Article: Free Agent Pitching: 20/20 Hindsight

[IdahoPilgrim]

I think over time some GMs prove to make more good decisions than than bad, and I think one reason for that could be they are better "artists" than other GMs, as well as or instead of being better "scientists."

 

Agree wholeheartedly with this. Sometimes people forget that life is not digital; sometimes the whole is greater than the sum of the parts.

 

I can take the sheet music and play the notes of a Beethoven Sonata, but that doesn't mean I can create music. There is an intangible that separates the true virtuoso performance from the mere repetition of notes in the correct sequence and tempo.

 

I think the same is true of being a GM in sports. You can't just plug numbers into an equation and get a correct answer. There is an indefinable quality that some GMs have and some don't (think David Kahn). In the past, TR seems to have had that quality. Whether he still does or not will be determined in part by how the rest of the season plays out

[PseudoSABR]

Way too early to judge most of last offsesaon's signings.

Welllllll........it was even earlier, when so many passed judgment on the Twins FA signings during the offseason.

[Oxtung]

So the premise of this article is that maybe, just maybe, Ryan knew some stuff before he signed those players, stuff that you can't find in FanGraphs. But, like any good experiment, you don't actually know your prediction is right until it is proven. The proof is in the performance of the pitchers.

 

No, the point of the article was to use 2013 stats to justify the 2012 signings. Which is a logical fallacy. There is nothing about what Ryan knew in the article. There is no, "We saw that Correia's breaking ball was pretty decent and if he just changes his grip slightly he will increase it's break by 3 inches."

 

Results should have no impact on whether a signing was a good decision. Look at this way, the future can never affect the past as we understand the concept of time today.

[Oxtung]

I would agree that in the strictest of senses, "what was known at the time the decision is made" is what's important in evaluating decision making. But I'm not at all sure "signing free agents in professional baseball" falls entirely into that circumstance.

 

GM'ing is part science, part art. Interpreting known data, and using it to predict the future, is part of it, and there's certainly some science to it.

 

But I think part of it also is the art of guessing how a player will fit into your roster, how his performance will translate into wins/losses, how he will be received by your customers and generate add'l revenue. The art of determining how much to offer a free agent, in terms of dollars and years, so that your offer has a chance to be the best, but not the best by so far as to be out of line. Even predicting the likelihood of health is as much art as science, IMO. We're also dealing with human beings here, so it's fair to say GMs must guess at human traits. "Will the guy continue to work hard if he gets financial security?" is one that comes to mind.

 

Bottom line, I think it's fair to evaluate GM decisions both at the time they are made, and--while keeping that in mind and tempering your conclusions--to also make some evaluations based on results.

 

I think over time some GMs prove to make more good decisions than than bad, and I think one reason for that could be they are better "artists" than other GMs, as well as or instead of being better "scientists."

 

But doesn't all of that "art" come before the decision is made to sign a player? Isn't that the important part? If you sign a player who has all those intangibles a GM is looking for and the player develops schizophrenia does that mean the GM made a bad decision? Of course not, there was no way he could know. What matters is what he did and should have known before an offer was extended.

 

My point isn't about stats vs. intangibles. My point is that the future can't affect the past. 2013 results, whether they're statistics or intangibles, have no bearing on 2012 signings.

[Guest USAFChief]

My point isn't about stats vs. intangibles. My point is that the future can't affect the past. 2013 results, whether they're statistics or intangibles, have no bearing on 2012 signings.

 

I'll quit with this: I disagree to some extent.

 

In some situations, it's valid to look at results, and in retrospect, go back and use that information to evaluate decisions. The future can't affect the past, but it can sometimes affect our understanding of it.

 

Drinking and driving is a bad decision no matter if you got home safely or not. No question, and results have no bearing on the decision.

 

Signing humans to play a game isn't quite so clearcut, and results might be useful later in evaluating those decisions. In my opinion.

[Oxtung]

I'll quit with this: I disagree to some extent.

 

In some situations, it's valid to look at results, and in retrospect, go back and use that information to evaluate decisions. The future can't affect the past, but it can sometimes affect our understanding of it.

 

Drinking and driving is a bad decision no matter if you got home safely or not. No question, and results have no bearing on the decision.

