FSN interview with Terry Ryan [July 15, 2013]

[ThePuck]

If you want to stop talking about Kevin Correia, then stop talking about Kevin Correia. I didn't bring him up in this thread. You did that.

 

No, I didn't bring him up in this thread. The OP brought Correia up. I responded about him and other points the OP brought up concerning Ryan's quotes. People focused on the Corriea part of my original post, which drove this. On top of that, I've tried to drop it and you keep coming back with him in replies to me, even when I don't even mention Correia.

 

In any event, I'll try this again, done with the Correia part of this on this thread.

[Brock Beauchamp]

I'm actually going to disagree with you here. I think the QS is a fine shorthand stat. Obviously it's imperfect (5 ER & 0 ER or 9 IP & 4 ER > 6 IP & 3 ER) but in the long run the 6IP/3ER barometer generally does a decent job of separating the good starts from the bad starts. It shouldn't be treated as gospel but it tends to give you a pretty accurate portrayal of how often the guy in question has pitched deep and given his team a solid chance to win.

 

I'm not saying the stat is without use entirely. If you get a large enough sample size, the metric's arbitrary nature will sort itself out over the long run, as many stats do.

 

But in a world where we have hundreds of metrics available to you, why use something that is so obviously flawed over the short term and unnecessarily boxes itself in? Any time you reduce the number of inputs in a metric and combine it with a shorter period of sample time, the more flawed your stat.

 

Quality Starts takes IP and Earned Runs, two stats that are broken down into the many pieces every start (usually 4-9 for IP, 1-infinity for ER), and reduces them to one boolean output in total (each input doesn't even get its own output, as both inputs must be met for a "yes" output). So, over the course of, let's say a month, you have reduced ~35 IP and ~17 ER (in other words, many input/output figures) into 5-6 boolean "yes" or "no" outputs. Instead of getting a more nuanced look at a guy's pitching (say, 36.2 IP over 5 starts with 14 ER allowed), you get an output of "4".

 

And that's not even bringing up the completely arbitrary nature of "6 IP, 3ER".

 

In my experience, QS is often used by people who want to prove a point with a single number, except every other stat says something differently so they can't use any conventional metrics to illustrate their point. It's way too easy to use the QS to cherry-pick whatever information you want to tell because it's so incomplete and arbitrary.

[Mike Sixel]

My issue with the other stats, Brock, is that they peanut butter starts. I think there is validity in the assumption that starts are a bit independent of each other. I kind of like QS, as a stat........but that's because I believe starts are more independent that probably others do.

[ThePuck]

My issue with the other stats, Brock, is that they peanut butter starts. I think there is validity in the assumption that starts are a bit independent of each other. I kind of like QS, as a stat........but that's because I believe starts are more independent that probably others do.

 

I like looking at the whole picture. To throw out quality starts is short-sighted since it is a good barometer of what you want your starting pitcher to do in today's pitch count dominated game. Go a good amount of innings, keep the team in the game.

[Brock Beauchamp]

Did I try and use QS as a single number, or did I mention multiple numbers like FIP, ERA, talking about innings pitched?

 

My point isn't that you didn't use any other numbers, it's that you leaned a bit on the worst of them. I'll argue all day on the merits of a guy's FIP, ERA, IP, whatever... But not QS, especially over a short period of time. IMO, it's a junk stat over a short period of time (and still mostly junk over a long period of time).

[Major League Ready]

ATTEMPTING to fix the rotation long term. Nothing wrong with that, but there's no guarantee there.

 

Is there a guarantee with a different approach?

[ThePuck]

My point isn't that you didn't use any other numbers, it's that you leaned a bit on the worst of them. I'll argue all day on the merits of a guy's FIP, ERA, IP, whatever... But not QS, especially over a short period of time. IMO, it's a junk stat over a short period of time (and still mostly junk over a long period of time).

 

Sorry, I just deleted that post...it wasn't going to take the debate anywhere. But, no, I didn't lean on any particular stat overall.

[Brock Beauchamp]

I like looking at the whole picture. To throw out quality starts is short-sighted since it is a good barometer of what you want your starting pitcher to do in today's pitch count dominated game. Go a good amount of innings, keep the team in the game.

 

Sure, in theory.

 

If the stat wasn't arbitrary and junk. If QS actually tried to reasonably gauge the actual quality of a start instead of tossing numbers on a board and saying "meet these qualifications exactly and you win!", I'd be more on board with it. I'm all for breaking down pitching starts into "good", "bad", and "mediocre"... Except QS doesn't even try to do that much.

 

Guys can easily pitch better than the metric demands and not get awarded the stat. That means it's a bad statistic.

 

That's not short-sighted. That's simple analysis. There's a reason why teams like the Rays and A's scoff at the metric (and others of its ilk). Given the amount of real stats available to us, it's like using a chisel to make a Ferrari while the competition is using a supercomputer and robots.

[ThePuck]

Is there a guarantee with a different approach?

