174 replies to this topic

[CA-ORIOLE]

I've never referred to the board as bandwagoners... I have only spoken about those relatively few who have clearly been on the bandwagon...

 

There is no reason for anybody who hasn't been on the "Cruz sucks" bandwagon to think I've been talking about them... just as there's no reason for you to assert that my input that is critical of a few somehow applies to everybody...

 

If you disapprove of painting others with a broad brush, I suggest you quit doing it...  this isn't high school, you know...

Last I checked, YOU isn't a broad brush, but hey, deflect and ignore stuff.... always a sound strategy. 

 

As always, feel free to have the last word. 

[PaulSporer]

Well, if I'm so impossibly wrong about it, then it should be pretty easy to find some actual evidence that shows I'm wrong... wouldn't you think?

 

My guess is you've been constantly provided with evidence to the contrary of your views, but still ignore it and either move the goalposts are just say you don't agree. Sometimes it's not an opinion. For example, it's a fact that you're not a #3 starter with a 1.42 WHIP. That is far too many base runners to be consistently relied upon as a #3. Expecting his ERA to maintain w/that kind of base runner volume is just not sounding reasoning on any level. 

[PaulSporer]

Lots of words and numbers,,, but nothing that shows that peripherals are the best predictor.... nothing.

 

However, there are several flawed and unsupported statements.... such as "works fairly well as a predictor"... that statement is either wrong or meaningless, depending on how generous you wish to be...

 

If you think anything there demonstrates that peripherals are the best predictors, I don't know what to tell you... because nothing in there does.  

 

However, every r^2 value cited is evidence of that item having only minimal predictive value... which, as I've said umpteen times, means that using it as a basis for personnel decisions is a recipe for making bad decisions.   (It would be different if we were talking about doing millions of automated stock trades, and thus could rely on the eventual normalization of results.... but we're not, so we can't...)

 

I'm happy to keep giving a look at whatever you can find... but I guarantee you that you won't find anything that demonstrates what you erroneously think these things demonstrate...  on the other hand, if you can find something that actually supports the notion that peripherals are the best predictor of future P performance, I will happily proclaim your victory..  

 

In the meantime, trust this:  if there was any such evidence, you wouldn't have to look hard to find it... they would be shouting it from the rooftops, and everybody would know about it... because it would be a major coup... not just in the world of baseball stats, but in the much broader domain of human performance research.  They're not doing that for exactly one reason: they don't have it... and the reason they don't have it is simple:  it doesn't exist...

LOL, is this real life? JFC. 

[Mackus]

My guess is you've been constantly provided with evidence to the contrary of your views, but still ignore it and either move the goalposts are just say you don't agree. Sometimes it's not an opinion. For example, it's a fact that you're not a #3 starter with a 1.42 WHIP. That is far too many base runners to be consistently relied upon as a #3. Expecting his ERA to maintain w/that kind of base runner volume is just not sounding reasoning on any level. 

 

Come on.  I think Shack's been difficult in dealing with this topic, but this is not what a "fact" is.

 

Agree with the rest, that you wouldn't expect someone with that type of WHIP to pitch like a #3.  But "#3 starter" isn't even a fact, its a subjective distinction.  So any subjective criteria required to meet a subjective distinction can't possibly be called a fact.

[PaulSporer]

If a guy walks too many guys, doesn't strike out guys, doesn't miss bats and gives up homers, the chances are far greater that he will struggle vs if he did those things well.

 

Is anyone going to dispute that?

 

My guess is Shack will. 

[PaulSporer]

I give up... this is hopeless.... you're gonna believe what you want to believe...

 

You can just say that in the mirror, you don't really have to post it here, we all already know this about you!   

[PaulSporer]

Come on.  I think Shack's been difficult in dealing with this topic, but this is not what a "fact" is.

 

Agree with the rest, that you wouldn't expect someone with that type of WHIP to pitch like a #3.  But "#3 starter" isn't even a fact, its a subjective distinction.  So any subjective criteria required to meet a subjective distinction can't possibly be called a fact.

 

Cool.

[SportsGuy]

My guess is Shack will.

But the Orioles pitchers of the 70s...

[RShack]

My guess is you've been constantly provided with evidence to the contrary of your views, but still ignore it and either move the goalposts are just say you don't agree. Sometimes it's not an opinion. For example, it's a fact that you're not a #3 starter with a 1.42 WHIP. That is far too many base runners to be consistently relied upon as a #3. Expecting his ERA to maintain w/that kind of base runner volume is just not sounding reasoning on any level. 

 

Well, let's just skip over what you said above about what is or isn't a fact about #3's, shall we?

