Defense: The Brewers and UZR

by Jordan M on Dec 7, 2008 12:27 PM CST Comment 21 comments

Fangraphs has added UZR to all of its player pages. It's not a perfect stat, but it is one of the best estimators of total defensive value, and now it's nice and easy to find. For the explanation, here's the fangraphs article:

RngR (range runs): The number of runs above or below average a fielder is, determined by how the fielder is able to get to balls hit in his vicinity.

ErrR (error runs): The number of runs above or below average a fielder is, determined by the number of errors he makes as compared to an average fielder at that position given the same distribution of balls in play.

UZR/150 (ultimate zone rate per 150 games): The number of runs above or below average a fielder is, per 150 defensive games.

So here's a chart of the Brewers' starters last year. Finally we have an accessible metric that balances errors and range!

Player	08 UZR/150
Prince Fielder	-7.5
Rickie Weeks	-5.3
J.J. Hardy	12.5
Bill Hall	9.5
Ryan Braun	-4
Mike Cameron	13.3
Corey Hart	-0.7

Craig Counsell (3B)	5
Gabe Kapler (OF)	9.3

Fielder has been about that bad his entire career, between 7-10 runs below average, so we can figure on him being a win below average in the field going forward.
Weeks was terrible in '05, better in '06, nearly average in '07 (-1 run), and slipped a bit last year. I'd expect him to be just slightly below average next year, I think he's better than most people think.
Another metric that really likes Hardy. I think we undervalue him. It's like 3/4 metrics have him as a very good defender and 1 doesn't like him.
I also think we underestimate Hall's D at third, and UZR reaffirms that. If Gamel is a Brewer 3B in '09, I wouldn't mind Hall playing a '07 Counsell role-- coming in after 6 innings when the Brewers lead. Note to Brewer managers: this assumes that Gamel is not hitting at a MVP-caliber level.
UZR doesn't like Braun's range in left, because he had no errors this year. That's understandable, and I'd expect him to improve next year with more experience. Contrast it with his -37.3 (!) UZR/150 at 3B in '07. That's over a 3-upgrade just by moving Hall and Braun and signing Cameron last season.
Another solid year for Cameron in center. Surprisingly, UZR had him at -7.2 with the Padres in '07. I wonder if the move from a big park like Petco to Miller Park helped him defensively and offensively.
Corey Hart just had a terrible year. After being 5 and 10 runs above average/150 in '06 and '07, he was slightly below average this year. If he gets back to normal levels defensively, that could be a half-win upgrade just by itself.
Counsell dropped off from his awesome '07, but still was a plus defender this year. I wonder if there is a deal on the table for him.
UZR likes Kapler as well. I doubt he's back, but some team will get a good player when they pick him up.

This is really fantastic that we have a catch-all metric available on player pages. Here's a few more random things I found playing with this:

Possible 3B Mike Lamb is a career -3.4 UZR/150 at third, leading me to believe there are better options for the lefty half of the platoon and that Gamel might be worth a shot if Lamb is the 1st option.
Manny Ramirez was consistently in the -20 run region with the Red Sox, but he turned in a -6 overall this year. Maybe there was something to his effort failures in Boston.
Mike Jacobs: -17 per 150 last year. That is some incredible suckitude to be 1 win worse than Prince Fielder.
Please sign Raul Ibanez, Cubs. For lots of money, preferably. -11 last year, -23 the year before.
via i27.tinypic.com (lookout landing)

via assets.sbnation.com (lookout landing)
Evan Longoria, 17.7 at 3B, is a monster.
Everybody's favorite Gold Glove Winner, Nate McLouth, was -15.5 by a metric that uses errors in its evaluation method.
Trade target David DeJesus is +5 career in CF, +20 in LF. I like him more now.
Adrian Beltre was worth 16 runs in the field last year. If we're going after a good 3B, make it be him.
Carlos Lee was only -2 in LF this year, and it appears he's not as bad as I might have thought he was. He's not in the league of Burrel, Dunn, Ibanez, and Ramirez.
The worst player I have found is Brad Hawpe, at -47.7 runs/150 games. That is unbelievable. 5 wins below average on defense. The difference between Carl Crawford for 150 games or Hawpe in the outfield is 70 runs or 7 wins.

0 recs | Comment 21 comments

Story-email Email Printer Print

Comments

I’m just learning about fielding statistics. Thanks for the information!

About Weeks being better than we think, yes, I’d believe the numbers, but is some of fan observation about Weeks being a bad defender related to how bad Prince is? (In other words, Weeks making what looks like a bad throw only because half the other 1B in the league could have got it.)

