Should we have a Sabermetrics forum in the Cafe?

[LBJackal]

How do you run statistical analysis with data points that have hundreds of independent variables and very few consistent correlations?

I don't know, and I don't have to know. What's your point? Like I said, why don't you use some of these skills you claim to have for good instead of just saying you have them?

ZR is one of the best measures of defensive range. It's accurate, despite having flaws. Every stat has flaws. Having a .333 BA doesn't mean you'll always get a hit every 3 at bats. Having a .893 ZR doesn't mean you'll always get to 89.3% of the balls in your zone during a given season. But it's a very good indicator. What do you suggest we use to judge range? Watching them play? How many data points and independant variables do you use for "watching them play".

I like sabermetrics and empirical information because it is objective. If you don't, then fine.

[Mookie4ever]

ZR is one of the best measures of defensive range. It's accurate

How do you know?

I don't like how James divides up the field. It's completely static and yet at the same time is entirely dependant on a James employee to make a decision on whether a ball is a fly or a line-drive.

I would argue that ZR is neither accurate nor useful. There's a reason that ZR is not an MLB official stat.

It's gimmicky and propagated by James to sell his product.

He doesn't have a professional opinion to sell like scouts so he sells stats and has to come up with newer and more gimmicky stats to sell.

"There are lies, damn lies and statistics"

[Mookie4ever]

I like sabermetrics and empirical information because it is objective.

Statistics are rarely objective. The person giving you the stat almost always has a story to tell and something to sell.

One thing I've learned is that you can get an "expert" to say anything you want so long as you pay him enough.

[LBJackal]

It's done by Stats, INC. And it's the only system that rates players based on play-by-play data. There may be an instance here and there wher ea ball is borderline line-drive/fly ball, but I believe they use a system that times how long the ball is in the air and how far it traveled, to determine the difference. Even still, this is negligible over the course of a season, and there will be borderline plays for everybody. And yeah the way they divide the field is static, but it's the same for everybody. UZR addresses this but I don't think it's been released yet this year. Also, history has shown that ZR correlates very well to UZR. There are always exceptions, but ZR is a much better measring stick for defensive range than any other stat out there that I know of... and better than just watching the game.

If somebody has a better stat, I'd love to see it. I'm not saying there isn't one, just none that I know of.

[LBJackal]

I like sabermetrics and empirical information because it is objective.

Statistics are rarely objective. The person giving you the stat almost always has a story to tell and something to sell.

One thing I've learned is that you can get an "expert" to say anything you want so long as you pay him enough.

Right, nothing is perfect. At least with stats you can see the rationale. With "watching them play" it's an opinion, and it's easier to lie with an opinion that with stats.

[FatGuyWithAMullet]

Also, history has shown that ZR correlates very well to UZR. There are always exceptions, but ZR is a much better measring stick for defensive range than any other stat out there that I know of... and better than just watching the game.

A) Here's a little tid bit from stat class: Correlation does not mean causation.

B) You have to understand that every single defensive statistic is inherently faulty since the model requires assigning points to things that are not direct outputs. It's a fallacious use of statistics that shouldn't be used.

[Mookie4ever]

B) You have to understand that every single defensive statistic is inherently faulty since the model requires assigning points to things that are not direct outputs. It's a fallacious use of statistics that shouldn't be used.

I'm not with you on this one FGWM. Almost every stat, and certainly every baseball stat, has got room for the exercise of discretion and is faulty to some extent. That doesn't mean that we shouldn't use them (otherwise there would be no FB), all it means is that we should give more weight to certain stats than to others.

ZR has got some use but I don't think that you should, on the basis of ZR alone, say that Andruw Jones or Pudge are poor defensive players. I don't give much weight to ZR.

[LBJackal]

Also, history has shown that ZR correlates very well to UZR. There are always exceptions, but ZR is a much better measring stick for defensive range than any other stat out there that I know of... and better than just watching the game.

A) Here's a little tid bit from stat class: Correlation does not mean causation.

B) You have to understand that every single defensive statistic is inherently faulty since the model requires assigning points to things that are not direct outputs. It's a fallacious use of statistics that shouldn't be used.

I've taken enough stats classes, but thanks for the little primer there buddy.

ZR is relevant, and the players who are viewed as perennial Gold Glovers usually have a history of great seasons with ZR early in their careers, and rightfully win Gold Gloves because of it. The problem with just watching them play, is that they keep that aura of being great throughout their career even when their range drastically declines, such as the case with Andruw, Edmonds, Vizquel, Bret Boone, and many other players.

