Article: Twins To Sign RHP Kevin Correia

In any case, the QS% argument is a little disingenuous because you're looking at 1 year of data in a 10 year career, and it happens to be Correia's career low in BABIP year, and career high in GB% year. Over the last 10 years, Correia's QS% is 46%, which is 4% worse than MLB average (probably 6-7% worse than the NL average). If you can prove Correia did something that should allow him to sustain a lower BABIP and higher GB rate, then you would have a point. All I see is an anomalous fluctuation in batted balls.

edit: and obviously that's before accounting for the fact that he's played in the NL, and in some pretty cavernous home ballparks for most his career.

Quote:

---

Originally Posted by whydidnt

I just saw on another blog that Baseball Reference has a cool little tool that allows you to adjusts a players stats for a year to other teams/league. Here are the results when I put Correia's 2012 line, his BEST season in 3 years, into the 2012 Twins:

2012

31

7

11

.389

4.91

163

186

98

89

21

49

85

3

1.442

32

Looks pretty ugly and not worth investing 2 years in the guy. I know these just calculations/estimates, but they are based upon sound mathematics anyway.

---

Did you bother to check that the past two seasons and see that that line beats Every pitcher we've run out there for the past two seasons for more than 100 innings apart from Pavano and Baker (22 starts) in 2011 and Diamond in 2012. It might be uninspiring, but Correia is an improvement.

Quote:

---

Originally Posted by Willihammer

In any case, the QS% argument is a little disingenuous because you're looking at 1 year of data in a 10 year career, and it happens to be Correia's career low in BABIP year, and career high in GB% year. Over the last 10 years, Correia's QS% is 46%, which is 4% worse than MLB average (probably 6-7% worse than the NL average). If you can prove Correia did something that should allow him to sustain a lower BABIP and higher GB rate, then you would have a point. All I see is an anomalous fluctuation in batted balls.

edit: and obviously that's before accounting for the fact that he's played in the NL, and in some pretty cavernous home ballparks for most his career.

---

I agree... And I cant prove that Correia did something different in 2012 and I'm not trying to. Fluctuation is consistent in all players from year to year... Month to month... Week to week...

I've only been backed into this location because I said that Correia wasn't that bad last year and here is why I think so and as I looked at his numbers from last year. They looked good and better than I thought.

And I only pointed out that his numbers were decent last year because the comments from posters have gone over the top.

Will he regress... Will he improve... Will he stay the same... I don't know... Bill James Doesn't know... Terry Ryan doesn't know... No poster on this site knows.

Correia was never my first choice but he is being stoned to death for no reason and I think QS is a decent stat with some flaws. That's what Im saying.

Quote:

---

Originally Posted by LoganJones

Did you bother to check that the past two seasons and see that that line beats Every pitcher we've run out there for the past two seasons for more than 100 innings apart from Pavano and Baker (22 starts) in 2011 and Diamond in 2012. It might be uninspiring, but Correia is an improvement.

---

I didn't, but not sure what difference it makes. This points to the divide among those that think it's an ok or good signing vs. those of us that don't like it. If the Twins lose 5-4 instead of 6-4 it's still a loss as far as I'm concerned. They should be trying to find guys that are going to help them win, not just help them lose by less. Correia is simply going to be a guy that might help them lose by less than what they have the last few years. Not satisfying to me. You can't build a quality rotation around guys that are 4-th 5th starters around the rest of the league and this points to trying to do that.

Riverbrian - Jackson checks in at a 4.40 ERA and a 1.288 WHIP on the baseball-reference tool for 2012. The ERA is higher than you'd like to see, but the WHIP is pretty good for a starter.

Quote:

---

Originally Posted by Riverbrian

Slippery slope when you talk about more flawed than others. 6IP 3ER or less in each start... Did he get there or didn't he.

