Exciting news this weekend, FanGraphs is now providing UZR data — Ultimate Zone Rating — on it’s website. UZR for those who don’t know, is one of the more heralded defensive systems out there. Consider UZR another handy tool to stick in your tool belt when it comes to player analysis. UZR splits the field into “zones” — hence the ‘Z’ in UZR — and calculates how many outs each player was expected to make compared to average. UZR looks at the probability of turning a ball in play into an out — how hard it was hit, where it was hit, what kind of hit (line drive, groundball, flyball) it was — and then makes a few adjustments. Adjustments such as: park factors, batter handedness, GO/AO ratio of the pitcher pitching, and base/out situation.
The Hardball Times has a nice article on UZR and some other defensive metrics here. You can also read some of MGL’s original posts on UZR which prove to be very informative if you are curious about the inner workings of the system. The bottom line is that UZR is one of the more respected defensive systems out there and to be able to get it for free, is awesome news. Let’s check out some 2008 UZR numbers for the San Francisco Giants.
2008 Giants Fielders by UZR (min. 300 innings)
| Name | Pos | Inn | UZR | UZR/150 |
| Randy Winn | RF | 1108.1 | 16.9 | 18.9 |
| Omar Vizquel | SS | 657.2 | 5.3 | 13.8 |
| Fred Lewis | LF | 905.2 | 7 | 12.1 |
| Rich Aurilia | 1B | 477 | -0.2 | -0.5 |
| Jose Castillo | 3B | 820 | -0.9 | -1.6 |
| Eugenio Velez | 2B | 449.2 | -0.9 | -3.2 |
| Ray Durham | 2B | 535.1 | -1.3 | -4.1 |
| Rich Aurilia | 3B | 427.2 | -1.2 | -4.9 |
| John Bowker | 1B | 550.1 | -1.9 | -5.9 |
| Aaron Rowand | CF | 1275.1 | -12 | -11.1 |
| Emmanuel Burriss | SS | 315 | -3.2 | -14.1 |
The table is sorted by UZR/150 or how many runs a player would save over 150 games. I’ve also limited my data to fielders with a minimum of 300 innings played at a defensive position.
A quick note on sample sizes and defensive metrics. Looking at Burriss’ -14.1 runs below average for position might freak you out but understand that 315 innings is but a tiny drop in the bucket. It doesn’t tell us anything about Burriss. It has no predictive value and all we can say is that in 315 innings played at shortstop in 2008, he was below average. Is that his true talent level? We don’t know. The sample size is just too small and ideally you’ll want a few seasons worth of data before you try and figure out the true talent level of a player. Dave Cameron makes a good point on USSM about comparing players across positions. Burriss may have been below average in his brief 315 innings but UZR will compare him to his peers; ie: other shortstops. This is good because we want to know how Burriss stacks up against other shortstops but also remember that shortstops, as a group, are some of the most athletic defenders on the baseball diamond. A -10 shortstop isn’t the same as a -10 left fielder or first baseman, who are on average, less athletic. Burriss is being compared to Omar Vizquel, Cezar Izturis, J.J. Hardy, and Jimmy Rollins. Fred Lewis is being compared to Adam Dunn, Carlos Lee, and Pat Burrell. Context is important when thinking about these rankings.
Is Randy Winn one of the most undervalued players in baseball? Every year he hits about the same, providing league average offense — wOBA+’s of 96, 107, and 108 over the last three seasons — while playing some of the best right field in all of baseball. Last season, Winn was the best defensive right fielder in the game by UZR. He was worth +18.9 runs by his defensive play alone, that’s nearly +2 wins. This isn’t anything new for Winn. From 2006-2007 he was worth +15.3 and +11.9 runs above average in right field. Winn’s game is extremely un-flashy but as an overall player, he’s been one of the better Giants over the last couple of years.