 

Instead, I would like to know what you believe is a good predictor of future pitching success... and why you think it is.  

 

I'd also like to see the body of evidence that supports it, but since I'm confident you don't have any, I won't ask for that part.  

 

It's pretty amazing that some folks are so sure there is compelling evidence... even though they can't quite find it... and when you ask where that body of evidence is, they just get mad and start being insulting about it.   I guess that's what some folks do when they don't have an actual answer.

[Mike in STL]

My guess is you've been constantly provided with evidence to the contrary of your views, but still ignore it and either move the goalposts are just say you don't agree. Sometimes it's not an opinion. For example, it's a fact that you're not a #3 starter with a 1.42 WHIP. That is far too many base runners to be consistently relied upon as a #3. Expecting his ERA to maintain w/that kind of base runner volume is just not sounding reasoning on any level.

Well, let's just skip over what you said above about what is or isn't a fact about #3's, shall we?

Instead, I would like to know what you believe is a good predictor of future pitching success... and why you think it is.

I'd also like to see the body of evidence that supports it, but since I'm confident you don't have any, I won't ask for that part.

It's pretty amazing that some folks are so sure there is compelling evidence... even though they can't quite find it... and when you ask where that body of evidence is, they just get mad and start being insulting about it. I guess that's what some folks do when they don't have an actual answer.

It's not rocket science. If a guy hits 25 HRs. Then 22, then 17, then 15, how many do you expect him to hit the next year? Over 25, under 15, in between?

Gallardo has given up 8.2, 9.0, 9.1, 9.4 H/9. What will it likely be in 2016? Lower than 8, or higher than 9, in between?

Gallardo K/9 rate went 9.0, 7.2, 6.8, 5.9. What will it likely be in 2016? Over 9 or under 6, in between?

Its just patterns. Elementary school stuff. Nothing more or less. ERA and WHIP have been up and down. But common sense should tell you if the patterns continue, his ERA will go up.

[RShack]

It's not rocket science. If a guy hits 25 HRs. Then 22, then 17, then 15, how many do you expect him to hit the next year? Over 25, under 15, in between?

Gallardo has given up 8.2, 9.0, 9.1, 9.4 H/9. What will it likely be in 2016? Lower than 8, or higher than 9, in between?

Gallardo K/9 rate went 9.0, 7.2, 6.8, 5.9. What will it likely be in 2016? Over 9 or under 6, in between?

Its just patterns. Elementary school stuff. Nothing more or less. ERA and WHIP have been up and down. But common sense should tell you if the patterns continue, his ERA will go up.

If you wanna look at a few numbers and form an opinion that's fine.... it's a free country (more or less)... everybody does it, including me...

What's not fine is to say that peripherals are the best predictors of future P success... because there is no body of evidence that shows that they are, while there is a body of evidence that shows no indication that they are. Claiming they are the best predictors is just as groundless as it would be for me to claim that the metric I prefer is the best predictor... if I wanted to claim that, I would have to demonstrate that it's true...

Bottom line: whether you like them or not is up to you. That's way different than claiming something that's just not true...

BTW, what answers would you give to the questions you asked?

[Mike in STL]

If you wanna look at a few numbers and form an opinion that's fine.... it's a free country (more or less)... everybody does it, including me...

 

What's not fine is to say that peripherals are the best predictors of future P success... because there is no body of evidence that shows that they are, while there is a body of evidence that shows no indication that they are.   Claiming they are the best predictors is just as groundless as it would be for me to claim that the metric I prefer is the best predictor... if I wanted to claim that, I would have to demonstrate that it's true... 

 

Bottom line: whether you like them or not is up to you.  That's way different than claiming something that's just not true...

 

BTW, what answers would you give to the questions you asked?

I didn't read all of this thread, but all it took was a simple google search to find this article.... or, a body of evidence that talks about the best predictors. Has this been brought up?

 

http://www.fangraphs...era-estimators/

 

It proves in multiple scenarios that the SIERA peripheral stat is the best predictor of future success. Is it right 100% of the time? No. Nothing is. Breaking news: Crystal balls don't exist. But you can make educated guesses based on the information available.

 

There is evidence here that suggests SIERA predicts future performance better than anything else out there. It also states that projections like ZIPS and PECOTA are above other estimators like xFIP, tERA, etc... because they take multiple seasons into account. I guess estimators do not. Thats what I gather from the article anyway.

 

But the last line of the article...the summary if you will....

 

"At this stage, using SIERA as a jumping off point appears to be the best method to project pitcher performance."