 

Signing humans to play a game isn't quite so clearcut, and results might be useful later in evaluating those decisions. In my opinion.

 

Can you give me an example? I can't come up with an instance where I have felt a signing was bad because of the results. I'm always disappointed because those results were predictable and the GM should have known before a player signed.

[jokin]

Welllllll........it was even earlier, when so many passed judgment on the Twins FA signings during the offseason.

 

And your problem with those noting that the Past is Probalistically Prologue...is what, exactly?

[Guest USAFChief]

Can you give me an example? I can't come up with an instance where I have felt a signing was bad because of the results. I'm always disappointed because those results were predictable and the GM should have known before a player signed.

In the specific case of the Twins, it's tough because they sign so few, and rarely venture from the bottom of the barrel, but I can think of two who probably worked out better than anyone had the right to expect:

 

Bob Tewksbury

Jim Thome

[Oxtung]

In the specific case of the Twins, it's tough because they sign so few, and rarely venture from the bottom of the barrel, but I can think of two who probably worked out better than anyone had the right to expect:

 

Bob Tewksbury

Jim Thome

 

Even in that example you say, "...worked out better than anyone had the right to expect...". So even here you are using the pre-signing expectations as your baseline.

[cmathewson]

No, the point of the article was to use 2013 stats to justify the 2012 signings. Which is a logical fallacy. There is nothing about what Ryan knew in the article. There is no, "We saw that Correia's breaking ball was pretty decent and if he just changes his grip slightly he will increase it's break by 3 inches."

 

Results should have no impact on whether a signing was a good decision. Look at this way, the future can never affect the past as we understand the concept of time today.

 

The problem with your stance is it means we are only allowed to evaluate GMs based on how they behaved relative to publicly available information such as FanGraphs. So you bias the whole conversation to that. As long as you stick to that information, and you can never be wrong, no matter how poorly the guys you favor do in the season.

 

Most of the things Ryan might use prior to signing a player he holds in confidence with his scouts. He doesn't want other teams using that information because it will tend to drive up the price of a player, or, worst case, entice a competitor to sign the player before he has a chance to make a counter-offer. After he signs the player, it does nobody any good to make the information public. It only prejudices the player against the organization.

 

The only things that have come out since the signings relate to durability and upside. Parker interviewed the Twins stats guy, who said he used Pitch F/X data combined with video to project a pitcher they signed. I.e. Carreia. They thought he had a chance to get more movement and be more effective than his numbers. The health thing was just about Correia not having health concerns and the other options having health concerns, which is why they passed on so many of them.

 

But every time I bring up stuff like that, you deny it because it is not available to you. As though the conversations on this site must be filtered through your lenses.

[The Wise One]

The problem with your stance is it means we are only allowed to evaluate GMs based on how they behaved relative to publicly available information such as FanGraphs. So you bias the whole conversation to that. As long as you stick to that information, and you can never be wrong, no matter how poorly the guys you favor do in the season.

 

Most of the things Ryan might use prior to signing a player he holds in confidence with his scouts. He doesn't want other teams using that information because it will tend to drive up the price of a player, or, worst case, entice a competitor to sign the player before he has a chance to make a counter-offer. After he signs the player, it does nobody any good to make the information public. It only prejudices the player against the organization.

 

The only things that have come out since the signings relate to durability and upside. Parker interviewed the Twins stats guy, who said he used Pitch F/X data combined with video to project a pitcher they signed. I.e. Carreia. They thought he had a chance to get more movement and be more effective than his numbers. The health thing was just about Correia not having health concerns and the other options having health concerns, which is why they passed on so many of them.

 

But every time I bring up stuff like that, you deny it because it is not available to you. As though the conversations on this site must be filtered through your lenses.

 

I love your comment. Count your blessings that you made it to the big league post level before you were on the receiving end of the argument used against a point they don't like but have no other way to disprove. "It is not available to me so therefore________________" When you ask for data, you well get rebuked for asking because you couldn't provide the information they wanted.

[Oxtung]

The problem with your stance is it means we are only allowed to evaluate GMs based on how they behaved relative to publicly available information such as FanGraphs. So you bias the whole conversation to that. As long as you stick to that information, and you can never be wrong, no matter how poorly the guys you favor do in the season.

