from the GArvey v Cey thread

[Jackaroo Dave]

This thread started in the HOF forum as an outgrowth of a discussion of the merits of Steve Garvey v Ron Cey. It has become a thread of its own, based on the discussion of sabermetric versus traditional methods, and is being moved to the stats forum if its participants wish to continue it.

That's the silly way my brain works; of course a HR is worth an RBI and a R because it's 100% of a run as opposed to 50% shared by each of two players; yet I've always subtracted HR totals from run production in my head because I think of it as the number of team runs an individual contributed to, like how a hockey player will get a point credit for a goal or an assist, but doesn't get two points if the goal is unassisted.

The additional silly thing is, it's Cey who should be penalized for having more HR, not Garvey. Oh well, it was an interesting theory, anyway. Either way, Garvey contributed slightly more to team run totals, whether you subtract HR or not. I stand by my belief that that can be significant. How significant I leave to individual judgement; as I said I still pick Cey.

As a traditionalist, I was very excited when the Runs Produced stat came along. At that time, Tony Oliva, one of my favorites, ranked very high, because he had a lot of inside-the-park power.

Tom Tango maintains that it is a better measure than R + RBI, but I have not followed his reasoning.

[Jackaroo Dave]

Very well said.

A lot of guys who were definite Hall of Famers when they played now seem to be getting the short shrift in the modern SABRmetrics area of selectively dissecting stats, that in many cases have very little relevant actual worth. (Sorry, everyone, I'm one of those "unenlightened" old guys that still places more value in what I saw transpire on the field than on the value of WAR+) Just because a new stat has been developed (OPS+) doesn't mean I am willing to accept it as being a valid way of measuring players. Are we really to assume that Garvey was some sort of Svengali that had all the press and fans under his spell in thinking he was a great player when in all actuality he was nowhere near being a Hall of Famer? 200 hits a season are merely "counting" stats, but meaningless in the larger picture? Because he never led the league in assists his putout totals should be ignored along with his fielding pct. because his range factor+ actually made him a poor fielder? I wish I could afford to buy everyone, if not time machines, than at least some boxed DVD sets.

Unlike the traditional stats, many contemporary ones, e.g. WAR and its components, linear weights, specifically claim to estimate a player's value via a chain of explicit reasoning at times expressed mathematically. If you think there's a mistake somewhere, it's easy: Everything is spelled out letter by letter and number by number from first premises to every conclusion that bothers you. All you have to do is peruse the literature and find the mistakes.

Also you'll find discussed problems of estimation, reliability, sample size, exceptions, and controversies about the value and demerits of various stats. If you look on the baseball literature forum, you'll find a thread on getting started in sabremetrics that refers you to an abundance of cheap or free resources that are geared and welcoming to new students. Aside from a little probability, there's no mathematics more difficult than, say, calculating ERA.

Dropping $4 for Lee Panan's Beyond Batting average and spending an evening or two going through it will make you much better equipped to actually argue that Garvey is better than Cey, rather than reassert it in various modes.

If your way of dealing with the arguments put forward here is just to dismiss them with the wave of a
hand, not even bothering to understand them, much less rebut them, then I have to conclude that you shy from the confrontation either because you anticipate an unwelcome result or you doubt your capacity to master the subject. That's OK; nobody's good at everything. You should hear me try to sing.

If, rather, you simply feel that it's a lot of work to learn something of no value, like mastering Ptolemaic astronomy, for example, I can understand that. But you can't expect people who are taking the trouble to understand the field to treat your judgments as those of a peer.

Seriously, why don't you guys take the trouble to work up some genuine arguments against the premises expressed or the steps of reasoning leading to the conclusions you mock? I've been here two years, and I haven't yet seen an argument against the sabremetric enterprise that gave any indication that the author had even read a book about it.

[The Commissioner]

Unlike the traditional stats, many contemporary ones, e.g. WAR and its components, linear weights, specifically claim to estimate a player's value via a chain of explicit reasoning at times expressed mathematically. If you think there's a mistake somewhere, it's easy: Everything is spelled out letter by letter and number by number from first premises to every conclusion that bothers you. All you have to do is peruse the literature and find the mistakes.