---

But, again, what does that mean? And why isn't 8 IP, 4 ER counted? The entire stat is based off a flawed stat (ERA) and an arbitrary number as a cutoff (IP). It makes no attempt whatsoever to factor in anything past earned runs with a random inning threshold thrown in for good measure. To boot, it unnecessarily constrains sample size by judging only an entire start as one entity instead of breaking it into small pieces, which makes it more prone to fluctuation. It's bad math any way you look at it and if you presented a similar metric to any statistics professor, he/she would laugh you out of the room.

Quote:

---

Originally Posted by Riverbrian

Correia was never my first choice but he is being stoned to death for no reason and I think QS is a decent stat with some flaws. That's what Im saying.

---

I think most people are stoning Jr.

Quote:

---

Originally Posted by Brock Beauchamp

Quote:

---

Originally Posted by Riverbrian

Slippery slope when you talk about more flawed than others. 6IP 3ER or less in each start... Did he get there or didn't he.

---

But, again, what does that mean? And why isn't 8 IP, 4 ER counted? The entire stat is based off a flawed stat (ERA) and an arbitrary number as a cutoff (IP). It makes no attempt whatsoever to factor in anything past earned runs with a random inning threshold thrown in for good measure. To boot, it unnecessarily constrains sample size by judging only an entire start as one entity instead of breaking it into small pieces, which makes it more prone to fluctuation. It's bad math any way you look at it and if you presented a similar metric to any statistics professor, he/she would laugh you out of the room.

---

A. The stat you are looking for doesn't exist. What stat will determine if a pitcher did good the day he started or didn't.
B. One Start is One Entity and the next one is another.
C. It's Basic... It's not advanced and most stats even advanced ones are based on the same basic information to the point of being superfluous.
D. Any statistic professor that laughs you out of the room for use of it. Would also laugh you out of the room for using 1 plus 1 equals 2 and I'm hopeful that any statistic professor would realize that the components used in the stat are the same components used to determine the majority of stats used in baseball. So if a professor laughs at you... Drop the class and find a more reasonable professor.

To make the Quality start stat a better indication of what a quality start actually is, it should become a function of ER/IP. Granted this must be a whole number stat, so I'd say a QS should be separate brackets beginning at 5 IP and 2 ER, 6-7 IP and 3 ER, 8-9 IP and 4 ER.

I really don't have a problem with the current stat, but recognize it for what it is -- a very MACRO snapshot of starting pitching success. It does that okay.

Quote:

---

Originally Posted by whydidnt

Riverbrian - Jackson checks in at a 4.40 ERA and a 1.288 WHIP on the baseball-reference tool for 2012. The ERA is higher than you'd like to see, but the WHIP is pretty good for a starter.

---

Thanks... Do you have a link? It looks like something fun to play with, I'd also love to see the methodology for better understanding.

Quote:

---

Originally Posted by Riverbrian

Any statistic professor that laughs you out of the room for use of it. Would also laugh you out of the room for using 1 plus 1 equals 2 and I'm hopeful that any statistic professor would realize that the components used in the stat are the same components used to determine the majority of stats used in baseball. So if a professor laughs at you... Drop the class and find a more reasonable professor.

---

No, the professor would laugh you out of the room because you're using an A+B=C stat that doesn't work 100% of the time.

The Quality Start stat essentially says that 1+2=3 but that 2+4/=/6.

Quote:

---

Originally Posted by Brock Beauchamp

Quote:

---

Originally Posted by Riverbrian

Any statistic professor that laughs you out of the room for use of it. Would also laugh you out of the room for using 1 plus 1 equals 2 and I'm hopeful that any statistic professor would realize that the components used in the stat are the same components used to determine the majority of stats used in baseball. So if a professor laughs at you... Drop the class and find a more reasonable professor.

---

No, the professor would laugh you out of the room because you're using an A+B=C stat that doesn't work 100% of the time.

The Quality Start stat essentially says that 1+2=3 but that 2+4/=/6.

---

lol... Brock... You are going to make me research the margin of error!!! It's my day off... I shouldn't have to look at stats that hard today. I do that when I'm working. lol.

