My blog- analysis of the San Francisco Giants, Baseball, and Sabermetrics.
The reason I think its flawed is because, Im just simply biased.
And, I dont get it.
Our QB > Yours
UZR should not ever be used at 1b.
When does range really become a factor at 1B? The plays some first baseman save with their reaction and their ability to scoop or save throws is > the range they cover.
It depends, are there more scoops than groundballs out of range? It's different with different teams obviously, when you compare having a poor arm in Jeter or an absolute cannon in Tulowitzki.
I've yet to come across some statistical model that isn't flawed in some way or another. The question really is whether it is less flawed (or closer to perfect) than other models - ie is it generally more reliable.