Also, with Craig Counsell around, everyone else looks crappy in comparison.

by morineko on Dec 7, 2008 2:32 PM CST reply actions 0 recs

Remember - just like hitting, it's possible for a player to have an up/down year on defense.

The best way to get a sense of where a player is at overall is to look at 2-3 years of data. In his career he’s -10 at second base per 150 games. (I really don’t feel like trying to do weighted averages at this time of the morning – I’m on vacation.)

I has a blog. | Ask a sabermagician!

by cwyers on Dec 8, 2008 10:14 AM CST up reply actions 0 recs

Hmmm

You know, I never thought about that.

"I will agree that the attitude [at BCB] is ridiculous and they have done so much to instigate animosity and then block us from responding. Real mature!"

by roguejim on Dec 8, 2008 2:35 PM CST up reply actions 0 recs

I wonder how much the numbers would change

If the Brewers used a different official scorer (since errors are involved).

Obscure baseball records and more at my blog, Recondite Baseball.

by TheJay on Dec 7, 2008 2:52 PM CST reply actions 0 recs

Right, it's a rough estimate

You have to take it for what it’s worth. But it’s the best we have right now, and most of the data passes the common sense test. If nothing else, it’s just combining zone rating and fielding percentage to get a more accurate overall fielding value. It could probably fluctuate 2 runs or so based on error calls, and that’s why you have to make generalizations instead of taking it is a specific run value.

And neck size to baby eating ratio.

by Jordan M on Dec 7, 2008 3:32 PM CST up reply actions 0 recs

What makes you say it is the best available now?

by ol Pete on Dec 7, 2008 6:23 PM CST up reply actions 0 recs

By "available"

I mean that I can easily find it for every player, for their entire careers, and it will update in-season.

And neck size to baby eating ratio.

by Jordan M on Dec 7, 2008 6:28 PM CST up reply actions 0 recs

Do you have a stat you like better?

And neck size to baby eating ratio.

by Jordan M on Dec 7, 2008 6:47 PM CST up reply actions 0 recs

for defense - runs allowed - work back from there

The underlying concept, that defensive plays are made in isolation, is flawed.. The idea that players have responsibility for zones is flawed. The raw data is unscientific. A player can play excellent defense and get a crappy rating. A crappy fielder can get a decent rating. There is no known (at least to me) measure of reliability. Given that estimates like PECOTA refer to themselves as extremely accurate because, using their own liberal standards, they get 2 out of 3 right, I wonder if most defensive metrics are worse.

Even the people who create enormously long attempts at estimating defense admit that it takes years of data to achieve a large enough sample size. Saying something like Counsell’s defense has dropped off last year is inaccurate. His defense may be unchanged and it would be more accurate to say that he has performed at the average of the two over the two years. I wouldn’t even do that.

I remember when most defensive metrics had Hardy as average. This year, I think I’ve read him as 1st, 2nd or 4th best depending on which metric you choose. Is he that different a fielder? Is his range different?

by ol Pete on Dec 8, 2008 8:54 PM CST up reply actions 0 recs

If you start with the conclusion that we can't measure individual defense, we're not going to get anywhere

My point is that UZR is a combination of the two things we know defensive players can do themselves— field lots of balls and not screw up once they’ve fielded them.

For a team, the best way to measure the overall defense is to find the percentage of balls in play converted into outs. But we can’t split that up into individual players.

UZR’s not perfect, but it gives us an idea, and it’s interesting to look at. I don’t pretend to have watched every player in the MLB extensively like national TV guys do, so I have to rely on some numbers to get an idea of how good a guy is defensively. These numbers— UZR in 08 for the Brewers— back up my observations pretty well.

And it isn’t smart to declare a guy a bad defender based on 1 metric. If you’re doing a study on a guy, you’d look at all the available metrics over the largest possible sample to determine how good he’ll be defensively moving forward. But if I’m just curious as to how good, say, Hank Blalock is on defense, because I’ve only seen him a few times, I check UZR to get an idea.

And neck size to baby eating ratio.

by Jordan M on Dec 8, 2008 9:42 PM CST up reply actions 0 recs

don't get me wrong

All the metrics, and I don’t put much stock in any of them (speaking only for myself) are fun, interesting and potentially illuminating. I think in the not too distant future, the raw data will radically improve and then so will the metrics.

I would also say that there aren’t many defensive plays that are made by individuals in baseball.

by ol Pete on Dec 9, 2008 12:00 AM CST up reply actions 0 recs

The big difference between UZR and ZR is that it looks at every zone seperately.

