I hadn't realized we allowed more runs than last year

[TGO]

All I'm saying is that for the 2009 season, Bergesen did indeed perform like those guys. I am not warranting that he will continue to do so. I don't know what he's gonna do next season. But I do think that saying he's a #3 based on what he actually did this season is rather odd. The only reason I can see for doing that is if you value peripherals more than you value things that mattered in the ballgames he pitched. My personal opinion is that the label "peripherals" is a good one, including the meaning of the word. However, it appears that some think those things are not actually peripheral but are central.

You aren't predicting, but others would like to find ways to do so. Predictive use of pitching stats has a long history going back a couple decades, and it has been shown that K/9, BB/9, K/BB, and HR/9 are better predictors of future performance than ERA, IP, or anything else. They are not perfect predictors but no one ever said they were, just that they have closer correlations to future performance than "things that mattered."

[RShack]

Well, it is up to you if you want to ignore the predictive tools that are out there...It is up to you if you want to ignore the data that supports a lot of the theories...But it is pretty ridiculous for you to slam others opinions on this just because they rely on a whole body of work and not some limited skill of being able to truly evaluate what you see over the course of a small sample size.

I'm not doing any of that. I'm just pointing out that the way we talk about who's a #1-#5 is goofy, sloppy, and based on not much. If you disagree with this, please explain to me how your thresholds for #1-#5 make sense, in light of the fact that the league doesn't have nearly enough SP's who meet *any* of them.

As for predictive tools, what are you talking about? When people rate SP's, I don't see anybody referring to predictive tools. I see people just pulling numbers out of their butt based mostly on unadjusted ERA. So, where are these predictive tools, and what basis do they use for saying who's a #1 and who's a #3?

I'm just trying to arrive at some kind of rational shorthand for ID'ing who falls into each category. If you have a better way to do it, that's fine, I'd love to hear it. Based on how unrealistic your thresholds are, I bet you don't. (This is not just you, pretty much everybody seems to do the same basic thing.)

[RShack]

You aren't predicting, but others would like to find ways to do so. Predictive use of pitching stats has a long history going back a couple decades, and it has been shown that K/9, BB/9, K/BB, and HR/9 are better predictors of future performance than ERA, IP, or anything else. They are not perfect predictors but no one ever said they were, just that they have closer correlations to future performance than "things that mattered."

OK, so based on this long history of useful analysis, how do you decide who's a #1, who's a #2, who's a #3, and so on?

The problem is that everybody wants us to get a TOR SP, which evidently means either a #1 or #2.

Yet when it comes to deciding which SP's should be ranked #1-#5, people just make up numbers based on not-much.

So, given all the history of demonstrating predictive value, how come we don't have a rational basis for saying which SP's are which number?

Or, if there is indeed a rational basis, how come nobody can tell me what it is?

[Stotle]

That's not what I'm asking.

I'm asking if you think the way to rate a SP's actual performance should be about peripherals, or whether it should be about actual results that directly matter in the outcome of ballgames, like his ERA+ and how deep he went into a game.

If you are trying to rank performances, then something like ERA+ or FIP is more important than SO/BB. If you're trying to figure out how someone will perform moving forward, stats such as SO/9 are much more important than something like ERA+.

All I'm saying is that for the 2009 season, Bergesen did indeed perform like those guys. I am not warranting that he will continue to do so. I don't know what he's gonna do next season. But I do think that saying he's a #3 based on what he actually did this season is rather odd. The only reason I can see for doing that is if you value peripherals more than you value things that mattered in the ballgames he pitched. My personal opinion is that the label "peripherals" is a good one, including the meaning of the word. However, it appears that some think those things are not actually peripheral but are central.

It isn't odd. This year he pitched like a front-end starter, but his peripherals, stuff, MiL history, etc. all suggest that he is unlikely to perform at this level moving forward. So when I give a general label for what type of pitcher he is, that is putting a number to the type of performance you should expect from him in the future. There isn't any point to assigning a label to past performance. Just pick a stat/metric and line up the pitchers. The point is to try and figure out what that list will look like in the future.

I'm not sure why this is so difficult. If batter A bats .200 but shows a high linedrive percentage and an incredibly low BABIP (say .240) then I'm much more inclined to look at the peripherals and project an improved average next year than I am to look at the production and use that as my guide.

It's the same principle. Variables affect production. What peripherals can do is help you to 1) isolate the things the player has control over, and 2) remove "luck" from the equation as much as possible. What you're left with is hopefully a truer picture of the actual performance capabilities of the particular player, free of extrinsic forces that might affect production.

