You can grapple with the statistic all you want, but the Balls in play data is pretty solid. It’s not like ESPN Gameday, where the “ball in play” information is some middle-aged dude sitting in the press box saying “looks like it landed HERE!” and clicking a box on a computer screen. MLBAM and BIS have been developing these tools for years, and there’s no internal bias at play there. Perhaps some inaccuracy, but it’s not human error. The calculations are done with computers, and if you look at the underlying ball-in-play data (which includes calculations of the speed and angle a ball comes off a player’s bat for every ball in play, among other things) you get a real sense of how extensive the data really is.

The big problem with defensive stats is that none of them isolate for defensive positioning, which makes it really hard to look at ball-in-play data and get an accurate assessment of a player’s range.

Saying that fangraph over-relies on UZR is both right and wrong. If you think that the FG writers quote WAR as if they think it’s a stat that’s 100% accurate and cannot be improved on, then you’re right. In all fairness, some of the articles come off that way sometimes. That’s really not how the authors see it, though. When they quote the final WAR number, it’s more like a shortcut, where it’s understood that if the WRAA+UZR+Replacement Level + Positional Adjustment formula calls a guy a 4-win player, there’s a built-in understanding that the imprecise defensive stat means his “true talent” range is somewhere between 3.5-4.5WAR. The FG authors used to build that range into their posts before UZR became publicly available on the site, and now they just use the UZR number as a quick-and-dirty approximation.

So yeah, nobody thinks UZR is spot-on. There’s little doubt, though, that it’s the best defensive metric that’s freely available and gives us the best way to approximate a player’s true value.

I am saying that if you take thousands of balls hit in a zone, and say throughout the majors x% were turned into outs, and player Y turned fewer than x% into outs, it is information, sure, but it doesn’t really tell you that the average player would have turned more of those particular balls hit to player Y into outs.

But i understand the concept that if 250 balls are hit to a player and he makes no errors and converts 220 on average, whereas another player makes 10 errors but converts 240, you take the guy who converts more balls into outs.

And I agree with your other post that they will eventually calculate the park effects better, and that in general the more seasons of UZR data you get the more reliable it is. I get a problem with the way fangraphs says, “oh, Jay Bay is worth x dolars, Franklin Gutierrez is worth $26.8 million this year” blithely tossed around. Maybe they are worth that, but i think it is a little more nuanced than that.

BTW, his UZR has improved last couple of years. It was +6.6 this year.

Did anyone see the Curb Your Enthusiam episode last week where Larry gets in the fight with the stonemason about Jeter? The stonemason says he is overrated, terrible fielder, etc. You know Larry doesn’t know from UZR, BABIP, clutch-hitting, etc.

My statement had nothing to do with any sort of statistical comparison.

Gonzalez also played good defense in center in ~530 innings in 2008 in Oakland.