 

Most of the things Ryan might use prior to signing a player he holds in confidence with his scouts. He doesn't want other teams using that information because it will tend to drive up the price of a player, or, worst case, entice a competitor to sign the player before he has a chance to make a counter-offer. After he signs the player, it does nobody any good to make the information public. It only prejudices the player against the organization.

Yup, that's the way the world works. It's not always fair or pretty. Even if someone doesn't like the outcome doesn't change the fact that logically 2013 results have no bearing on decisions made in 2012. They are two independent events.

 

 

The only things that have come out since the signings relate to durability and upside. Parker interviewed the Twins stats guy, who said he used Pitch F/X data combined with video to project a pitcher they signed. I.e. Carreia. They thought he had a chance to get more movement and be more effective than his numbers. The health thing was just about Correia not having health concerns and the other options having health concerns, which is why they passed on so many of them.

That's just not true. The talk was not about Correia in Parker's article and there was nothing about making the pitch better. Here is alink if you wish to brush up on it.

 

Twins Daily - Where are the Twins at with statistical analysis?

 

But every time I bring up stuff like that, you deny it because it is not available to you. As though the conversations on this site must be filtered through your lenses.

 

You're right. When you make things up I'm going to call you on it. Just like in this post.

[tobynotjason]

Worley is 86th in the league in SIERA among qualified pitchers to lead the Twins. (I assume you're interested in what's likely going forward, not what SSS BABIP and HR/FB numbers have posted.)

 

Correia's 88th. Pelfrey 104th (of 108).

 

Sanchez is 3rd. Jackson's 45th. You might have wanted Saunders, but "ew". (He's 103rd, btw.) Haren's 63rd. McCarthy's 51st. Dempster's 11th.

 

So per the best one-number statistic we have to measure how well a pitcher's done the things he can actually control and how well he's likely to do going forward: yeah, Terry Ryan picked poorly and the one guy who's done superficially "well" (if you ignore everything lately) hasn't posted any numbers that suggest he's anything but thoroughly mediocre. (His ERA actually outpaces his BABIP, by the way, which itself ranks THIRTY TWO places higher than his SIERA. Unsustainable much?)

[old nurse]

Worley is 86th in the league in SIERA among qualified pitchers to lead the Twins. (I assume you're interested in what's likely going forward, not what SSS BABIP and HR/FB numbers have posted.)

 

Correia's 88th. Pelfrey 104th (of 108).

 

Sanchez is 3rd. Jackson's 45th. You might have wanted Saunders, but "ew". (He's 103rd, btw.) Haren's 63rd. McCarthy's 51st. Dempster's 11th.

 

So per the best one-number statistic we have to measure how well a pitcher's done the things he can actually control and how well he's likely to do going forward: yeah, Terry Ryan picked poorly and the one guy who's done superficially "well" (if you ignore everything lately) hasn't posted any numbers that suggest he's anything but thoroughly mediocre. (His ERA actually outpaces his BABIP, by the way, which itself ranks THIRTY TWO places higher than his SIERA. Unsustainable much?)

 

Using any of the metrics is fine if you understand the limitations of them. You sign a player to play in your baseball park 81 games a year. What that player does should suit your park. All of the metrics attempt to get rid of that factor. Any metric is a look backwards wether or not it is sustainable no metric measures.

[cmathewson]

Yup, that's the way the world works. It's not always fair or pretty. Even if someone doesn't like the outcome doesn't change the fact that logically 2013 results have no bearing on decisions made in 2012. They are two independent events.

 

 

A scientist makes a prediction based on a theory. He then tests the prediction in the lab. The test works as predicted (within a margin of error). Therefore, the theory has some validity.

 

A GM makes a decision based on a theory and some facts. He then tests the decision on the field in the form of giving the player an opportunity. The prediction holds, as the player was successful. Therefore, the decision was right.

 

A scientist makes a prediction based on a theory. He then tests the prediction in the lab. Unfortunately, the test fails to show what he expects. This invalidates the theory.

 

A GM signs a player based on a theory and some facts. He then tests the theory on the field in the form of giving the player an opportunity. Unfortunately, the test fails as the player fails to perform as well as the GM's theory predicted. Therefore, the decision was wrong.

 

This is the kind of reasoning we use in everyday life. I fail to see that it is invalid or it's "just not how the world works."