Besides Randy Winn, Omar Vizquel and Fred Lewis are the only other players who scored as above average fielders in UZR. Vizquel’s bat might be dead but his defense is still very good. If Vizquel would have somehow played 150 games — that’s not going to happen any more — he would have saved 13.8 runs. A team who needs a late inning defensive replacement might consider Vizquel, just don’t let him hit too much. Fred Lewis did well in LF in ’08, he was worth +7 runs above average for his playing time and would have been closer to +12 runs in 150 games played. Lewis’ good foot-speed helped him to overcome sometimes awkward routes in LF to be a above average defender.
The Giants have to be worried about Aaron Rowand’s defensive performance in 2008. Rowand has been a good defender in CF but he saw a complete reversal of fortune last year. Check out his UZR’s from 2002-2007: +12.6, +9.2, +9.6, +16.5, +4.6, and +7.9. Those are solid defensive numbers to be getting from your CF. I’m not sure Rowand is the type to age well — plays hurt all the time, runs into walls, GAMER, etc. — and the Giants have to be hoping he’s not hitting a wall (no pun intended) at 30-years-old. If the team does have to move him out of center, I think he’ll be headed to LF. Playing right field in AT&T is almost like playing a second center field and if Rowand can’t hack it in CF, I can’t see him doing much better in RF. Next year will be a big year for Rowand and if he continues to slide, his contract starts to look much worse. I’m hoping he was hurt but we can’t just throw out all of his 2008 data and wish that it didn’t happen. For what it’s worth, CHONE’s defensive projections for 2009 see Rowand bouncing back to a +7 CF.
The rest of the list has the usual suspects. Rich Aurilia is an average defender at first base but takes a dip at third. John Bowker struggled at first base in ’08 and I’m guessing the Giants won’t revisit that expirmanet anytime soon. Velez rated better than I thought he would have, but remember, the sample size is too small to draw any conclusions from. Jose Castillo played better at third than I expected and Durham is closer to an average defender than I expected. Ray might have a couple of years left in the tank if he wishes to pursue them. New guy Edgar Renteria isn’t on my list, but he scored as a +1.1 shortstop for the Tigers in 2008. PMR hated Renteria and plus/minus had him as below average, but he might not be as bad as we thought. If you check out his fielding data over the past four years, you can see he looks to be pretty steady, fielding around average each year.
Also, a quick comment on the new FanGraph UZR’s is that they are using BIS data instead of STATS data. I believe they each define their zones slightly differently so you might find some variation between the older UZR stuff and the new stuff on FanGraphs.
Still, an amazing set of data to be able to browse though freely.
I am curious, why is Burriss’s stats too small to have predictive value? I don’t know all the statistical terminology (or the answer), but know that when a value is at an extreme, then the possibility that it represents something predictive improves a lot, the null hypothesis could be calculated and determined to be true N% of the time. Typically, you want to see N to be greater than or equal to 90 or 95 percent, but it would also be significant to know if, say, maybe it was 75-80%, which is still a strong chance, just not the slam dunks 90-95% is.
Burriss’s UZR is at the opposite end of the spectrum of Winn and Vizquel. He might only have 315 innings but that is half of what Vizquel played. Is there a study that shows at what size the sample has to be? Or a way to analyze sample size with UZR and get a guess at what the true level might be? I know with stats that one can do that, just don’t know if that applies to UZR.
About Rowand, I have to think that there is either an adjustment period or permanent problem with playing CF in AT&T. Take a look at Winn’s numbers for CF after joining the Giants. After years of mildly positive UZR, he comes to the Giants and post negative numbers. Per your point with Burriss, could be small samples, but for him, it is multiple seasons, and always negative. And if hitters have to adjust to hitting there, I would assume that fielders would have to adjust as well, given the odd shape of the ballpark. Too bad we can’t see the UZR for each month, see how it jumps around, or by home vs. road.
300 innings is just too small to draw conclusions from because of the randomness of BIP distribution.
USSM has a nice defensive primer and it touches on sample size for defensive stats:
http://ussmariner.com/2006/01/24/evaluating-defense/
As for this:
>>About Rowand, I have to think that there is either an adjustment period or permanent problem with playing CF in AT&T.
