Mark,
I think we are in agreement.
The problem is that it is impossible to come up with strict criteria and a formula for calculating a rank.
At least in U.S. News & World Report's college rankings, you can see the data that was compared and how it was weighted. There are books on the topic. Of course, we're still talking about trying to quantify the unquantifiable, in my opinion (declaring Harvard is "better" than Princeton or Williams or Virginia is lunacy). However, it is easy to see how those results are reached, factoring in endowment, selectivity, etc.
I don't see how you could ever develop a comparable scientific method for rating architecture, because that ranking depends on personal preferences rather than discrete data. Because different people prefer different things (I might like an engineered look, my friend might favor a more natural style), those rankings aren't really helpful to us.
Obviously, there's no real harm in it, but in response to the title of the thread, I don't think there is another way of ranking that would offer more "valid" results.