 

No, but then again, I'm not the one saying Ryan fixed the future rotation with the players he got as if it's a fact. If you can tell me, for sure, our rotation is going to be fixed with those three, I'll rescind my comment. Mind you, I'm not the only one who said it. Also understand, I was very happy with those trades as I've said multiple times. But we don't know for sure he's fixed it. We know he's made a good effort to do so, which is not the same thing.

[ThePuck]

Sure, in theory.

 

If the stat wasn't arbitrary and junk. If QS actually tried to reasonably gauge the actual quality of a start instead of tossing numbers on a board and saying "meet these qualifications exactly and you win!", I'd be more on board with it. I'm all for breaking down pitching starts into "good", "bad", and "mediocre"... Except QS doesn't even try to do that much.

 

Guys can easily pitch better than the metric demands and not get awarded the stat. That means it's a bad statistic.

 

That's not short-sighted. That's simple analysis. There's a reason why teams like the Rays and A's scoff at the metric (and others of its ilk). Given the amount of real stats available to us, it's like using a chisel to make a Ferrari while the competition is using a supercomputer and robots.

 

okay

[Kwak]

Liriano is what I term a "Good Time Charlie" guy--when things are going well, he does really well. If he's on a good team that's playing well--he looks great. The problem occurs when things get tough and the pressure is on, then Liriano folds-up like a road map. Liriano is not the pitcher to place your destiny on--he will disappoint!

[beckmt]

Might want to sign Pelfrey for the next couple of years if he keeps improving. Otherwise, might as well flip him and get something (hopefully a C+ prospect) in return. Do not expect a big return on any Twins pitchers with the exception of Perkins.

[DaveW]

Few things:

 

They should sign pelfrey for two years now.

 

Anyone who blames the twins for Gomez is ridiculous, they gave him plenty of time and had a better in house option in Span, would u have preferred to hang onto him as a 4th OF?

 

It's like blaming the twins for RA dickey or the Rangers for Chris Davis. Competitive teams can't afford to waste that many at bats on a guy who just doesn't produce

[ThePuck]

Few things:

 

They should sign pelfrey for two years now.

 

Anyone who blames the twins for Gomez is ridiculous, they gave him plenty of time and had a better in house option in Span, would u have preferred to hang onto him as a 4th OF?

 

It's like blaming the twins for RA dickey or the Rangers for Chris Davis. Competitive teams can't afford to waste that many at bats on a guy who just doesn't produce

 

No one on this thread was blaming the Twins for trading Gomez were they?

[nicksaviking]

No one on this thread was blaming the Twins for trading Gomez were they?

 

I am. That Gomez/Santana trade was terrible. Or were you guys talking about the other one.

[BabyJesusBuxton]

Personally, one of my favorite pitching stats to break down a start is Bill James' game score. Most stats have some sort of flaw in them but I find the game score is effective in showing the overall "quality" of a SP's performance with a numerical value. Therefore, possibly an average game score would be a good stat to compare pitchers over a season.

[snepp]

No one on this thread was blaming the Twins for trading Gomez were they?

 

This.

 

There are plenty of real disagreements going on already, let's not add any phantom arguments to the mix.

[ThePuck]

I am. That Gomez/Santana trade was terrible. Or were you guys talking about the other one.

 

They didn't trade Gomez twice did they, Nick? :-)

[Joe A. Preusser]

If a guy (pitcher A) goes 5 innings every single game and give up 2 ER every single game, and makes 30 starts: 150 IP, 3.60 ERA, 0 QS.

 

Now, if a player (pitcher goes 6 innings every single game and gives up 3 ER every single game, also 30 starts: 150 IP, 4.50 ERA, 30 QS.

 

Pitcher A is a potential All Star, Pitcher B is below league average.

 

QS is a completely worthless stat invented by agents in an effort to make their clients more money.

[Joe A. Preusser]

I am. That Gomez/Santana trade was terrible. Or were you guys talking about the other one.

 

I hated most of Bill Smith's moves, but even I can admit he had zero leverage on this one. Dealing from a position of weakness doesn't bode well for a positive return.

 

Now, if you want to talk about the Ramos trade, I will tear Smith a new one all over again for a week straight.

[Joe A. Preusser]

Few things:

 

They should sign pelfrey for two years now.

 

Anyone who blames the twins for Gomez is ridiculous, they gave him plenty of time and had a better in house option in Span, would u have preferred to hang onto him as a 4th OF?

 

It's like blaming the twins for RA dickey or the Rangers for Chris Davis. Competitive teams can't afford to waste that many at bats on a guy who just doesn't produce

 

My favorite will always be Big Papi. I still have friends who will talk about what a terrible move it was trading him, as if he wasn't injured and/or unproductive for most of his stint with us.

[Major League Ready]

I have to agree with you completely. Meyer and May were a good start but it would be hard to argue they fixed the situation. Worley did not look like part of the long-term solution even before he tanked. I was hoping Morneau, Willingham, and Doumit would have big years and return something at the deadline. As much as I hate to see Perkins go, he is now the only chance of bringing on another quality SP prospect.

 

I understand where you are coming from now.

[snepp]

Ortiz was never "unproductive" in his time with the Twins.