[El Gordo]

You aren't predicting, but others would like to find ways to do so. Predictive use of pitching stats has a long history going back a couple decades, and it has been shown that K/9, BB/9, K/BB, and HR/9 are better predictors of future performance than ERA, IP, or anything else. They are not perfect predictors but no one ever said they were, just that they have closer correlations to future performance than "things that mattered."

You do realize that this statement is oxymoronic to the point of negating every thing else you say in your post.:laughlol:

[Sports Guy]

I'm not doing any of that. I'm just pointing out that the way we talk about who's a #1-#5 is goofy, sloppy, and based on not much. If you disagree with this, please explain to me how your thresholds for #1-#5 make sense, in light of the fact that the league doesn't have nearly enough SP's who meet *any* of them.

As for predictive tools, what are you talking about? When people rate SP's, I don't see anybody referring to predictive tools. I see people just pulling numbers out of their butt based mostly on unadjusted ERA. So, where are these predictive tools, and what basis do they use for saying who's a #1 and who's a #3?

I'm just trying to arrive at some kind of rational shorthand for ID'ing who falls into each category. If you have a better way to do it, that's fine, I'd love to hear it. Based on how unrealistic your thresholds are, I bet you don't. (This is not just you, pretty much everybody seems to do the same basic thing.)

This is such an uneducated post.

A lot of us routinely talk about K rate, BB rate, HR rate, K/BB rate, FIP, DIPs, xFIP,. BABIP, LD%, LOB%, HR/FB%, gb rate....All of those are good ways to evaluate pitchers, both in the now and more importantly, the future.

They aren;'t things pulled out of our butts..They are legit ways...Just because it is over your head and you don't understand it, doesn't make it worthless.

[Stotle]

I'm not doing any of that. I'm just pointing out that the way we talk about who's a #1-#5 is goofy, sloppy, and based on not much. If you disagree with this, please explain to me how your thresholds for #1-#5 make sense, in light of the fact that the league doesn't have nearly enough SP's who meet *any* of them.

As for predictive tools, what are you talking about? When people rate SP's, I don't see anybody referring to predictive tools. I see people just pulling numbers out of their butt based mostly on unadjusted ERA. So, where are these predictive tools, and what basis do they use for saying who's a #1 and who's a #3?

I'm just trying to arrive at some kind of rational shorthand for ID'ing who falls into each category. If you have a better way to do it, that's fine, I'd love to hear it. Based on how unrealistic your thresholds are, I bet you don't. (This is not just you, pretty much everybody seems to do the same basic thing.)

What you are looking for is a neat, easy and precise solution to vastly complex question. I've laid out my criteria -- it's similar to scouting criteria I've seen from ML scouts and generally matches up with pitchers currently and previously in the game and is reasonably reliable when it comes to trying to predict the levels of success a pitcher might enjoy.

And it should be the case that there are more players that meet the criteria of mid-rotation starters than front-end starters, more back-end starters than mid-rotation starters, and more bullpen arms than back-end starters.

[Frobby]

Dang, I was hoping it was something someone else said in the heat of a debate. There are certain people I wouldn't put it past (past whom I would not put it? ).

Oh well, big ups to Frobby for sure for coming clean. No way anyone would have reminded him.

Actually, Tony Soprano reminded me of this at least twice this season. I'd like to think that I would have paid off anyway.

[RShack]

It isn't odd. This year he pitched like a front-end starter, but his peripherals, stuff, MiL history, etc. all suggest that he is unlikely to perform at this level moving forward. So when I give a general label for what type of pitcher he is, that is putting a number to the type of performance you should expect from him in the future. There isn't any point to assigning a label to past performance. Just pick a stat/metric and line up the pitchers. The point is to try and figure out what that list will look like in the future.

I'm not sure why this is so difficult. If batter A bats .200 but shows a high linedrive percentage and an incredibly low BABIP (say .240) then I'm much more inclined to look at the peripherals and project an improved average next year than I am to look at the production and use that as my guide.

Here's why it's difficult. I don't see how you can predict who's gonna be a #3 if you can't tell me what a #3 is.

I swear, this is why it's mostly pointless to try to discuss interesting things about baseball around here. You just ask a damn question and people have a fit and get snippy if you're not standing in line and saying "Mooo" about the goofy things people do with numbers.

Look, it there's some good way to say who's a #1, #2, #3, etc., that's great, just tell me what it is. And if there isn't a good way to say what those things mean, how in the hell are you gonna predict whether somebody is gonna be one?

[Sports Guy]

OK, so based on this long history of useful analysis, how do you decide who's a #1, who's a #2, who's a #3, and so on?

The problem is that everybody wants us to get a TOR SP, which evidently means either a #1 or #2.

Yet when it comes to deciding which SP's should be ranked #1-#5, people just make up numbers based on not-much.

So, given all the history of demonstrating predictive value, how come we don't have a rational basis for saying which SP's are which number?