Also you'll find discussed problems of estimation, reliability, sample size, exceptions, and controversies about the value and demerits of various stats. If you look on the baseball literature forum, you'll find a thread on getting started in sabremetrics that refers you to an abundance of cheap or free resources that are geared and welcoming to new students. Aside from a little probability, there's no mathematics more difficult than, say, calculating ERA.

Dropping $4 for Lee Panan's Beyond Batting average and spending an evening or two going through it will make you much better equipped to actually argue that Garvey is better than Cey, rather than reassert it in various modes.

If your way of dealing with the arguments put forward here is just to dismiss them with the wave of a
hand, not even bothering to understand them, much less rebut them, then I have to conclude that you shy from the confrontation either because you anticipate an unwelcome result or you doubt your capacity to master the subject. That's OK; nobody's good at everything. You should hear me try to sing.

If, rather, you simply feel that it's a lot of work to learn something of no value, like mastering Ptolemaic astronomy, for example, I can understand that. But you can't expect people who are taking the trouble to understand the field to treat your judgments as those of a peer.

Seriously, why don't you guys take the trouble to work up some genuine arguments against the premises expressed or the steps of reasoning leading to the conclusions you mock? I've been here two years, and I haven't yet seen an argument against the sabremetric enterprise that gave any indication that the author had even read a book about it.

Wow, what a truly offensive post. Seriously. You are assuming that because we do not agree with many of the SABRmetrics that we do not understand them. Many of us do understand them and that is why we see the inherent flaws in many of them. I have read how these statistics are derived and that is why I find flaws, many flaws, in the supposed "chain of explicit reasoning at times expressed mathematically". I believe it is you that are dismissing with the wave of a hand anyone that doesn't agree by throwing us all into the category of unenlightened and/or lazy fools. As for not considering me a peer, thank you for the compliment. I do not, however, feel the need to prove to you, nor anyone else here, my credentials regarding a knowledge of the game. If you insist, though, I would welcome a public debate in as large of a forum as possible on any given baseball related topic of your choice. You find the TV and or radio time and let's make it happen.

You are expecting us to refute books worth of "reasoning" within the confines of an online forum. It would take at least a book's worth of arguments to lay out and dissect all the flaws. I'm not ruling out the possibility of one or many of us doing so in the future, but until that opportunity arises to write and publish one you will have to put up with the uninformed and intellectually lazy rantings of such infidels and troglodytes as myself. One day when I am capable of sitting down to read a whole big book 'bout baseball, maybe I can work hard enough to understand what all of you with a superior understanding of the game are saying.

In the meantime, JR, do you wanna go break things or look at pop-up books like others in our peer group do?

[JR Hart]

Unlike the traditional stats, many contemporary ones, e.g. WAR and its components, linear weights, specifically claim to estimate a player's value via a chain of explicit reasoning at times expressed mathematically. If you think there's a mistake somewhere, it's easy: Everything is spelled out letter by letter and number by number from first premises to every conclusion that bothers you. All you have to do is peruse the literature and find the mistakes.

Also you'll find discussed problems of estimation, reliability, sample size, exceptions, and controversies about the value and demerits of various stats. If you look on the baseball literature forum, you'll find a thread on getting started in sabremetrics that refers you to an abundance of cheap or free resources that are geared and welcoming to new students. Aside from a little probability, there's no mathematics more difficult than, say, calculating ERA.

Dropping $4 for Lee Panan's Beyond Batting average and spending an evening or two going through it will make you much better equipped to actually argue that Garvey is better than Cey, rather than reassert it in various modes.

If your way of dealing with the arguments put forward here is just to dismiss them with the wave of a
hand, not even bothering to understand them, much less rebut them, then I have to conclude that you shy from the confrontation either because you anticipate an unwelcome result or you doubt your capacity to master the subject. That's OK; nobody's good at everything. You should hear me try to sing.

If, rather, you simply feel that it's a lot of work to learn something of no value, like mastering Ptolemaic astronomy, for example, I can understand that. But you can't expect people who are taking the trouble to understand the field to treat your judgments as those of a peer.