Are you saying that CF at AT&T is so hard for Rowand to play, that he went from about a +5-7 run defender all the way down to a -11 run defender? He’d played CF at AT&T before with the Phillies and how many countless games has he played at the position? How many different parks has he played in? Heck, Steve Finely at 41-years-old played an average defensive CF for the Giants in 2006 (UZR of 0.8). Winn was close to average too in ’06 and above average in ’05. I bet he could play an average CF in ’09. If those guys can handle the fielding why can’t Rowand? Maybe he lost a step. Maybe it was an injury. I don’t know, but with his age, if he continues to play below average defense, the Giants have a problem.
UZR data by month would fall to the same problems of small samples, BIP distribution is going to make it useless to draw any conclusion from.
Rowand is actually projected to +7 in the corner, 0 in center. Same as Winn.
@Baetown
I agree with CHONE, I’m sure that Rowand would be above average in a corner spot but I think he would struggle in RF at AT&T. That makes him a LF if/when he ever gets that bad on defense. I mainly worry about a player like Rowand who has been historically hard on his body and how he’ll age.
Yeah, I’m worried about those years too. Maybe Sabes unloads him before he falls off dramatically.
Like OGC, Rowand worries me. This is all observational, but CF at ATT seems like death for guys who don’t have blazing speed (and don’t be fooled by his age, even at 41 Trash could fly). I just don’t think Rowand has the wheels to cover that kind of turf.
In Philly, not only was he playing in a bandbox, he also was basically playing Left-Center since Victorino was able to cover about 3/5 of the outfield. His success in CHI, I think, was a combination of playing with Posednick, who at that point was covering more ground than just about any LF in baseball, and the fact that Rowand was still at peak age.
The problem now is that if Rowand can’t be an above average defensive CF, he ceases to have any real value. He doesn’t have the arm to be an effective RF, and he doesn’t hit enough to be a worthwhile LF.
I’m still a huge fan of the GAMERness, and yes, I do believe that he provides some “intangibles” that go beyond the numbers, but I’m not optimistic that his defensive numbers will rebound.
MrLomez,
Rowand may have played with some good fielders in the past, but don’t forget that he played with two excellent fielders last season. Winn was the best RF in all of baseball by UZR. He’s a top-3 RF but other metrics and Lewis was above average in LF, too. The Giants basically had three CF’s playing OF last year — not that I think Lewis can play CF, but he’s played the position in the minors.
I definitely agree that the Giants need him to play at or around average defense in CF to have good value. I just fear that Rowand is a player-type that might age poorly because he’s always hurt it seems.
I was going to pull up a list of comparable players off of BP but I think they are down right now.
Also, UZR adjusts for park. So, it should make the proper adjustments for playing in a bandbox like Philly.
So, I think, we’re more or less in agreement. Rowand is suspect going forward, the reasons for which (injury-ness or incompatibility with ATT) are moot. Rowand is the CF for the foreseeable future, which is somewhat troubling.
I do think it’s inaccurate to say that Lewis and Winn are the equivalent of CF’s. Lewis certainly not. He has the speed, but oh my, it makes me cringe to imagine him running in circles in center. Minus drastic improvement in reading the ball off the bat, Lewis would be a tragedy in CF.
Winn’s prowess, I think, is more a consequence of his utter mastery of ATT’s RF dimensions and quirks than innate ability. He is probably the most cerebral outfielder I’ve ever watched, but throw him in the boundless terrain that is CF and all that cerebralness and mastery of angles is negated.
Yes Rowand is surrounded by two guys who are good at playing their respective spots in the OF, but it’s different than playing with true CF’s flanking you, like a Carl Crawford, or Victorino, or Posednik circa 2004.
Yeah, it might be different but the question is “how much does it matter”? Winn is an average CF and Lewis is probably slightly below, but my point was that Rowand wasn’t flanked by immobile fielders like Barry Bonds or Carlos Lee, he’s playing between two plus-defenders.
So, to that extent, it doesn’t really matter if they can play CF or not, they are playing their positions quite well. If Rowand looks this bad playing between plus-defenders, I don’t like his prospects for the future (as I’m sure you don’t either).