 

I do web marketing. We build experiences and test them in multivariate settings. The one that works the best gets adopted. We then continue to test that experience and iteratively improve it.

 

The whole premise is that you can't really know all that much about how an experience will work in advance until you test it. When you test it, you can say, "I guess option B was best."

 

This is a lot like what GMs do. They have limited data. Humans are not robots. So they take their best shot. If it doesn't work out, they say, "Well, I guess I was wrong." If everybody knew the right answer before trying it, it wouldn't be a game. The guys who play the game the best win. But even the best GMs are right a little better than half the time, which is why they hedge their bets by getting depth.

 

Oh, and the stats guy never said it was Correia. He said it was a player they were considering acquiring. It seems likely that if he did it for one player, he did it for every player they acquired.

[Mike Sixel]

Just because a decision works, does not mean the decision was good. Lots of lucky things happen that lead to a good outcome, that does not mean the actual decision was a good decision. For example, if you drive drunk and don't kill someone, that does not mean you made a good decision. It happened to work, but that does not make it a good decision. And, more importantly, just becaue it worked once does not mean it will work again.

 

Look, Ryan bought cheap pitchers for whatever reason. He passed, as he has every year he has been GM, on more expensive options. The players he signed (and traded for) have not been effective this year, as a whole. They seem to be getting worse as the year goes on. Can we all just agree on a couple of things:

 

1. They have utterly sucked at drafting and developing pitchers for years.

2. They have not spent money to make the team better, for whatever reason.

3. This staff is awful, again this year, and they continue to stick with the staff.

4. Something needs to change.

[snepp]

http://www.baseballprospectus.com/u/images/Process%281%29.png

[jharaldson]

maybe free agency is such a crap shoot that you never know what you're going to get

 

I thought I would do some math to prove or disprove this point. To start with, I took the top 25 contracts from the ESPN Free Agent Tracker and ordered them from largest to smallest:

 

2012 MLB Baseball Free Agent Tracker - Major League Baseball - ESPN

 

I then went to FanGraphs and created a custom report for these players and ranked them by ERA:

 

Major League Leaderboards » 2013 » Pitchers » Custom Statistics | FanGraphs Baseball

 

I put both of these together in a spreadsheet:

 

[TABLE=width: 303]

Contract

Rank

ERA

Rank

Name

ERA

1

1

Zack Greinke

1.62

25

2

Francisco Liriano

1.69

10

3

Hisashi Iwakuma

1.84

2

4

Anibal Sanchez

2.05

9

5

Hiroki Kuroda

2.31

15

6

Scott Feldman

2.53

6

7

Jeremy Guthrie

2.82

13

8

Kevin Correia

3.35

4

9

Hyun-Jin Ryu

3.4

20

10

Jason Marquis

3.49

5

11

Ryan Dempster

3.75

12

12

Andy Pettitte

3.83

18

13

Roberto Hernandez

4.43

19

14

Bartolo Colon

4.56

11

15

Dan Haren

4.76

7

16

Brandon McCarthy

5.63

23

17

Jeff Francis

6

3

18

Edwin Jackson

6.02

22

19

John Lannan

6.14

8

20

Joe Blanton

6.46

17

21

Mike Pelfrey

6.57

14

22

Brett Myers

8.02

16

23

Scott Baker

[/TD]

20

24

Jeff Karstens

24

25

Jair Jurrjens

[TD]

[/TABLE]

 

One interesting thing is 7 of the top 11 guys in ERA are from the top 11 contracts signed in the offseason. When I ran the correlation numbers it came up as a .39 correlation which is not high but shows there is a correlation between contract size and performance.

 

In short, free agent pitchers are not a crap shoot. There is a moderate correlation between the amount you spend and the performance you get.

[cmathewson]

http://www.baseballprospectus.com/u/images/Process%281%29.png

 

Sure. So you can talk about whether they had a good process or not. Problem is, much of that is proprietary to the Twins. So from an outsiders' perspective, it might seem like a bad process. But that doesn't mean it is.