Seriously, why don't you guys take the trouble to work up some genuine arguments against the premises expressed or the steps of reasoning leading to the conclusions you mock? I've been here two years, and I haven't yet seen an argument against the sabremetric enterprise that gave any indication that the author had even read a book about it.

I don’t like adding OBP and slugging together. They aren’t compatible stats. If a player walks a lot, his ABs will lower, and thus he has a better chance to raise his slugging. There’s too much borrowing from Peter to give to Paul. I don’t know why sabers hate TBs so much, but they show durability also. If we must throw in walks, I would be for a stat like TB + (BB x .60). That would factor in OBP and also show durability.

What don’t I like about WAR? Virtually everything! I don’t like fielding runs and think that they are a guessing game at best. Do we really know Jack Glasscock’s range in 1886? I don’t like positional adjustments and feel that values assigned for positions are highly hypothetical. I don’t like that positional adjustments are used in batting WAR.

Having said that, I think that using a single value to rate players, is just lazy. I see it all the time here, and when I point it out, I get a wave of protest. But it happens here all the time. I think that we have to look at everything, awards, all star games, post season, durability, team success, peak, and career stats. AND I don’t want a statistic thast factors all that in, because it will enivitably make assumptions that can’t be proven.

Look, I know that baseball lends itself to all kinds of mathematical analysis. But Joe Carter was a player, not a math problem. His manager decided that the team could sacrifice outs, in order for him to drive the ball. It worked.

[filihok]

I don’t like adding OBP and slugging together. They aren’t compatible stats. If a player walks a lot, his ABs will lower, and thus he has a better chance to raise his slugging. There’s too much borrowing from Peter to give to Paul.

It's not a great stat. wRC+ is much better.

I don’t know why sabers hate TBs so much, but they show durability also. If we must throw in walks, I would be for a stat like TB + (BB x .60). That would factor in OBP and also show durability.

I don't know many swords that are able to discuss baseball...
There's nothing wrong with total bases. They are what they are.
You should like wOBA (and wRC and wRC+) since it's similar to total bases, just making some needed adjustments.

What don’t I like about WAR? Virtually everything! I don’t like fielding runs and think that they are a guessing game at best. Do we really know Jack Glasscock’s range in 1886?

So, how should we consider defense? Should we just pretend it doesn't exist? I'd like you to proffer an actual solution to the real problem of measuring defense.

I don’t like positional adjustments and feel that values assigned for positions are highly hypothetical.

You're absolutely free to ammend those numbers to a

I don’t like that positional adjustments are used in batting WAR.

Do you understand why they are used?
Major league shortstops hit .310/.378 in 2012
Major league right fielders hit .327/.434 in 2012.

This difference should be considered, should it not?

Having said that, I think that using a single value to rate players, is just lazy. I see it all the time here, and when I point it out, I get a wave of protest. But it happens here all the time. I think that we have to look at everything, awards, all star games, post season, durability, team success, peak, and career stats.

Why should we use subjective factors (awards) to make a subjective decision?

Look, I know that baseball lends itself to all kinds of mathematical analysis. But Joe Carter was a player, not a math problem.

Has ANYONE ever claimed that Joe Carter was, in fact, a math problem?
Statistics are simply a measure.
Saying we should't use stats to decide what player is better than another is like saying we shouldn't use a scale to decide what person is heavier than another.

[JR Hart]

So, how should we consider defense? Should we just pretend it doesn't exist? I'd like you to proffer an actual solution to the real problem of measuring defense.

But why make evaluations that involve guessing, then pretend that they have merit

Do you understand why they are used?
Major league shortstops hit .310/.378 in 2012
Major league right fielders hit .327/.434 in 2012.

This difference should be considered, should it not?
.

But doesn't overall WAR adjust for that? Regardless, I just don't think that we can quantify what that positional difference is.

Why should we use subjective factors (awards) to make a subjective decision?
.

Because we need to consider how they were regarded at the time. On this board, a common theme is to say the people who gave awards were stupid and we (years later) know so much better. Thst's arrogant and shortsighted. We can't just look back at the numbers

Has ANYONE ever claimed that Joe Carter was, in fact, a math problem?
Statistics are simply a measure.
Saying we should't use stats to decide what player is better than another is like saying we shouldn't use a scale to decide what person is heavier than another.