Quote:

---

Originally Posted by Riverbrian

Quote:

---

Originally Posted by Brock Beauchamp

Quote:

---

Originally Posted by Riverbrian

Any statistic professor that laughs you out of the room for use of it. Would also laugh you out of the room for using 1 plus 1 equals 2 and I'm hopeful that any statistic professor would realize that the components used in the stat are the same components used to determine the majority of stats used in baseball. So if a professor laughs at you... Drop the class and find a more reasonable professor.

---

No, the professor would laugh you out of the room because you're using an A+B=C stat that doesn't work 100% of the time.

The Quality Start stat essentially says that 1+2=3 but that 2+4/=/6.

---

lol... Brock... You are going to make me research the margin of error!!! It's my day off... I shouldn't have to look at stats that hard today. I do that when I'm working. lol.

---

Margin of error is a different argument entirely. Quality Start uses a minimum threshold (6 IP, 3 ER), which is fine. I may not agree with it but for the sake of this argument, I'll ignore it. Where the stat goes horribly wrong is that it uses only half of the same implementation for its maximum threshold. IP continues to scale but ER does not.

A guy can pitch a six inning game, give up 20 hits, but manage to keep the damage to three runs so he gets the quality start.

Another guy can pitch a nine inning game, give up two hits, a walk, and a homer, get the win, and not be awarded a quality start. That's a bad stat.

As has already been stated, QS is acceptable at the macro level, either team as a whole, or player career. But for a single season player? Yuck, stay away, for every reason Brock as been continually stating.

Quote:

---

Originally Posted by Brock Beauchamp

Quote:

---

Originally Posted by Riverbrian

Quote:

---

Originally Posted by Brock Beauchamp

Quote:

---

Originally Posted by Riverbrian

Any statistic professor that laughs you out of the room for use of it. Would also laugh you out of the room for using 1 plus 1 equals 2 and I'm hopeful that any statistic professor would realize that the components used in the stat are the same components used to determine the majority of stats used in baseball. So if a professor laughs at you... Drop the class and find a more reasonable professor.

---

No, the professor would laugh you out of the room because you're using an A+B=C stat that doesn't work 100% of the time.

The Quality Start stat essentially says that 1+2=3 but that 2+4/=/6.

---

lol... Brock... You are going to make me research the margin of error!!! It's my day off... I shouldn't have to look at stats that hard today. I do that when I'm working. lol.

---

Margin of error is a different argument entirely. Quality Start uses a minimum threshold (6 IP, 3 ER), which is fine. I may not agree with it but for the sake of this argument, I'll ignore it. Where the stat goes horribly wrong is that it uses only half of the same implementation for its maximum threshold. IP continues to scale but ER does not.

A guy can pitch a six inning game, give up 20 hits, but manage to keep the damage to three runs so he gets the quality start.

Another guy can pitch a nine inning game, give up two hits, a walk, and a homer, get the win, and not be awarded a quality start. That's a bad stat.

---

You are absolutely right... That's why the stat is not perfect.

But can we agree... That the 20 hit thing is going to be very limited in occurance to almost nil because most managers are going to pull the guy with that much traffic on the bases. Therefore the chances of reaching 6 IP in that scenerio is highly improbable with Pitch Counts and Bullpen usage.

And if a pitcher does make it through 6 innings under that scenerio. 3 runs crossed the plate is still 3 runs across the plate regardless and the end result is his team has been given a chance to win. Therefore it's a quality start.

If you are looking for degrees of quality. That stat doesn't exist and you are better off looking at other stats and you are saying that is exactly what you do and thats ok with me... whatever floats your boat.

With Quality starts... I'm looking at one thing and one thing only... Does the pitcher give your team a chance to win... QS does that just fine for me. I love advanced metrics but I ain't afraid to hang with the basics and I do so all the time because the basics is the portal to the advanced. I think a lot of people skip that basic step.

