Why should anyone be surprised that Teixeira ranks lower than his reputation on defense? You can’t “see” a guy’s range, especially at 1b. So we use measurements like UZR to measure range. Turns out Tex is overrated.
No, just that UZR doesn’t see Tex as a good defender (though MGL would say that’s not precisely true either). It could mean that Teix is overrated. Or it could mean that UZR doesn’t handle first-basemen well. Or it could not handle first basemen in new Yankee Stadium well.
Excellent article. One comment I would make is instead of “WAR component” (you use both “offensive” and “batting”), I think it would be clearer if you consistently used RAA, or spelled it out like with Uggla. When I see something like “25.4 WAR…”, I immediately think it is an aggregate WAR over years, rather than a single-season RAA. After rereading the line I see, “oh, just offense”. Perhaps I’m the only one…but I think it can help avoid confusion.
I don’t a ton of stock into UZR for 1B because I think alot of times firstbasemen don’t get to balls to their right that they could because it actually makes it a tougher play vs. just letting the 2B go and get it.
Comment by Dwight Schrute — February 9, 2011 @ 1:55 pm
I think for both 1B and especially catcher, a lot of the value is not strictly in fielding range, so it’s hard to measure reliably.
Mike K’s suggestion and Austin’s corrections have been implemented. Both make too much sense.
Comment by Joe Pawlikowski — February 9, 2011 @ 2:18 pm
So Ozzie will therefore play Dunn at 1st and Konerko will be the DH???
I love the defensive rankings (UZR especially) – nice to have metrics. Does a 1 baseman’s “range” count less for UZR than range for a SS? Seems intuitive… but though to put a value on range for different positions, no?
Comment by Pistol Pete — February 9, 2011 @ 2:41 pm
I don’t think UZR is useless, but I do think it’s next to useless for firstbasemen. Just like it can’t tell us much about catchers, even though catcher defense is really important, I think it’s missing a lot of the picture with firstbasemen. Mostly, the responsibility of the first baseman is to simply cover first base.
He needs to assume the second baseman can get to most balls to his right. If a firstbaseman makes a play at the absolute edge of his range, he has turn around and throw with his momentum taking him the wrong direction, trying to hit a moving target. A lot of the time he’s better off not even making the throw because pitchers aren’t always the most reliable targets. If the firstbaseman lets the ball go through and covers the bag, the secondbaseman still has a chance to field it at the edge of HIS range, and his momentum is taking him in the proper direction and he’s got a good target to throw at for the out. Worst case, it rolls through to the right fielder who presumably has the best outfield arm and can still prevent a double.
I know for a fact that there’s a “Scoop” factor as well, which Tango and MGL have looked at, and I still am not sure it’s completely accounted for because they looked at errors, primarily. There are certainly plays where a good fielding firstbaseman can help coax an “out” on a bang-bang play on which there would have no error assessed if the runner had reached, but a more indifferent fielder at first wouldn’t have earned the out. These are rare, obviously, but since there are like 1500+ plays at first base every year, the very rare still comes up often enough.
I obviously don’t know everything there is to know about UZR, so I was wondering if pre-pitch position over the year(s) effects a player’s UZR? The reason I ask is because Wright’s UZR was above average up until ’08 where it plummeted. It just so happens that 2009 was the first year Jerry Manuel was at the helm full-time and he always made Wright guard the line. Would this effect his overall UZR? It’s just odd to me when you look at his UZR over the years.
Comment by The Nicker — February 9, 2011 @ 3:22 pm
I guess to clarify, if you took the difference between offensive RAA and defensive UZR runs, I’m sure he’d make the top 5.
Comment by The Nicker — February 9, 2011 @ 3:23 pm
I believe positioning will affect UZR only as to whether or not the fielder makes the play.
That’s one of the complaints people have about it, that it doesn’t know anything about positioning. However, for one, there isn’t any data there. For two, how do we know that the player didn’t make the change to their positioning on their own? If they did, it’s a skill. For three, are we worried about how much ground the player covered to get the ball, or whether he got it? Of course the answer is, “it depends”. For projection purposes, we may want to know how much ground they covered. But for looking at last year, we only care about how many balls they got too.