I never said that we shouldn't use statistics, nice straw man. I said that we should a variety of factors to rate players.

[filihok]

But why make evaluations that involve guessing, then pretend that they have merit

So you can't proffer a better solution?
Should we ignore defense completely?
Any type of evaluation is going to come with certain assumptions, right?

But doesn't overall WAR adjust for that? Regardless, I just don't think that we can quantify what that positional difference is.

See above.
Because the positional adjustment may not be perfect should we pretend that there is no need for a positional adjustment?

Because we need to consider how they were regarded at the time. On this board, a common theme is to say the people who gave awards were stupid and we (years later) know so much better. Thst's arrogant and shortsighted. We can't just look back at the numbers

It is almost certain that we are better able to accurately measure a player's contributions now than we were 50 years ago.

I never said that we shouldn't use statistics, nice straw man. I said that we should a variety of factors to rate players.

The guy saying that Joe Carter isn't a math problem calling 'strawman'...

[JR Hart]

So you can't proffer a better solution?
Should we ignore defense completely?
Any type of evaluation is going to come with certain assumptions, right?


See above.
Because the positional adjustment may not be perfect should we pretend that there is no need for a positional adjustment?


It is almost certain that we are better able to accurately measure a player's contributions now than we were 50 years ago.



The guy saying that Joe Carter isn't a math problem calling 'strawman'...

In other words, you are conceding the flaws, but saying that even though we are guessing with defensive measurements, and can't quanitfy positional adjustment; we should just accept this stat anyway? Prepostorous!

You're also saying that now we have so much better statisical analysis, that we know how good Dizzy Dean was compared to people who covered the league day in and day out? I don't buy it.

[filihok]

In other words, you are conceding the flaws,

Yes. There are flaws.

but saying that even though we are guessing

Nope. We are certainly not 'guessing'

Again, can you proffer a better solution?
Should we just pretend that defense and positional differences don't exist?

You're also saying that now we have so much better statisical analysis, that we know how good Dizzy Dean was compared to people who covered the league day in and day out? I don't buy it.

Yes. That is what I'm saying.

How do you think the people who covered Dizzy Dean back in the day made judgements about his abilities? In part by watching him? Yes. In bigger part by looking at his statistics? Yes.

The statistics that we have now are certainly better than the statistics that we had then.

[Jackaroo Dave]

I don’t like adding OBP and slugging together. They aren’t compatible stats.

This is absolutely on target: OPS is a mathematical abomination. It does correlate well with offensive production. It is popular because it's easy to calculate and understand and accessible on BBREF for group studies. In other words, for the same reason AB, RBI, and BA were popular because they were in the Sunday paper. When BBREF starts using Runs Created Plus, watch references to OPS+ vanish.

don’t know why sabers hate TBs so much, but they show durability also. If we must throw in walks, I would be for a stat like TB + (BB x .60). That would factor in OBP and also show durability.

I don't know that "sabers" DO hate TB. I think your Example would be a very good counting stat; you could devide by PA to get a rate stat, then put it through the park and era adjustment mill to get a relative rate stat, if you wanted. But watch out; you've taken the first steps down the road that leads to linear weights.

I don’t like fielding runs and think that they are a guessing game at best. Do we really know Jack Glasscock’s range in 1886?

Again, this is on the money. No one I know is saying, "Oh Boy, we've got fielding roped and tied." It's a work in progress, and I have my doubts about reaching a satisfactory level of coverage and reliability. This, however, is something to be debated, not grounds for dismissal. The amount of defensive information available now compared to just a few years ago is huge, and much of it does not depend on long strings of inference. But the most cogent argument against using fielding runs is that the most popular form is based on proprietary information, and that was pointed out by Ubiquitous, the stats moderator. When defensive ability is a matter of moment, I look at the available advanced defensive stats, and I'm sure my peers do as well. For example, looking at Garvey's DP-started and assists-not-at-first, it was easy to see that the rumors about his throwing were well based. When I asked BBREF, "How much confidence do you have in your defensive evaluations?" I got the reply, "We can distinguish among good, average, and poor defenders at this point."

