The UZR Statistic

[Dipre]

So what happens when your actual game-viewing experience is in contradiction to your statistical analysis?

 

Hasn't happened once.

 

I use both UZR and Rf/G. If i have my doubts about a certain position (Like 1B) and a certain player (Like Tex) i turn to RF/G and use comparable players.

[yankees228]

Average glove work + above average arm =/= Above average 2B. I've watched Cano quite a bit on both the WL and Yankee games. It's not bashing him' date=' i'm simply combining what my eyes and the stats say. Just how i believe it's supposed to be done.[/quote']

 

Well, what other components are you looking for? If he has average glove work, an above average arm, and above average range, (with the occasional mental lapse) I consider that to equate to an above average defender. Admittedly, I'm completely going against the statistics here, so take it for what it's worth.

 

Also, I know you're not bashing him. I find your opinion to be reasonable and objective.

[Dipre]

Well, what other components are you looking for? If he has average glove work, an above average arm, and above average range, (with the occasional mental lapse) I consider that to equate to an above average defender. Admittedly, I'm completely going against the statistics here, so take it for what it's worth.

 

Also, I know you're not bashing him. I find your opinion to be reasonable and objective.

 

The problem with Cano are exactly his mental lapses which include poor positioning in some instances and a lot of questionable decisions.

[Gom]

What about when your viewing of the player is wrong and the Stats are right?

 

Average glove work + above average arm =/= Above average 2B. I've watched Cano quite a bit on both the WL and Yankee games. It's not bashing him' date=' i'm simply combining what my eyes and the stats say. Just how i believe it's supposed to be done.[/quote']

 

Cano.

 

Ok...now look at Tex and Youkilis when it comes to UZR/150. If that doesn't make you throw up, I don't know what will.

 

Here is a question...why not simply look at Range Factor, and ignore UZR?

[Dipre]

Cano.

 

Ok...now look at Tex and Youkilis when it comes to UZR/150. If that doesn't make you throw up, I don't know what will.

 

Here is a question...why not simply look at Range Factor, and ignore UZR?

 

UZR is much better at evaluating all three OF positions, third base, and second base.

[ORS]

Do question the validity of a formula and it's application in determining the value of the formula? To answer both your question and mine' date=' no. Yet in the players I've watched, I'm more accurate than UZR/150 when it comes to comparing players. [/quote']

You keep bringing Teixeira up, so let's jump right in. In your mind, it's impossible for him to be below average by any defensive measure. Here are the components of his UZR:

 

DPRns: -0.4

RngRns: -5.2

ErrRns: 1.9

 

The big negative for him is range. Now he's got very sure hands (see his Error rating), and this is commonly the thing that laymen (that's you) mistake for excellent defense. I'm sure you'll blather on about how there's no way that a stallion (horse pun intended) like Mark Teixeira could have below average range, but he's done that 3 other times in the past. But what's really important is understanding how few plays it takes to get to a -5.2 Runs rating. The run value of a single is 0.48 by linear weights. Assuming all balls that get past him are regarded as singles (and from memory, they count likely doubles as doubles, which would be common for balls up the line missed by a 1B ), it only takes 11 balls missed relative to average to get to a -5.2. Just 11. Over a season, that's 1 missed chance every two weeks.

 

Now, you are expecting me to buy into the notion that a guy who only attends half of his home games, and doesn't document any of his observations, is going to know how accurate a stat is when the difference only manifests itself once every two weeks, and the difference is relative to a league average that the observer is ignorant to.

 

Nope, sorry, not buying.

 

Of this, I have ZERO doubt.

We've already established the infallibilty, in your mind, of your opinion. This is a wasted sentence and impresses nobody.

[ORS]

Cano.

 

Ok...now look at Tex and Youkilis when it comes to UZR/150. If that doesn't make you throw up, I don't know what will.

 

Here is a question...why not simply look at Range Factor, and ignore UZR?

Look up range factor to inform yourself about what it measures, and tell me what's wrong with it.

 

Quickly, your free thinking throne is in jeopardy.

[Dipre]

Look up range factor to inform yourself about what it measures, and tell me what's wrong with it.

 

Quickly, your free thinking throne is in jeopardy.

 

As stated before, the only positions where i don't trust UZR by itself are SS and 1B, confidentially, Gom thinks that because RF better coincides with his "opinion" it's the better stat. Absolutely hilarious.

