I don't think it's a terrible idea. Obviously it has inherent limitations, but I get the logic.
Makes more sense for, say, New York State than Rhode Island obviously.
What I've never understood, however, is why Vermont thinks, "active cases per million" needs to be normalized to Vermont's data. I dont know if I'm missing something obvious, but I fail to see the logic there. I also dont know what they mean by, "known metrics", and AFAICT, they never explain it, which seems pretty important.
At any rate I'll be able to go to Vermont again in a few days, because the Vermont model has a significant recency bias which is pretty obvious to figure out if you know your counties' data per day, and our "bad data" for my county was from last week.
On the "by-county" thing, it's probably the last-bad solution they could come up with that had any reasonable chance of implementation. It also meets an important goal: allowing travel to/from upstate New York without lumping the city into the same bucket. At this point, that seems less important, but when the city rates are significantly higher, it allows some level of protection (if the rules are enforced, of course).
I suspect the normalization to Vermont's rates is a "well, allowing travel there shouldn't make things worse"-type theory. Or it may be political in the form of "we're gonna get raked over the coals if we say people doing as well or better than us are a danger to local residents."