That's good enough for a Cowtipper Coot Veal v Cot Deal thread, but sometimes you want to dig deeper . . . . and some of those times you can't. Still better than relying on fielding average tout court.

[Jackaroo Dave]

I don’t like positional adjustments and feel that values assigned for positions are highly hypothetical. I don’t like that positional adjustments are used in batting WAR.

The positional adjustments are not hypothetical; rather, they are based on the league average offensive production at each position. This is itself a little iffy to me, but doing it this way makes the accounting balance. BREF's wrapping them with batting runs in a category called oWAR is just a presentation technique. A lot of people don't like and they just ignore it.

Having said that, I think that using a single value to rate players, is just lazy.

No argument here. The nice thing about WAR is that you can look at the components and see where it comes from and how reliable the sources are. Some people on both sides of the debate don't get this.

I see it all the time here, and when I point it out, I get a wave of protest.

I think a lot of the time you get a wave of other stats.

AND I don’t want a statistic thast factors all that in, because it will enivitably make assumptions that can’t be proven.

In the context of the discussions here, every stat is used with the assumption that it is an index of player value. With BA, HR, and RBI, this assumption has been covert, because it's pointless to defend what is hallowed by tradition and not under attack. The controversial contemporary stats, on the other hand, have to make their assumptions explicit. Thereby, they are much more vulnerable to cogent criticism, which in turn can be used to improve them. These arguments are going on all the time. Look at the blogs in Fangraphs, BBREF itself, Tom Tango's site, etc etc.

[Jackaroo Dave]

Wow, what a truly offensive post. Seriously. You are assuming that because we do not agree with many of the SABRmetrics that we do not understand them. Many of us do understand them and that is why we see the inherent flaws in many of them. . . I believe it is you that are dismissing with the wave of a hand anyone that doesn't agree by throwing us all into the category of unenlightened and/or lazy fools.

Grouping all critics together and addressing them in a response to you was offensive. I apologize. I hate it when JR treats "sabers" as a uniform regimented group (as I suspect he realizes), so I should have been more careful. In particular, when I mentioned "you," sometimes it was singular, sometimes plural. Very annoying, and I'm sorry. What follows is addressed to singular you.

I have read how these statistics are derived and that is why I find flaws, many flaws, in the supposed "chain of explicit reasoning at times expressed mathematically".

Now you are confusing me. It was you who introduced the ab ignoratio argument:

(Sorry, everyone, I'm one of those "unenlightened" old guys that still places more value in what I saw transpire on the field than on the value of WAR+) Just because a new stat has been developed (OPS+) doesn't mean I am willing to accept it as being a valid way of measuring players. Are we really to assume that Garvey was some sort of Svengali that had all the press and fans under his spell in thinking he was a great player when in all actuality he was nowhere near being a Hall of Famer? 200 hits a season are merely "counting" stats, but meaningless in the larger picture? Because he never led the league in assists his putout totals should be ignored along with his fielding pct. because his range factor+ actually made him a poor fielder

When you describe yourself as an unenlightened old guy, refer to OPS+ as a "new" stat, and deprecate two non-existent stats "WAR+" and "range factor +"*, I think I can be pardoned for taking you at your word. This is the language not of a stats expert but of someone who does not see value in such expertise.

(Just by the way, as a 66-year old, I take exception to your implication, albeit ironic, that the intellectually reprehensible unwillingness to learn things that may disturb our deeply cherished notions is a characteristic of "old guys.")

As for not considering me a peer, thank you for the compliment. I do not, however, feel the need to prove to you, nor anyone else here, my credentials regarding a knowledge of the game. If you insist, though, I would welcome a public debate in as large of a forum as possible on any given baseball related topic of your choice. You find the TV and or radio time and let's make it happen.

You have subtly changed the subject: I am certainly not challenging your knowledge of the game, where I am sure I am nowhere close to your equal, but your knowledge of the stats you disparage. Think about it: Suppose I burst into a discussion on the memorabilia forum saying, "I don't know anything about the provenance of this baseball card, but when you've been around as long as I have, you'll realize that those cards are usually fake, so your discussion is beside the point." Wouldn't you find that arrogant, if not . . . um . . . offensive? If you are qualified to criticize the arguments here from a statistical or logical point of view, let's have it. It will improve the discussion and lead to more knowledge--and better stats--all around.

