There are several things wrong with what this guy says. Just in case the article gets taken down from PSA or something, here's the original: http://itsaboutthemoney.net/archives/2011/09/06/is-war-the-new-rbi/
UZR and flyball rates
UZR is a counting stat, of sorts, so it is true that outfielders with more opportunities have a chance to pad their UZR more. However, bad fielding outfielders will see their UZR get even worse with more opportunity, so this conjecture is completely wrong. Even if a team's pitchers had 99% groundball rates, that wouldn't cause their outfielders to have a negative UZR. If they had great outfielders, they would still most likely have a positive UZR, it just wouldn't be as large a number as it should.
The author tries to justify this by saying that a couple messed up plays is the cause of this, but then unknowingly destroys this theory by showing that it is a systematic shift. A couple missed plays that cause good fielders to have negative UZR would be a random shift. The fact that it is a systematic shift says something else. Maybe GMs know that they have extreme flyball pitchers, so they look for good fielding outfielders? Or maybe they know they have extreme groundball pitchers, so they aren't afraid to put a lumbering power hitter in right field.
Basically, lack of opportunity will not systematically make good outfielders see negative UZRs, nor would more opportunity make bad outfielders have positive UZR.
Edit: I just went through and graphed team FB% against team outfield UZR in excel for every year from 2002-2011. Here's the graph:
You'll notice (or maybe you won't since it came out so fuzzy, just trust me) that the R^2 of the linear regression line is .0234, so flyball rates are only accounting for 2.34% of the variation in outfield UZR, which is statistically insignificant.
Try selling this: According to WAR, in 2011, Carlos Lee has had as much defensive value as Troy Tulowitzki
Yeah, except this isn't true at all. They have the same UZR, sure, but your UZR score does not equal your defensive value. You can't say that a SS with a 9 UZR is equal in defensive value to a LF/1B with a 9 UZR, you have to include the positional adjustments. This is a really common mistake I see all the time with people trying to find faults with UZR, when in fact the fault is with their understanding of the stat. UZR is how many runs you save above average for your position. The average SS saves a lot more runs than the average LF. If you include the positional adjustments, Lee is worth 1.4 runs on defense, and Tulo is worth 15.1, or a difference of about 1.4 wins in defense alone.
Plus, the fact that Lee has a higher UZR this year than he's seen in a few years has nothing to do with him playing several positions, and has everything to do with the fact that his errors are down, his range is better, and he has almost twice as many outfield assists as last year despite only playing 60% as many innings in the outfield. He is playing better. Sure, with more opportunity he might screw it up and have his UZR fall, but his UZR being positive isn't just because of small sample size, it's because he has played better.
There are two types of utilitymen, those who are given the job because they play many positions well and those who are given it because they play no position well. As yet, WAR struggles to distinguish between the two.
This isn't correct either. If you were throwing Adam Dunn around through all the fielding positions, WAR would most definitely reflect the fact that he sucks at all of them. His positional adjustment would be higher, yes, but his UZR would be a lot lower, compensating for that. The writer of this article obviously doesn't know that UZR by itself is only valid when comparing two players who play the same position. You can't expect to take a first baseman with a UZR of 0, plug him in shortstop, and expect his UZR to still be a 0. That is not how UZR works.
Saying that Ian Kinsler provides more value than Miguel Cabrera isn't saying it's better to have him. Part of Kinsler's value is in that he plays a premium defensive position. If you want to know which player it would be better to have, you have to look at their own standings within their position. If Kinsler was the best second baseman in baseball, and Cabrera was only like the 5th best first baseman, I might say it would be better to have Kinsler, since you would have the best player in baseball at a position. Of course, this also depends on who you would be putting at the other position that you didn't fill with one of these stars. Really, WAR is not addressing the question of "who would you rather have," that is a complicated question with more variables.
WAR suggests, by the way, that Morgan has been more valuable on a per game basis than Fielder
This is an irresponsible manipulation of WAR. WAR was never meant to be broken down into a per-game basis because part of the value of a player is how many games he plays. Jesus Montero has .075 WAR/game, on pace for 12.15 WAR over a whole season, so he must be the most valuable player in baseball, right? (Well, maybe that's true). Breaking down WAR into a per-game basis should never be done, it's not a valid analysis.
Now, to address the example of Morgan and Fielder you have to take into account that Morgan is having a great year. The reason why his WAR/game (sigh) is higher is because he is a gold-glove caliber center fielder who has been 22% better than the average major league hitter, which is much much rarer (and thus MORE VALUABLE) than a poor fielding, bad baserunning first baseman who is 53% better than the average major league hitter. Good hitting fist basemen are a dime-a-dozen, but a good hitting plus defense center fielder is very hard to find. <em>However,</em> Fielder ends up being the more valuable player so far because he has played in 42 more games than Morgan.
it’s beyond asinine to conclude that Ellsbury is twice as valuable as Fielder.
It's really not, this year. Part of the point of WAR is saying how hard that player would be to replace. There are a lot of first basemen who can hit. Maybe not as well as Fielder, but the Brewers could find a good hitting player in their minor league system and call them up, or they could move one of their corner outfielders to first for a few games if Fielder got injured. The point is that first base is the easiest position to replace. Center field is not as easy. Not many players can play above average defense in center field, and only a fraction of them can do so while putting up MVP caliber offensive numbers. Yes, Ellsbury has been far more valuable than Fielder. His offense has been almost as good, his defense has been far far better, and he plays a premium position. I don't see how this is even a debate right now.
Even WAR’s adherents, like Dave Cameron, generally admit the margin of error is at least 15%. When we stubbornly suggest that 0.5 WAR means anything, we are grossly exaggerating the statistic’s accuracy, even according to its creators.
What a .5 WAR means, with a 15% error margin, is that the player is between a .425 and .575 WAR. Either way you slice it, that is a below average player. Even when you get into the higher WAR ranges, say a 10, the range is between 8.5 and 11.5. This is still an MVP caliber player. The 15% error range doesn't diminish what WAR is telling you, it just means that you have to be careful in judging two players whose WAR are close to each other. If the error ranges of two players' WAR don't overlap, the higher WAR player was almost definitely better. If they do overlap, there is some probability that the higher WAR player was better, based on the amount of overlap.
Misunderstood data is misrepresented and polemicized.
Well said. How ironic this sentence is, considering how little this author apparently understands about the stats he tries to criticize.