would you answer? How confident would you be in your answer?

According to DIPS, there IS an answer that can be derived from those numbers; you have enough information, technically. You might think, ok, that ERA is pretty low so the bar for sustainability is probably kind of high. That K rate seems alright. Would have been above average a few years ago but with the way strikeouts are up now, I guess it isn't super high. Not terrible. I wonder what average is, maybe around 8? Probably a good sign, then, if I have it right. Okay that definitely seems like more walks than you want, but maybe not egregious -- do the strikeouts cancel out the walks? Are the walks a bigger problem than the strikeouts are a benefit? The homerun rate is a bit low but not crazy low. I imagine that will probably go up and affect his ERA a bit. But how much? Is it a run here or there?

Or, you could look at FIP, which takes all of those factors, weighs them mathematically based on how they've been related to rates historically, adjusts for offensive environment, adjusts for league, and then gives you a number than can be compared directly to ERA. Then, it's simply a matter of saying either "Okay, that's relatively close," or "Whoa, that's WAY off. There's reason to believe that that performance will produce a different result."

The purpose of FIP is to put those underlying (K/BB/HR) numbers into context so that they can most precisely (and as objectively as possible) answer the question "is this pitcher's performance likely to continue?"

This doesn't make the counting stats obsolete or useless -- this isn't a matter of whether or not a stat is "better" or "worse" -- it's just a matter of a stat being derived to take the guesswork out of a specific question in which those stats are applied.

In that particular case (and boy do I hope it is real so I can see the outcome...) I would just happen to guess that this pitcher is outperforming his FIP.  The K's and HR/9 rate look pretty good, the walks are high.  The walks are too high.  They are something that just looking at, you say, he has to fix that or it is going to cause him problems.  I would guess a guy like that would have a FIP of about low to mid 3 or so. 

Edited May 15, 2018 by turnin' two

[ptatc]

1 hour ago, Eminor3rd said:

Further down in the same article:

"Essentially, FIP is an attempt to measure how well a pitcher actually performed independent of factors outside of his control that contribute to runs allowed based statistics."

Here is also part of the issue. It says it in the definition. It is an attempt to look at only the factors that the pitcher can control. Does looking at only these factors control enough of the variance of all factors involved in scoring runs.

[Eminor3rd]

59 minutes ago, ptatc said:

Here is also part of the issue. It says it in the definition. It is an attempt to look at only the factors that the pitcher can control. Does looking at only these factors control enough of the variance of all factors involved in scoring runs.

I'm not aware of any method/study that solves that.

You've actually hit on the central point of the great debate between fWAR and bWAR. 

fWAR (which is FIP-based) is generally considered more accurate, but less complete. This is because it includes on the factors that we're can statistically "prove" are under the pitcher's (nearly) sole control. FanGraphs decided they would rather accept that there are performance factors missing in the model -- and leave it up to the reader to decide how much of an influence these factors have -- than to make a gut-based guess as to how much credit to assign to different components.

bWAR (which is ERA-based) is generally considered vastly less accurate, but totally complete. It's less accurate because it gives the pitcher all the blame/credit for things the pitcher cannot control (defense, conditions, luck, etc.), which everyone knows is incorrect, but ultimately includes every descriptive aspect of run scoring. 

Which one is better, as with all of these stats, depends upon the context in which it is being used. In a Cy Young contest, for example, those who feelthe definition of "valuable" should be more closely related to what actually happened during the season, regardless of whether luck or teammates affected it, would much prefer the complete story that bWAR tells. Those looking to evaluate a pitcher for his "true talent," however, are incentivized to reduce error and strip extraneous and non-repeatable factors out of the equation entirely. If pitcher X had a great ERA simply because he played in front of a great defense, you may not want him if your defense isn't that good.

Edited May 15, 2018 by Eminor3rd

[Eminor3rd]

2 hours ago, turnin' two said:

https://www.fangraphs.com/library/pitching/xfip/

 

Again, I don't see it in there.  It uses a league average HR/FB ratio.  Averaged, not park factor.