ZR looks at every zone in a fielder’s area of responsibilty together. UZR looks at how many plays above/below average a player makes in each zone seperately, and then consolidates this information. (Technically then no ball is “out of zone”, which is a big help as well).

MGL did a nice writeup of UZR a few years ago on BBTF, here and here. Some details have changed since then, of course (the one I remember off the top of my head is excluding plays where the typical wishbone shift is probably on, determined by who the batter is). MGL also talked about making adjustments for when the 3B is presumably guarding the line against doubles or when the infield is likely playing in, although I don’t know if he went through with them or not.

I has a blog. | Ask a sabermagician!

by cwyers on Dec 8, 2008 10:19 AM CST up reply actions 0 recs

is there a generic ZR any more?

Seems like there are about a dozen zone ratings.

by ol Pete on Dec 8, 2008 8:35 PM CST up reply actions 0 recs

No, no post, just showing that again.

by HRF on Dec 7, 2008 4:18 PM CST reply actions 0 recs

This is why I wouldn't be against trading Fielder

In a perfect world we’d trade him and sign Teixeira, but that’s obviously not going to happen.

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by Jack Moore on Dec 7, 2008 4:39 PM CST reply actions 0 recs

I wonder how Dunn would handle 1B full-time

And neck size to baby eating ratio.

by Jordan M on Dec 7, 2008 5:37 PM CST up reply actions 0 recs

Couple things.

Adrian Beltre was worth 16 runs in the field last year. If we’re going after a good 3B, make it be him.

Not to say that Beltre wouldn’t be better than Hall offensively—he probably would—but he’s still the kind of offensive player that the Brewers have too many of: good power, below average OBP. If Hall is worth a win on defense already, and Beltre is worth maybe 1.5 wins there, is it worth picking up Beltre’s contract, or should we take our chances that Hall can post a .320 OBP in a platoon split with someone better than Lamb? I guess signing Beltre would at least eliminate the need for a platoon. I don’t know the answer to this question. I think it matters who the platoon partner for Hall would be.

It’s cool to see that the Brewers did little things right in 2008 (to make up for the Riske and Gagne flops, among other things), like having two great defenders on their bench who also had their merits offensively (Counsell can draw a walk and work a count, Kapler flashed some power). Kapler will be too expensive, but I bet we can get Counsell back for a cheap year — I don’t think he wants to leave Milwaukee at this point in his career.

"Nothing is working right now with a lot of the guys. We're trying to see pitches and see what we can do. ... I'm not going to sit there and walk, though. I'll eventually find it, and hopefully we'll still be in it." -Corey Hart

by baumann on Dec 8, 2008 11:23 AM CST reply actions 0 recs

Seattle hates righty power

Away 08: .292 .349 .512
Away 07: .288 .320 .538
Away 06: .280 .351 .462

And don’t forget the last time he was in a walk year, with the 04 Dodgers, even though I don’t really believe in the “performance spike”

.334 .388 .629

And remember, UZR is only looking back. I’d put Beltre at +2 wins going forward and Hall around +.7 on defense.

And neck size to baby eating ratio.

by Jordan M on Dec 8, 2008 12:54 PM CST up reply actions 0 recs

good point.

i thought about that ‘performance spike’ too, but then i thought ‘steroids’ and forgot about it.

those away numbers look good. of course, if we could guarantee he’d have a .850 OPS and be a 2-win defensive player, i’d ask what we need to trade to get him.

also: Hall/Counsell platoon at 2B.

just kidding.

by baumann on Dec 8, 2008 4:01 PM CST up reply actions 0 recs

how do they measure range?

Does it take play-to-play positioning into account?

by tcyoung on Dec 8, 2008 11:08 PM CST reply actions 0 recs

read this comment above

And neck size to baby eating ratio.

by Jordan M on Dec 9, 2008 8:36 AM CST up reply actions 0 recs

Comments For This Post Are Closed

Search

FanPosts

Community blog posts and discussion.

Recent FanPosts

+ New FanPost All FanPosts >

NL Central Standings

	W	L	PCT	GB	STRK
St. Louis Viva El Birdos	91	71	.561	0	Lost 6
Chicago Bleed Cubbie Blue	83	78	.515	7.5	Lost 1
Milwaukee Brew Crew Ball	80	82	.493	11	Won 3
Cincinnati Red Reporter	78	84	.481	13	Won 2
Houston The Crawfish Boxes	74	88	.456	17	Lost 3
Pittsburgh Bucs Dugout	62	99	.385	28.5	Lost 2