But that's Glenn's point, too; that the number can be artificially inflated or deflated. That it assumes that the .921 would have the same save percentage on more or less shots--and he's showing that it can be deflated. Otherwise, you're suggesting that a goalie with a low GAA and average save % would see the .921 goalie come in and lower the GAA some more even though that team won the Jennings, for example. It assumes the transposition exercise that is taking Ryan Miller from the Sabres to the Blues would see his save percentage skyrocket instead of tank.
You've made a great post I'm not trying to knock it, but taking raw save percentage as gospel independent of team effects is folly.
Oh I know that's his main point, I wasn't trying to discredit that. But the specific post seemed to suggest that the original difference would be negligible, and on top of that, we should adjust for his theory of shots affecting save percentage. Now, assuming the theory is true, it might elevate the .916 performance to the .921 performance, but it does need to be elevated for it to be similar, because the difference isn't negligible.