I'm sorry -- it's the index versions of the stats that have the league/park factors. I misspoke.

[Eminor3rd]

1 hour ago, turnin' two said:

In that particular case (and boy do I hope it is real so I can see the outcome...) I would just happen to guess that this pitcher is outperforming his FIP.  The K's and HR/9 rate look pretty good, the walks are high.  The walks are too high.  They are something that just looking at, you say, he has to fix that or it is going to cause him problems.  I would guess a guy like that would have a FIP of about low to mid 3 or so. 

It's not, I just made up an example of stats that seemed reasonable to illustrate that you can't make the conclusion because it isn't an apples to apples comparison. Just trying to find ways to illustrate its use.

[Eminor3rd]

2 hours ago, chitownsportsfan said:

You'll like this eminor.  This guy ran a pretty large sample using common DIPS stats to predict future RA/9 and FIP came out the best over a larger sample, over SIERRA, K/BB rate and xFIP, among others.  But it still wasn't very good, so he changed some of the inputs (and this, again, as you linked, goes back to linear weights and Tom Tango) to make what he called "pFIP", which was more predictive over the sample he ran using historical data.

I googled and didn't really find much work on continuing development of pFIP, although there does seem to be some similar efforts underway.  I'm sure the clubs have long had stuff like this, at least the smarter ones.

Anyways, it's a really small sample for Lopez.  Gio I'm not even touching.  The eye test on him is puke, you don't need no saber stats.  Lopez is more nuanced in his "struggles".

Super interesting -- I remember this study being referenced a lot in articles and podcasts at the time, but I don't think I ever actually read it.

I think you're spot on with Lopez/Giolito. With Lopez, FIP should be referenced more as a snapshot than anything else -- by the time a trend sets in, his arsenal and skill will have changed the inputs.

[BamaDoc]

Some great discussions and thanks for everyones input.  A different question I haven't seen specifically is what has happened to Giolitto?  In spring training his numbers and reports had me expecting him to turn the corner and be an anchor in our rotation.  Not so in season.  I haven't seen him on tv etc.  Is there anything anyone has spotted mechanically?  Physical?  Mental?

[Balta1701]

18 minutes ago, BamaDoc said:

Some great discussions and thanks for everyones input.  A different question I haven't seen specifically is what has happened to Giolitto?  In spring training his numbers and reports had me expecting him to turn the corner and be an anchor in our rotation.  Not so in season.  I haven't seen him on tv etc.  Is there anything anyone has spotted mechanically?  Physical?  Mental?

Someone posted his "release point" data in the last game that he was pitching and there was a clear distinction between where his hand was releasing his fastball from his curveball, something hitters can likely pick up on.

I think you've got a combination of not being ready for the weather and a guy struggling for consistency with his delivery. His stuff is good enough to beat AAA level hitters, but he has to work through these difficulties and develop consistency and better ways to attack hitters before he can turn into a big league pitcher. 

[ptatc]

1 hour ago, Balta1701 said:

Someone posted his "release point" data in the last game that he was pitching and there was a clear distinction between where his hand was releasing his fastball from his curveball, something hitters can likely pick up on.

I think you've got a combination of not being ready for the weather and a guy struggling for consistency with his delivery. His stuff is good enough to beat AAA level hitters, but he has to work through these difficulties and develop consistency and better ways to attack hitters before he can turn into a big league pitcher. 

That actually is very common and one reason pitchers have gone away from the true curveball IMHO. To throw a true 12 to 6 curve you need to come straight over the top to pull straight down on the ball. Very few pitchers throw the fastball from straight over the top because they don't want it stay straight.

[Dam8610]

1 hour ago, ptatc said:

That actually is very common and one reason pitchers have gone away from the true curveball IMHO. To throw a true 12 to 6 curve you need to come straight over the top to pull straight down on the ball. Very few pitchers throw the fastball from straight over the top because they don't want it stay straight.

Can a 12-6 curve be adapted to, say, an 11-5 curve without losing effectiveness?