[Gom]

You keep bringing Teixeira up, so let's jump right in. In your mind, it's impossible for him to be below average by any defensive measure. Here are the components of his UZR:

 

DPRns: -0.4

RngRns: -5.2

ErrRns: 1.9

 

The big negative for him is range. Now he's got very sure hands (see his Error rating), and this is commonly the thing that laymen (that's you) mistake for excellent defense. I'm sure you'll blather on about how there's no way that a stallion (horse pun intended) like Mark Teixeira could have below average range, but he's done that 3 other times in the past. But what's really important is understanding how few plays it takes to get to a -5.2 Runs rating. The run value of a single is 0.48 by linear weights. Assuming all balls that get past him are regarded as singles (and from memory, they count likely doubles as doubles, which would be common for balls up the line missed by a 1B ), it only takes 11 balls missed relative to average to get to a -5.2. Just 11. Over a season, that's 1 missed chance every two weeks.

 

Now, you are expecting me to buy into the notion that a guy who only attends half of his home games, and doesn't document any of his observations, is going to know how accurate a stat is when the difference only manifests itself once every two weeks, and the difference is relative to a league average that the observer is ignorant to.

 

Nope, sorry, not buying.

 

 

We've already established the infallibilty, in your mind, of your opinion. This is a wasted sentence and impresses nobody.

As stated before' date=' the only positions where i don't trust UZR by itself are SS and 1B, confidentially, Gom thinks that because RF better coincides with his "opinion" it's the better stat. Absolutely hilarious.[/quote']

Ok...I should just sit back and watch. Have at it boys.

 

I've watched Tex all season, I see about 140 games, of which 40 are live. His range is amazing. More than Tino, and every bit as good as Don Mattingly.

[Gom]

Dude' date=' you just said Gary Mathews was a good outfielder. That takes everything away from your argument.[/quote']

 

Why? I don't buy UZR as a valid formula. I haven't seen him this year, but according to his RF/G he was decent in CF. Maybe he's lost a step or so, but the guy could flat out go get the ball. I don't think he's lost that much.

 

UZR is useless, in my opinion. As it is constructed.

[yankees228]

Why? I don't buy UZR as a valid formula. I haven't seen him this year, but according to his RF/G he was decent in CF. Maybe he's lost a step or so, but the guy could flat out go get the ball. I don't think he's lost that much.

 

UZR is useless, in my opinion. As it is constructed.

 

What are your issues with the structure of UZR, besides the fact that, in your opinion, it missed on three Yankees this year?

[Dipre]

Why? I don't buy UZR as a valid formula. I haven't seen him this year, but according to his RF/G he was decent in CF. Maybe he's lost a step or so, but the guy could flat out go get the ball. I don't think he's lost that much.

 

UZR is useless, in my opinion. As it is constructed.

 

He barely played there.

[ORS]

Gom, they are all flawed.

 

Again, what is the main flaw of RF? Come on, anyone with a brain should be able to determine this within 5 minutes of looking at how it measures defense. I did, and my criticism was confirmed in subsequent reading about it.

[Dipre]

Gom, they are all flawed.

 

Again, what is the main flaw of RF? Come on, anyone with a brain should be able to determine this within 5 minutes of looking at how it measures defense. I did, and my criticism was confirmed in subsequent reading about it.

 

Easiest formula for a stat ever.

[Gom]

I saw him as a CF. Anyways, irrelevant to the debate that UZR is useless.

[Dipre]

I saw him as a CF. Anyways' date=' irrelevant to the debate that UZR is useless.[/quote']

 

It is. Your eyes tell you he was a good CF and he wasn't. His RF was solid because it can be easily deviated on a SSS.

[bsox0407]

I saw him as a CF. Anyways' date=' irrelevant to the debate that UZR is useless.[/quote']

 

How many times did you see him play CF last Year.

[ORS]

How many times did you see him play CF last Year.

Not enough times to answer my question, apparently.

[bsox0407]

Not enough times to answer my question' date=' apparently.[/quote']

 

He knows he is beat and just does not want to satisfy us with a answer that would actually make sense which would say, " i am to busy watching Youporn than to pay attention to the games"

[Dipre]

As I've said all along' date=' a combination of watching the games and statistical analysis paints the best picture. [/i']

 

This is not what you have said. At all.

 

It is what i have said, what ORS has said and what Y228 has said.