I believe MGL’s position is that he investigated scoops by first-basemen, and didn’t find that it was worth more than a run or two at the extremes. And since it is doesn’t add anything to the equation, and is a lot of work, he skipped it. I’m not sure if TotalZone includes scoops, and if it does if Rally found the spread to be greater.
Pretty sure UZR does *not* dock a 1B, if the 2B makes the play. So if the 1B is letting balls go by that are in a shared-zone, and the 2B could not make the play, I think it is indeed correct to dock the 1B. Sure, it’s tough to decide in a split second if the 2B has a better play on it, but if some first baseman are better at doing that than others, they should get credit for it, right?
Where I think Teix may lose a few points in UZR (and potentially get them back in TZ?), is popups. UZR I believe ignores popups, since 1) they are usually no brainers and 2) who catches the pop has nothing to do with skill – if Jeter and Cano can both get to a pop, who do you think is going to call for and catch it? I think Teix is particularly good at catching pops in foul territory that only he could get to, but he gets no credit for it. Those are also some of his flashiest plays, which is one of the reasons he gets recognized as a great defender. So sure, maybe Teix is around average outside of popups and scoops, but they add 10 or so runs to his defense?
I think Braun (a good athlete) is actually improving toward respectability as a LF
Comment by westcoast hero — February 9, 2011 @ 4:28 pm
I think Dwight hit part of the issue on the head… Tex will let a routine ball go to 2nd instead of making a tough play and pray Burnett or CC can cover 1st. Then you have the Cabrera’s of the world (not that he has great range) who will range as far as possible to nab any ground ball even if the 2nd baseman will easily field (see last out/hit of near perfect game). I know UZR corrects for that with OF’s but I’m not sure if it is done with infielders (specifically 1st and to a lesser extent 3rd).
The other thing I’m not sure about is how double play runs are dealt with… does it include all DP possibilities (3-6-1) and are you potentially factoring in the capabilities of the pitching staff defensively?
Also how are foul balls dealt with? Are they part of the rangeR component and is it adjusted for relative foul territory. A guy playing in Oakland will be able to rack up more foul popups, and even if they are only adding small amounts to UZR if they are considered to be in a routine (high probability out) zone, it is still tacking on range.
I’ve watched Wright play a ton, and while he may have slipped a little, there’s no way he’s the worst in the league like UZR may have it. He’s one of the best at coming on slow rollers, plus it’s not like he’s Jorge Cantu over there. He can move a bit. UZR has a long way to go to gain credibility.
I agree that Wright is excellent on charging bunts and slow rollers but his throwing has really deteriorated. He relies on the sidearm flip way too much and makes routine plays more difficult. It drives me nuts to watch because he has excellent range but not a premium arm. Still, I am not sure where that puts him league-wide besides being behind the obvious choices like Zimmerman and Beltre.
Comment by AccusedWizard — February 9, 2011 @ 6:34 pm
Ah, citing one year of UZR data as fact. Excellent. It is a very dangerous thing to look at one year of fielding data and consider it as concrete as hitting data. If anything, I would like to see UZR discoutned by 50% or so in the WAR rankings, because of its inherit flaws and fluctuations. For all those people who belive UZR is actually the way God intended man to scrutinize defense: http://www.baseballprospectus.com/article.php?articleid=11476
Scoops are only one part of a 1Bmans job though. His reach and footwork can mean the difference between a hit and an out on throws to 1B, or a 2 base error and an out.
As an example, a 1Bman with poor reach and/or footwork may get pulled off the bag on a less than perfect throw while Teixeira makes that play and gets the out. Depending on the play, it is either scored a hit or an error. On close plays, the ability of the 1Bman to stretch for the ball and his reach can be the difference between an IF hit and an out.
Since Teixeira came on board, Yankee IF’ers make far fewer errors
Big difference between throwing to Teixeira and Giambi. I always felt part of Renterias problems in Boston were Kevin Millars lack of reach. When Olereud played 1B Renteria made few errors, but he made a ton when Millar was at 1B (thinking you have to make a perfect trhow can also lead to fielding errors and not just throwing errors). Having confidence in your 1Bman may be one of those intangibles we can not measure.