I've conceded the whole time it's not perfect, but don't go and say it has no relevence, and no defensive stat has relevence (most don't however, since only a couple - ZR and UZR - are based play-by-play data). There is obviously a margin for error which is why it shouldn't be used to rank range exactly as the ZR's go. But when your ZR declines steadily, do you not think there is a reason behind that, which is a decline in the player's range?

How does Andruw go from top tier in 1997, 1998, and 1999 (avg ZR of .910) to dead last over the span of 2 years, 2003-2004 (avg ZR of .838)? Sure that doesn't mean he's for sure the worst for range, although he very well could be, but it's a pretty solid indication that he's gotten a lot worse, and that he isn't nearly the best OF in the league like many people make him out to be.

[FatGuyWithAMullet]

B) You have to understand that every single defensive statistic is inherently faulty since the model requires assigning points to things that are not direct outputs. It's a fallacious use of statistics that shouldn't be used.

I'm not with you on this one FGWM. Almost every stat, and certainly every baseball stat, has got room for the exercise of discretion and is faulty to some extent. That doesn't mean that we shouldn't use them (otherwise there would be no FB), all it means is that we should give more weight to certain stats than to others.

ZR has got some use but I don't think that you should, on the basis of ZR alone, say that Andruw Jones or Pudge are poor defensive players. I don't give much weight to ZR.

While it's true that statistics are open to interpretation, some statistics are abominations to mathmatics and logic. I've explained numerous times why defensive statistics are fershit. If a statistic is a simple mathmatic formula and mathematically sound it is simple to buy into. Batting Average is a perfect example of this; it is a straight ratio. In contrast, OPS isn't. Instead, it is the addition of two fractions with different denominators without creating a common denominator, so it's validity is zero. Just to clarify - A player with an OPS of 1.000 doesn't perform in an arithmatically predictable manner better than someone with an OPS of .800. The end value isn't a mathmatical calculation.

People can show me as many formulas as they want, and I really don't care. They should know as well as I that you can manipulate numbers to say just about anything you want. I could easily plug numbers into a formula and come up with some arbitrary number that is assigned a positive or negative connotation.

I enjoy baseball statistics as a whole, they are the backbone of the game, as long as they are used appropriately, and are mathematically significant to the analysis they are being used to support.

[FatGuyWithAMullet]

Also, history has shown that ZR correlates very well to UZR. There are always exceptions, but ZR is a much better measring stick for defensive range than any other stat out there that I know of... and better than just watching the game.

A) Here's a little tid bit from stat class: Correlation does not mean causation.

B) You have to understand that every single defensive statistic is inherently faulty since the model requires assigning points to things that are not direct outputs. It's a fallacious use of statistics that shouldn't be used.

I've taken enough stats classes, but thanks for the little primer there buddy.

ZR is relevant, and the players who are viewed as perennial Gold Glovers usually have a history of great seasons with ZR early in their careers, and rightfully win Gold Gloves because of it. The problem with just watching them play, is that they keep that aura of being great throughout their career even when their range drastically declines, such as the case with Andruw, Edmonds, Vizquel, Bret Boone, and many other players.

I've conceded the whole time it's not perfect, but don't go and say it has no relevence, and no defensive stat has relevence (most don't however, since only a couple - ZR and UZR - are based play-by-play data). There is obviously a margin for error which is why it shouldn't be used to rank range exactly as the ZR's go. But when your ZR declines steadily, do you not think there is a reason behind that, which is a decline in the player's range?

How does Andruw go from top tier in 1997, 1998, and 1999 (avg ZR of .910) to dead last over the span of 2 years, 2003-2004 (avg ZR of .838)? Sure that doesn't mean he's for sure the worst for range, although he very well could be, but it's a pretty solid indication that he's gotten a lot worse, and that he isn't nearly the best OF in the league like many people make him out to be.

Why won't you open your eyes and see that stats don't always tell the whole story and sometimes they even lie? It's like you refuse to believe there is anything more to this game than a stat book, Moneyball, Bill James and BP analysis. It's futile attempting to discuss anything with you because all you are able to even comment about is statistical analysis, derivitives, and equations. You offer no individual analysis, but rather reguritate everything you've read regarding statistics. Put down the books and watch a game, you might learn something.

I'm done with this stupid topic. I'm sick of you repeating everything you read and spouting off like it's the newly discovered universal truth of baseball. It's just going in circles now, I'm done.