 

Some here on this board are a lot more qualified to chime in on this. Maybe I'm wrong, maybe I'm interpreting something here incorrectly because I don't have a Masters Degree in Statistics. 

 

But you can't keep saying that there is no body of evidence, when someone put the time and effort into creating one with math, and facts to back it up. It took 10 seconds to find. Google is a great thing.

 

ZIPS is a projection, which is not the same as estimators like SIERA, xFIP, etc..., which I gather from the article (I'm not entirely sure what the difference is since a projection is an estimation, IMO). ZIPS projects Gallardo with the following. 

 

(This was before he signed with the Orioles. Not sure if these numbers are adjusted later now that he is here, pitching in the AL East).

 

166 IP, 108 K, 61 BB, 176 H, 78 ER.

 

Do some math, that is...

 

5.8 K/9 - Worst of his career, lower than last year

3.3 BB/9. - His career average

9.5 H/9 - Worst of his career, higher than last year

4.23 ERA - Worst of his career, almost a full run worse than last year.

1.428 WHIP - Worst of his career, 1.416 last year. 

166 IP - Lowest since becoming a full time starter in 2009.

 

Now...I didn't even need to seek out ZIPS projections to formulate my opinion that I made probably a month ago. The patterns I saw just by looking at the back of his baseball card lead me to believe, pretty much exactly what the projections ended up saying. 

 

I posed the question before, Over/Under on Gallardo. 180 IP, 3.75 ERA, 1.450 WHIP. I went Under, Over, Over. The projection calls it Under, Over, Under. 

 

If SIERA, the leading estimator, puts Gallardo at a 4.59 ERA. And ZIPS, the leading projector puts Gallardo at 4.23, I feel comfortable saying the truth lies somewhere in the middle.

 

Call it 4.41. Add the potential shoulder issue, the decrease in velocity, the AL east ballparks, the internal handling of pitchers, I'll go out on a limb and say he posts damn near a 5.00 ERA in 2016. I'll be utterly shocked, and eat my hat if his ERA is under 4.00 in 2016. This is just my opinion. Maybe I'm wrong. For the O's sake I hope I'm dead wrong.

[CA-ORIOLE]

Cliff notes version FIP versus ERA: 

 

1. ERA and FIP are measures of pitcher effectiveness.

2. FIP is more predictive of FIP than ERA is predictive of ERA.

3. FIP is more predictive of ERA than ERA is predictive of ERA (this appears marginal)

4. FIP an ERA will generally regress towards each other over time and sample size (defense. luck and BP's tend to balance over time). 

5. Some pitchers appear to have the ability to outpitch their FIP, even over time (See note 4). Most of this appears insignificant (Jim Palmer certainly excluded). If this capability was identifiable/available, it would be a highly sought after commodity.   

6. Large variances between FIP versus ERA should raise red flags and probably lean towards FIP being truer. That said, all factors should be looked at.

 

The FIP detractors will minimize FIP as a measure of pitcher performance because of  its limited data scope and counter that FIP does not take into account in game data and generally unidentified pitcher skills (while at the same time mitigating the value of luck/randomness, defense and bullpens etc.that tend to make ERA more variable). This is particularly true in short term evaluations.   

[Matt_P]

2. FIP is more predictive of FIP than ERA is predictive of ERA.

 

Being able to predict FIP is mostly meaningless without point 3. I suppose it would have some value if FIP and a Defensive_Metric would measure Runs Allowed better then ERA. Come to think of it, we should be doing a regression using those variables instead of a correlation using just ERA and FIP.  Whoops.

 

5. Some pitchers appear to have the ability to outpitch their FIP, even over time (See note 4). Most of this appears insignificant (Jim Palmer certainly excluded). If this capability was identifiable/available, it would be a highly sought after commodity. 

 

Given that FIP is an imperfect predictor, shouldn't this be expected?

[CA-ORIOLE]

1. Being able to predict FIP is mostly meaningless without point 3. I suppose it would have some value if FIP and a Defensive_Metric would measure Runs Allowed better then ERA. Come to think of it, we should be doing a regression using those variables instead of a correlation using just ERA and FIP.  Whoops.

 

 

2. Given that FIP is an imperfect predictor, shouldn't this be expected?

1. I'm not sure that it's meaningless. Depends on how you value FIP/DIPS versus ERA. ERA based WAR integrates BIS data/ defensive performance if you want that angle, otherwise FIP is true to DIPS theory. I see no reason to make it something it isn't supposed to be, especially since it was designed almost wholly for the purpose of measuring value while excluding defense. I guess SIERA does some of that. 

2. Yes, obviously.