[nat]

"A lot of them, simply by stealing chances from his centerfielder on routine plays"

This is a known issue. There's some philosophical differences over whether it's a problem or not - it really depends on what's you're trying to measure. If you have Greg Luzinski in the outfield, having him play next to someone who can "steal" chances from him is a really valuable thing to have. If you have other really good outfielders it doesn't matter so much.

As for Mattingly: he was a really good defensive first baseman. I'm not objecting to his gold gloves. dWAR (the defensive component of WAR) itself has two components: a measure of how many runs a player saved based on the plays that he made, and a positional adjustment. Mattingly saved a lot of runs, but, like all first basemen, he's hit with a big negative positional adjustment. But you really do need a positional adjustment, otherwise a short stop who could hit like Mattingly, and a first baseman who could hit like Mattingly, would end up with equal WAR. Whereas obviously the SS is more valuable, since it's so much harder to find someone who can both hit like Mattingly and play short stop. We can argue over how large the positional adjustment should be (how large it should be for DH's is a matter of rather extreme controversy), but it's clear that there's got to be some, and it's clear that it's got to be large and negative for first basemen. But again, that's not a knock on Mattingly's defense, he was a very good defensive player for a first baseman. It's just a reminder that it's easier to find good hitters who can play first than it is to find good hitters who can play short.