Yes the 2 hit... homer... walk scenerio is also a good example. This is exactly where my main problem with the stat lies..however... I think if we looked thru start after start game logs... We will find that this isn't all that prevalent either.

We are basically discussing the potential for corruption and I don't deny that potential... Its onvious... However... We don't have actual proof of corruption despite the potential for it. That would take a little work to prove or not prove... I contend that it will be a fairly equal spread if I or we or anyone put the work into it.

If the skewing is for the most part universal. I will place my faith in the stat with the understanding that there is potential for error in the result in certain cases.

Quote:

---

Originally Posted by Riverbrian

If you are looking for degrees of quality. That stat doesn't exist and you are better off looking at other stats and you are saying that is exactly what you do and thats ok with me... whatever floats your boat.

---

The thing is, that stat does exist.

Game score - Wikipedia, the free encyclopedia

Quote:

---

Originally Posted by snepp

As has already been stated, QS is acceptable at the macro level, either team as a whole, or player career. But for a single season player? Yuck, stay away, for every reason Brock as been continually stating.

---

You would have to stay away from almost every single season stat. Something that I actually agree with to a point because things fluctuate.

Quote:

---

Originally Posted by Riverbrian

Quote:

---

Originally Posted by snepp

As has already been stated, QS is acceptable at the macro level, either team as a whole, or player career. But for a single season player? Yuck, stay away, for every reason Brock as been continually stating.

---

You would have to stay away from almost every single season stat. Something that I actually agree with to a point because things fluctuate.

---

It's all about quantity of data, Brian. On top of being a flawed stat based on flawed stats using arbitrary cutoff levels, Quality Start also uses a very small data size (broken into 27-33 pieces for a full season). Most other stats use hundreds of data points to reach their full season total.

The smaller the quantity of data, the more each failure in recording impacts the statistic. One missed call by Quality Start impacts the player by 3-4%. One missed Earned Run impacts the player by .75-1.25%. Each missed Strikeout impacts the player by .33-1.0%. Etc, etc.

Quote:

---

Originally Posted by Riverbrian

Quote:

---

Originally Posted by YourHouseIsMyHouse

Quote:

---

Originally Posted by Riverbrian

Quote:

---

Originally Posted by Brock Beauchamp

Is Kevin Correia actually Jason Marquis? I'm looking through their stat lines and they're basically the exact same guy.

What on earth is Ryan thinking?

---

Blackburn gave up 81 runs in 2012... Correia gave up 80 and Marquis gave up 74 in 127.2 innings between Minnesota and San Diego.

Blackburn pitched 98.2 innings and Correia threw 171 innings. If posters want to cry Blackburn with this signing... Please don't ignore the 72.1 innings of shutout ball in comparison between the two... In Comparison... Thats quite the difference.

---

Good for Correia that he wasn't capable of the other worldly and historically colossal decrepitude that Nick Blackburn is. For a metaphorical visual of that comparison, a 4 looks a lot sexier when standing next to a 1....and yet it's still a 4.

---

Yeah... That's kinda my point... But we have posters here who are calling this signing Blackburn esque. I was pointing out that... that is embellishment at a fairly large degree.

---

Fair enough. But I see their point in a way. Even Blackburn might have been a 4 prior to 2012 where he gained a lot of weight and refused to wear any makeup. It's possible that Correia falls apart in that way he and Marquis did at Target Field.

Quote:

---

Originally Posted by whydidnt

I just saw on another blog that Baseball Reference has a cool little tool that allows you to adjusts a players stats for a year to other teams/league. Here are the results when I put Correia's 2012 line, his BEST season in 3 years, into the 2012 Twins:

2012

31

7

11

.389

4.91

163

186

98

89

21

49

85

3

1.442

32

Looks pretty ugly and not worth investing 2 years in the guy. I know these just calculations/estimates, but they are based upon sound mathematics anyway.

---

Do you have a link for this? That seems like it would be quite addicting.