[bsox0407]

This is not what you have said. At all.

 

It is what i have said, what ORS has said and what Y228 has said.

 

I totally missed that hahhahha. Blatant lies and making more stuff up shows the real truth.

[Gom]

I've said it all along. In many posts. In many threads. I also said that UZR is useless, and that defensive metrics lag behind offense and pitching ones.

[Dipre]

I've said it all along. In many posts. In many threads. I also said that UZR is useless' date=' and that defensive metrics lag behind offense and pitching ones.[/quote']

 

You've said UZR and defensive metrics are useless, you also said defensive metrics lag behind the others. However, not once have you expressed faith in them or a reasonable understanding, if you did, we wouldn't be having this argument. I can prove i've said that, and unless you can prove it too, no one is buying it.

[yankees228]

I don't understand why you think someone is treating defensive statistics as gospel just because they mention them. As Dipre alluded to, many of us have said that the best way to evaluate defense is with a combination of defensive statistics and visual observation.

 

You continue to falsely perceive other people's opinions.

 

EDIT: Obviously, at Gom.

[Gom]

RF has many errors. Dependent on many factors, least of all the type of pitcher on the mound. Groundball or strike out pitchers will yield a smaller range factor. That is one of many factors involved.

 

228, I have no problem with people quoting the stats, but when faced with someone who states that scouting reports differ, they start with "watch the gamezz" and other illogical crap that shows they are incapable of independent thought. So I take them to task every time. No one has yet told me how something like UZR [mind you, we've moved on from UZR to RF] can have such variety in the numbers. See it's based on multiple formulas...so an aberration in one will lead to wildly changing figures, to the point to render the formula mainly useless.

[ORS]

RF has many errors. Dependent on many factors' date=' least of all the type of pitcher on the mound. Groundball or strike out pitchers will yield a smaller range factor. That is one of many factors involved. [/quote']

It has one glaring error, and you are close, but the error is a bit more general and not as specific as K pitchers vs. contact pitchers. The problem is that it is a counting stat. It counts the number of outs a fielder generates (put outs + assists) per game. It makes an assumption that all fielders at the positions on the field will generally receive the same number of opportunities over a full season. To give you a comparable example on the offensive side, it's like judging hitters by the number of hits they got prior the use of BA, which actually happened in the early part of the last century.

 

228, I have no problem with people quoting the stats, but when faced with someone who states that scouting reports differ, they start with "watch the gamezz" and other illogical crap that shows they are incapable of independent thought. So I take them to task every time. No one has yet told me how something like UZR [mind you, we've moved on from UZR to RF] can have such variety in the numbers. See it's based on multiple formulas...so an aberration in one will lead to wildly changing figures, to the point to render the formula mainly useless.

No, "we" haven't moved on. "We" understand why you like, it matching your opinion (which you overvalue).

 

RF was the first attempt by a statistician to look beyond fielding percentage when rating defensive players. When Bill James created this he was working from his house on his own dime and didn't have the resources for something more comprehensive, which would require a staff, access to game tape for all games, and time. When Stats Inc. was started up by John Dewan, you had someone who had the resources to start a venture that would provide the information that he knew there was a market for. Along came ZR (zone rating), which determined how many opportunities each player had inside his fielding zone. Of course, the flaw here was that it assumed each opportunity was equal, and the reality is that some opportunities are more difficult than others. UZR adds the context of evaluating the quality of the opportunties.

 

Obviously, it isn't perfect, but if you consider the methodology, it's attempting to measure defensive range in a fair manner. The main problem is that it still remains subjective, like your opinion, but they at least they have an established set of parameters they use when applying the subjective qualification of each opportunity.

 

The ideal stat would remove the subjectivity. For example, in the future, a sensor system (with miniature transponders on each players uniform - hat, belt, somewhere) could track how far a player had to move to field each ball and how much time they had to get there. The distance/time relationship would be an objective average for each position based on measurable data.

 

IMO, we'll see this, or something similar, in the future as the appetite for meaningful information has not been sated, but where we stand right now is pretty good.

[Gom]

IMO, we'll see this, or something similar, in the future as the appetite for meaningful information has not been sated, but where we stand right now is pretty good.

At this point, I think we can agree that both defensive metrics and scouting have their flaws. I'm of the thought process that scouting is much more effective than defensive metrics. I guess we can leave it at that.