[ThePuck]

If a guy (pitcher A) goes 5 innings every single game and give up 2 ER every single game, and makes 30 starts: 150 IP, 3.60 ERA, 0 QS.

 

Now, if a player (pitcher goes 6 innings every single game and gives up 3 ER every single game, also 30 starts: 150 IP, 4.50 ERA, 30 QS.

 

Pitcher A is a potential All Star, Pitcher B is below league average.

 

QS is a completely worthless stat invented by agents in an effort to make their clients more money.

 

Well, first, that hypothetical situation never happens. Of the top 10 guys with QS, the highest ERA is 3.42. The next 10, only one is over 3.50. The next 10, only two over 4.00 (barely).

 

Second, it discounts that one of a starting pitcher's jobs is to not burn the bullpen.

 

Third, when these two phypothetical pitchers left, both pitchers gave their team a decent chance to win, but one was able to save the manager from having to have 30 extra innings tacked onto an already overworked bullpen.

[nicksaviking]

Ortiz was never "unproductive" in his time with the Twins.

 

Or traded. He had two more arbitration years left.

[Nick Nelson]

If a guy (pitcher A) goes 5 innings every single game and give up 2 ER every single game, and makes 30 starts: 150 IP, 3.60 ERA, 0 QS.

 

Now, if a player (pitcher goes 6 innings every single game and gives up 3 ER every single game, also 30 starts: 150 IP, 4.50 ERA, 30 QS.

 

Pitcher A is a potential All Star, Pitcher B is below league average.

 

QS is a completely worthless stat invented by agents in an effort to make their clients more money.

This is always amusing to me. People come up with some extreme scenario that would never actually occur in real life, then use it to prove that the statistic is "completely worthless."

[MichiganTwins]

I do not understand people's annoyance with the quality start stat. It is just a benchmark and an easy stat to judge someone's start. We can argue that it should have higher standards, but no one takes into account quality starts in any serious discussion i.e. when voting for the cy young.

[Brock Beauchamp]

This is always amusing to me. People come up with some extreme scenario that would never actually occur in real life, then use it to prove that the statistic is "completely worthless."

 

Funnily enough, I agree with this... I don't like the QS stat, as I've made abundantly clear to but clarify my point further, it does have merit when used in larger numbers. Given its arbitrary threshold and limited numbers of inputs, it's very flawed in the short term (because, as mentioned previously, a guy can go 8 IP, 4ER and not get credit). That situation can happen over the short-term and seriously skew the stat over a month-long period or even a half season because of the limited number of QS opportunities a pitcher receives in a season (we're talking about a metric that allows no more than ~33 outputs in a season).

 

But, like most metrics, the more inputs you throw at it, the more stable it becomes because the arbitrary outliers are offset by sheer numbers.

 

Honestly, it's not unlike the win stat when you get right down to it. Flawed and pretty crappy over the short term but over multiple years and/or a career, it starts to form a solid idea of who that guy was as a player.

 

Obviously, I also dislike the win statistic so there you go.

[ThePuck]

Funnily enough, I agree with this... I don't like the QS stat, as I've made abundantly clear to but clarify my point further, it does have merit when used in larger numbers. Given its arbitrary threshold and limited numbers of inputs, it's very flawed in the short term (because, as mentioned previously, a guy can go 8 IP, 4ER and not get credit). That situation can happen over the short-term and seriously skew the stat over a month-long period or even a half season because of the limited number of QS opportunities a pitcher receives in a season (we're talking about a metric that allows no more than ~33 outputs in a season).

 

But, like most metrics, the more inputs you throw at it, the more stable it becomes because the arbitrary outliers are offset by sheer numbers.

 

Honestly, it's not unlike the win stat when you get right down to it. Flawed and pretty crappy over the short term but over multiple years and/or a career, it starts to form a solid idea of who that guy was as a player.

 

Obviously, I also dislike the win statistic so there you go.

 

So now the QS start is so bad, it's reached the depths of how bad the win/loss stat is for pitchers?

[MichiganTwins]

Funnily enough, I agree with this... I don't like the QS stat, as I've made abundantly clear to but clarify my point further, it does have merit when used in larger numbers. Given its arbitrary threshold and limited numbers of inputs, it's very flawed in the short term (because, as mentioned previously, a guy can go 8 IP, 4ER and not get credit). That situation can happen over the short-term and seriously skew the stat over a month-long period or even a half season because of the limited number of QS opportunities a pitcher receives in a season (we're talking about a metric that allows no more than ~33 outputs in a season).

 

But, like most metrics, the more inputs you throw at it, the more stable it becomes because the arbitrary outliers are offset by sheer numbers.

 

Honestly, it's not unlike the win stat when you get right down to it. Flawed and pretty crappy over the short term but over multiple years and/or a career, it starts to form a solid idea of who that guy was as a player.

 

Obviously, I also dislike the win statistic so there you go.

Brock, I have to say, "I love your profile picture." It makes me laugh every time I see it. I really have nothing to say about what you posted, but I have been meaning to tell you that for a creepily long time.