Or, if there is indeed a rational basis, how come nobody can tell me what it is?

Why are you so concerned over the label and the # associated here?

[Sports Guy]

Here's why it's difficult. I don't see how you can predict who's gonna be a #3 if you can't tell me what a #3 is.

I swear, this is why it's mostly pointless to try to discuss interesting things about baseball around here. You just ask a damn question and people have a fit and get snippy if you're not standing in line and saying "Mooo" about the goofy things people do with numbers.

Look, it there's some good way to say who's a #1, #2, #3, etc., that's great, just tell me what it is. And if there isn't a good way to say what those things mean, how in the hell are you gonna predict whether somebody is one?

I agree...you should just leave and start your own board.

[RShack]

Why are you so concerned over the label and the # associated here?

Because it's a constant reference point for how people talk about SP's, that's why. People use that about SP's in a way that's similar to how they use OPS to talk about hitters. Yet, while we know what OPS means, we don't really know what the SP numbers mean. The ones you came up with were flat-out unrealistic. You say you want the O's to get a TOR guy, yet your numbers about what constitutes that were goofy. According to the numbers you came up with, there's only like 20 guys in the whole dang league who make up the population fo #1's, #2's, and #3's combined.

I can't believe people are challenging the notion that there should be some actual basis for the quality-labels we use for SP's all the time. But that's exactly what's happening.

[Mackus]

Because it's a constant reference point for how people talk about SP's, that's why. People use that about SP's in a way that's similar to how they use OPS to talk about hitters. Yet, while we know what OPS means, we don't really know what the SP numbers mean. The ones you came up with were flat-out unrealistic. You say you want the O's to get a TOR guy, yet your numbers about what constitutes that were goofy. According to the numbers you came up with, there's only like 20 guys in the whole dang league who make up the population fo #1's, #2's, and #3's combined.

I can't believe people are challenging the notion that there should be some actual basis for the quality-labels we use for SP's all the time. But that's exactly what's happening.

The numbers system for starting pitchers is no different that saying a hitter is a #4 hitter. Its just a very quick and dirty way of describing a player. There are dozens if not more of ways a pitcher could be a #1 or #2 or whatever, just like some #4 hitters are typical 3-outcome hitters, some are more well-rounded, some don't hit a ton of HR but hit for good average and walk a lot.

Its a generic term. It doesn't specify any certain ERA or innings pitched or set of stuff or anything. Its just a general term for how good a pitch is to get a basic idea of how he compares to other starters.

[Stotle]

Here's why it's difficult. I don't see how you can predict who's gonna be a #3 if you can't tell me what a #3 is.

I swear, this is why it's mostly pointless to try to discuss interesting things about baseball around here. You just ask a damn question and people have a fit and get snippy if you're not standing in line and saying "Mooo" about the goofy things people do with numbers.

Look, it there's some good way to say who's a #1, #2, #3, etc., that's great, just tell me what it is. And if there isn't a good way to say what those things mean, how in the hell are you gonna predict whether somebody is gonna be one?

From a predictive/quasi-scouting standpoint, I've given my definitions plenty of times, including a general explanation in our last convo. I'm not retyping it -- feel free to search it.

From a statistical standpoint, it's simply a matter of lining up past performers and establishing tiers. How you break down those tiers might vary depending on the metric you are using. In practice, the lines between tiers are closer to blurry than they are to "bright line". Here are a series of articles that I found useful in getting familiar with some of ideas re: labeling starters (from 2006). None are exactly what you are looking for but maybe they point you in the right direction:

http://www.baseballthinkfactory.org/files/primate_studies/discussion/starting_rotation_analysis/

http://www.hardballtimes.com/main/article/how-good-is-your-4-starter/

http://www.hardballtimes.com/main/article/more-fun-with-rotation-numbers/

http://www.hardballtimes.com/main/article/another-look-at-starting-rotations/

[Sports Guy]

Because it's a constant reference point for how people talk about SP's, that's why. People use that about SP's in a way that's similar to how they use OPS to talk about hitters. Yet, while we know what OPS means, we don't really know what the SP numbers mean. The ones you came up with were flat-out unrealistic. You say you want the O's to get a TOR guy, yet your numbers about what constitutes that were goofy. According to the numbers you came up with, there's only like 20 guys in the whole dang league who make up the population fo #1's, #2's, and #3's combined.

I can't believe people are challenging the notion that there should be some actual basis for the quality-labels we use for SP's all the time. But that's exactly what's happening.

Oh no, we do know it...You just don't agree with it..You just choose to ignore it and just to play the semantics game, like you did with the whole "blow it up" fiasco.

The rest of us are ok with it, so at least we can sleep at night.