Teixeria made over 40 outs on foul PO. Many of these were down the RF line, and he led MLB in this category. Some were spectacular. He was kind of like Beltre at 3B in his ability to go back on popups, but he was easier on other players ribs
Also, the issue of positioning for a 1Bman is far more complicated than other positions. Depending on the runner and a teams philosophy on positioning , a 1Bman can be positioned very differently than the average 1Bmans position. UZR is unlikely to handle this very well.
I would prefer instead to average it with DRS and Total Zone. This will smooth out the outliers we see from time to time in UZR (upside and downside) and keep the proper weighting for defense Never use only 1 model when you can average several similar models, since no model is 100% accurate, and in fact, their accuracy is unkown.
While I’m not going to debate that Konerko does not get to a ball more than 2 feet one way or the other, the problem I have with using metrics like UZR is they don’t take a couple factors into account. Konerko has very good hands around the bag and is very skilled at picking balls in the dirt. He saved Alexi and Mark (I should be playing for the Joliet Jackhammers) Teahen, countless errors during the season by picking low throws. Also, Konerko throws well for a first baseman. He is very skilled at making throws to start the 3-6-1 DP, or cutting down lead runners on the bases. These are a couple of elements that don’t get factored into advanced metrics obviously. Again, I’m not trying to excuse Konerko’s non-existent range, but I clearly don’t think he should be labeled as the worst defensive first baseman in the league.
That theory isn’t really supported if you look more closely at the data. The reason for the big drop off in his UZR the past few years has been a big drop off in his range (RngR), the other components have remained relatively consistent. He’s always had somewhat of an erratic arm, I don’t think that has gotten all that much worse in recent yrs, it just seems for whatever reason his range has dropped off a lot.
I agree, and his UZR has improved each year he has been in left field. Just from watching the games, though, while I think he’s about average in range and getting to balls, I don’t think he has a particularly strong or accurate arm.
Great point about pop-ups. Also, Teix seems to stand out in his ability to get outs by throwing home or 2nd base in force-situations. Giambi, by contrast, would rarely ever throw to another base and would instead take the easier out at first. Not sure how UZR calculates that, as only “one out” is made, but in Teix’s situation, he’s saving a situational run or preventing a guy from getting into scoring position.
I think a lot of defensive metrics are subject to a lot of variation. Not totally sure exactly how it’s calculated, but from what I do know, it seems that a lot is generalized. Zones, starting position.
Here’s what I want (if I’m an idiot and this already exists then kill me), I want a way to track the distance a fielder runs to get a ball and the time it takes the ball to reach his glove. That’s one metric, and the other is time it takes from him to transfer the ball from his glove to the first baseman. You divide each of those not into “routine” and “range” but what is “routine” for the fielder and what is his outer range.
The reason I say you say what is routine for a certain fielder is because I really think the best fielders are the ones that you don’t see on top plays. David Wrights one handed dive was a bad play to me, spectacular yes, but catching it with his bare hand and diving backwards? How about taking a proper route to the ball and catching it with your glove?
I don’t know if current metrics take into account how hard a ball is hit or how far the fielder had to move, but they should if they don’t.
Also, Tex is infinitely better at making the throw to 2nd for the force or the 3-6-3 or 3-6-1 DPs than Giambi. Half the time, such a throw from Giambi wound up in short left center, and another 30-40% of the time, the throw got to 2nd too late to turn two if it was near enough to the bag to get a force.
You can talk about a few guys problems on defense if you want to, but the problem is throughout the MLB, and it stinks, not unlike some of the garble you create. But, this is only one of the skills of the game that is lacking in todays players, and a lot of it has to do with physical conditioning, too fat.
Comment by ropespinner — February 11, 2011 @ 6:45 am
Metrics are easy. Problem is data. There is no way to determine where fielder started and how hard was the ball hit. Recently, accuracy of the spot where the ball was fielded / hit the ground was also questioned. Stringers aren’t marking starting position or hang time.
If FieldFX becomes available to people doing these metrics, it should be easy to make a metric that would include all that information and that would be more accurate than anything we have now.