You are expecting us to refute books worth of "reasoning" within the confines of an online forum. It would take at least a book's worth of arguments to lay out and dissect all the flaws.

This is simply not so. JR did a very good job of putting up some cogent objections, which, of course are well known and thoroughly argued by proponents of modern stats. St Anselm designed the ontological argument in 1078. After hundreds of years and thousands of pages, Kant overturned it in a paragraph.

In your first post, you present yourself as too worldly wise and reality-based even to bother with sabremetrics. Now you know too much about it to voice your criticisms in the narrow confines of this forum. Instead, you offer to debate on national TV. Of course I believe you when you say you have learned so much about sabremetrics and you see so many flaws that you wouldn't know where to begin, but some might think it evasive.

I'm not ruling out the possibility of one or many of us doing so in the future, but until that opportunity arises to write and publish one you will have to put up with the uninformed and intellectually lazy rantings of such infidels and troglodytes as myself. One day when I am capable of sitting down to read a whole big book 'bout baseball, maybe I can work hard enough to understand what all of you with a superior understanding of the game are saying.

Under the slathering of irony, you have again shifted the topic from a knowledge of the statistics you dismiss to a general knowledge of baseball. No one disputes your expertise there. I am curious about your claim to expertise about the statistical arguments you dismiss without addressing.

Back to Cey v Garvey: One way you could show up the smarty-pants is to point out the advanced batting stat, a very important one, where Garvey outshines Cey. "Even if sabremetrics were a valid way of judging Garvey and Cey's worth, Garvey beats Cey all hollow in ________, which even you guys have to admit is pretty important." No reason why you should, of course. You certainly don't have anything to prove to me.

---------------------------------------------------------------------------------------------------
*Unless you are talking about an obscure stat invented by a blogger called "Foment" for evaluating hall of fame WAR values, or Contextual Range Factor+, a nonce stat in a BaseballThinkFactory blog about Bill Mazeroski.

[filihok]

But weeee knooooow here on Baseball Fever far better than those who watched every game. We have the superior knowledge. LOL

I see a lot of major league baseball teams hiring analysts and statisticians

http://www.fangraphs.com/blogs/index...lytics-intern/
The internship will provide exposure to all departments within baseball operations. Responsibilities include overseeing daily updates to our baseball information system, monitoring the quality of external data feeds, maintaining databases and spreadsheets, providing technical support for info system users, maintaining our simulation database, operating ball tracking system(s), carrying out ad hoc data analysis projects for various baseball operations staff members, designing and conducting advanced analytical projects, reviewing new research from various external analytical sources, and supporting the department in other ways as needed.

Applicants must have a passion for baseball, strong work ethic, excellent communication skills, attention to detail, ability to work quickly while balancing multiple priorities, strong working knowledge of SQL, keen analytical skills, and a thorough understanding of advanced statistical techniques that are commonly used for baseball research. Other programming and database skills are a plus.

I don't see a lot of teams hiring sports writers to work in the front office.


I wonder whyyyyyyyyyyyyyyyyyy?

[brett]

Look, I know that baseball lends itself to all kinds of mathematical analysis. But Joe Carter was a player, not a math problem. His manager decided that the team could sacrifice outs, in order for him to drive the ball. It worked.

Carter drove in 1 run for every 5.0 outs he cost. An average throughout his career drove in about 1 run for every 5.5 outs cost, despite the fact that Carter batted in the middle of the lineup. In a simulation I saw, if an average major leaguer hit in his spots with his opportunities and hit randomly he would have driven in slightly more runs than Carter AND cost 10% fewer outs over the course of his career.

[JR Hart]

Carter drove in 1 run for every 5.0 outs he cost. An average throughout his career drove in about 1 run for every 5.5 outs cost, despite the fact that Carter batted in the middle of the lineup. In a simulation I saw, if an average major leaguer hit in his spots with his opportunities and hit randomly he would have driven in slightly more runs than Carter AND cost 10% fewer outs over the course of his career.

Of course, Cito Gaston was stupid for batting him cleanup. They could have won with anybody. Thanks for making my point.