[Mike Sixel]

Great post jharaldson

[Siehbiscuit]

I think the underlying problem is is that most of us do not like the team's philosophy on pitching, period. I don't think spending in FA is always the answer. How many times have the Oakland A's turned out great pitching prospects!??? Thy may not stay in an A's uniform past 6 years (when FA kicks in) but they flip it for more young pitching and they always seem to have it. The A's are always acquiring pitching via trades and the Twins (up until this off season) seldom think like this. Its dumpster diving every offseason and hoping something sticks. The soft tossing, pitch-to-contact guys are cheap for a reason. They get hit and too often hit hard. IMO, the Twins organization needs to seriously change both the type of pitcher it is targeting and also it acquiring philosophy. Enough with the soft-tossers!

[70charger]

jharaldson - a .39 correlation!? That's way lower than I would have guessed. Thanks for running the numbers.

[drjim]

Great post jharaldson

 

+1. I would add a couple thoughts.

 

1. Contract rank might not be the best indicator of talent as this will vary based upon age of pitcher or certain intangibles (ie desire to play for contender or certain geographic location). Perhaps the rankings of MLBTR?

 

2. Despite Seth's comments, I would think of FA as a relative crapshoot. Few, if any, would argue that Greinke was #1, Sanchez #2, and Dempster #3 last offseason. The crapshoot starts in the next tier and I would imagine if you knock out the clear top performers that the correlation would be that much smaller.

[Mike Sixel]

But is FA more of a crapshoot than drafting and developing players? I'd think that success (or heck, reaching) in the majors indicates a higher likelihood of success than being a 3rd round pick in the draft. That's the issue for some of us, ignoring a channel, for a channel that is less likely to work (and that the team has utterly failed out lately).

[drjim]

1. They have utterly sucked at drafting and developing pitchers for years.

2. They have not spent money to make the team better, for whatever reason.

3. This staff is awful, again this year, and they continue to stick with the staff.

4. Something needs to change.

 

I would accept this list with some slight modifications.

 

On point 1 I would emphasize a bad run instead of sucking. The front office has developed talent before and probably will again.

 

On point 2 I would cut out "for some reason". It adds nothing to the point and I personally think there are several good reasons why they didn't spend this offseason.

[drjim]

But is FA more of a crapshoot than drafting and developing players? I'd think that success (or heck, reaching) in the majors indicates a higher likelihood of success than being a 3rd round pick in the draft. That's the issue for some of us, ignoring a channel, for a channel that is less likely to work (and that the team has utterly failed out lately).

 

They don't ignore it, they just signed two free agents this offseason.

[jharaldson]

Correlation is measured between -1 and 1. Anything above 0 is considered correlation and I generally think of it this way:

 

.66-1 = High correlation

.33-.66 = Moderate Correlation

0-.33 = Low correlation

 

So contract size is moderate to low correlation but still well above a crap shoot which I interpreted as no correlation.

[70charger]

But is FA more of a crapshoot than drafting and developing players? I'd think that success (or heck, reaching) in the majors indicates a higher likelihood of success than being a 3rd round pick in the draft. That's the issue for some of us, ignoring a channel, for a channel that is less likely to work (and that the team has utterly failed out lately).

 

The relative downside is not even remotely close to the same. If you sign a guy like Willingham for example, which by all means was NOT an expensive free agent contract, you're on the hook for $21 million. If he implodes, that's serious money. If you sign a guy in the third round and he never makes it out of AA, you've paid him maybe $300,000 over the life of his contract and given the minor league fans something to watch.

 

I think it's terribly obvious that signing good players in free agency is far more likely to get you a good player than signing a kid in the draft, waiting five years, and crossing your fingers. But when one fails, future teams suffer. When the other fails, people yawn.

[Mike Sixel]

Unless it fails year, after year, after year, and you have zero, not one, starter in your rotation you drafted, and the ones you had the year before were bad (not Baker, he was good), and the same as the year before.......the fear of downside risk shown by the fans of this team astounds me, it really does. You are not wrong to have that fear if you want, people have different appetities, but it does amaze me.

[cmathewson]

Unless it fails year, after year, after year, and you have zero, not one, starter in your rotation you drafted, and the ones you had the year before were bad (not Baker, he was good), and the same as the year before.......the fear of downside risk shown by the fans of this team astounds me, it really does. You are not wrong to have that fear if you want, people have different appetities, but it does amaze me.

 

Um, 2010 says Hi.