Fangraphs has UZR data now

[Transmogrified Tiger]

Awesome stuff. Some players of note, the numbers are UZR/150 for 2008, which prorates to a full season, meaning guys with less PT at some positions could see some more extreme results. The numbers you see are in runs.

 

Fukudome: +13.4 in RF

Edmonds: -17.6 in CF

Johnson: -12.0 in CF (career +19.6 in LF, -15.6 in CF, -3.2 in RF)

Soriano: -3.1 in LF (+4.4 in '07)

 

Ramirez: -4.2 at 3B (+13.5 in '07)

Theriot: -0.5 at SS (+2.7 in '07)

DeRosa: -8.8 at 2B (+3.5 in '07); +20.3 in RF (-18.6 in '07)

Fontenot: +16.5 at 2B (+7.6 in '07)

Cedeno: -5.1 career at SS, +22.6 career at 2B

 

 

Here's some current names being bandied about:

 

Ibanez: -11.0 in LF (-23.9 in '07)

Abreu: -25.9 in RF (-3.9 in '07)

Dunn: career -10.1 in LF, -64.4 in RF, -13.3 at 1B(grain of salt position alert)

Bradley: +21.9 in RF (+15.5 career)

Hermida: -3.0 in RF (-3.9 in '07)

DeJesus: -7.4 in CF (+7.5 in '07, +5.5 career)

Roberts: -0.3 at 2B (+3.7 in '07, +4.2 career)

[CubColtPacer]

Does UZR have questions with first base defense? I noticed you hadn't posted Lee so I looked him up. According to UZR, Lee had his best defensive season of his career last year, and it wasn't even close. Overall, he's in the negative for his career.

Are there particular positions that it is supposed to be better at than others? I know that is the case for some defensive rating systems.

[Transmogrified Tiger]

Does UZR have questions with first base defense? I noticed you hadn't posted Lee so I looked him up. According to UZR, Lee had his best defensive season of his career last year, and it wasn't even close. Overall, he's in the negative for his career.

Are there particular positions that it is supposed to be better at than others? I know that is the case for some defensive rating systems.

 

I didn't include Lee because I don't think we're to the point with any system that can accurately get a grip on 1B defense because of how many throws they receive. I think it does as good a job as other positions with regards to fielding batted balls, but when you get maybe 2 of those per game compared to 10+ throws a game, it doesn't paint a very accurate picture.

[haltz]

Does UZR have questions with first base defense? I noticed you hadn't posted Lee so I looked him up. According to UZR, Lee had his best defensive season of his career last year, and it wasn't even close. Overall, he's in the negative for his career.

Are there particular positions that it is supposed to be better at than others? I know that is the case for some defensive rating systems.

 

I didn't include Lee because I don't think we're to the point with any system that can accurately get a grip on 1B defense because of how many throws they receive. I think it does as good a job as other positions with regards to fielding batted balls, but when you get maybe 2 of those per game compared to 10+ throws a game, it doesn't paint a very accurate picture.

John Dewan tried to tackle this looking at game video. It looks pretty subjective, but according to this method, Lee is good, Pujols is a god (and/or Eckstein skipped 900 baseballs at him in 2005), and most 1B fall within a few plays of each other.

 

Anyway, fangraphs is amazing. A WAR stat would be sweet and pretty simple with UZR and wOBA data.

[Transmogrified Tiger]

Does UZR have questions with first base defense? I noticed you hadn't posted Lee so I looked him up. According to UZR, Lee had his best defensive season of his career last year, and it wasn't even close. Overall, he's in the negative for his career.

Are there particular positions that it is supposed to be better at than others? I know that is the case for some defensive rating systems.

 

I didn't include Lee because I don't think we're to the point with any system that can accurately get a grip on 1B defense because of how many throws they receive. I think it does as good a job as other positions with regards to fielding batted balls, but when you get maybe 2 of those per game compared to 10+ throws a game, it doesn't paint a very accurate picture.

John Dewan tried to tackle this looking at game video. It looks pretty subjective, but according to this method, Lee is good, Pujols is a god (and/or Eckstein skipped 900 baseballs at him in 2005), and most 1B fall within a few plays of each other.

 

Anyway, fangraphs is amazing. A WAR stat would be sweet and pretty simple with UZR and wOBA data.

 

How did he define "bad throws" though? Guys like Pujols and Lee are going to make a lot of throws look a lot easier than Scott Hatteberg(or even Todd Helton) would have because of their stride.

[Colin Wyers]

It's not as much of a skill as you might think. Something like two runs per year for a guy like Lee who's good at it.

[sneakypower]

fangraphs makes life worth living

[SouthSideRyan]

Does UZR have questions with first base defense? I noticed you hadn't posted Lee so I looked him up. According to UZR, Lee had his best defensive season of his career last year, and it wasn't even close. Overall, he's in the negative for his career.

Are there particular positions that it is supposed to be better at than others? I know that is the case for some defensive rating systems.

 

I don't believe it takes into account receiving throws.

Edited December 8, 2008 by SouthSideRyan

[Colin Wyers]

Does UZR have questions with first base defense? I noticed you hadn't posted Lee so I looked him up. According to UZR, Lee had his best defensive season of his career last year, and it wasn't even close. Overall, he's in the negative for his career.

Are there particular positions that it is supposed to be better at than others? I know that is the case for some defensive rating systems.

 

I don't believe it takes into accounts receiving throws.

 

It doesn't; MGL seperately accounts for that in his SuperLWTS system.

[Mephistopheles]

So, here, we have a statistic that has an extremely large error in terms of actual fielding "production", and that even if we considered that it had no error, it's still 80% randomness.

 

Really there's no use in wasting our time quantifying such data.

[Vanilla Ice]

So, here, we have a statistic that has an extremely large error in terms of actual fielding "production", and that even if we considered that it had no error, it's still 80% randomness.

 

Really there's no use in wasting our time quantifying such data.

what defensive metrics do you like the best?

[nilodnayr]

So, here, we have a statistic that has an extremely large error in terms of actual fielding "production", and that even if we considered that it had no error, it's still 80% randomness.

 

Really there's no use in wasting our time quantifying such data.

 

to a very large degree, i agree with this, however, taking a weighted average of the last 3 years of uzr IMO is better than nothing

 

the thing with advanced defensive metrics is that their results vary drastically year to year, so they obviously aren't measuring true talent. Its a bit that the stats are flawed (which even the advanced ones are to a degree), but mostly theres a huge amount of randomness. Is Crisp really the best defender one year and below average the next year...well maybe. So that begs the question...whats the point of valuing defense if it can be so random from year to year. However, of course there are the guys that are consistantly bad (lots of the corner OF types) or consistantly good (endy chavez, franklin gutierrez), but there are a lot of guys that vary in that year to year.

 

So, don't look at the 2008 UZRs and take them as dogma for what the players defense will be in 2009. Look at the last 3 years, take a weighted average, add in some aging if necessary and regress it to the mean. You'll find that guys like Abreu/Ibanez/Dunn are pretty darn bad.

[Tyrant]

So, here, we have a statistic that has an extremely large error in terms of actual fielding "production", and that even if we considered that it had no error, it's still 80% randomness.

 

Really there's no use in wasting our time quantifying such data.

what defensive metrics do you like the best?

 

Whichever players can best field a rat off a DeMarini.