"No god" and "god who does nothing" are very different metaphysically but have the exact same observable consequences, and evidence for or against one will equally be evidence for or against the other.

While I agree that there will be no observable difference, you're talking about two different axes here. One axis is "how much God does", and most of your spectrum is running along this axis. The other axis is "whether God exists", and treating that as part of the same axis is an error. (Admittedly, the axes are related - the idea of a universe where God doesn't exist yet is nonetheless active is rather absurd - but they are still not the same axis).

My perspective is a little different.

...the way you've phrased #3 reduces your argument to a tautology on close reading (specifically, #1 and #3 must always match regardless of #2). I think (and feel free to correct me if I'm wrong) that a better phrasing for #3 would be "how the universe should look if God exists" (instead of "...if #2 is correct"). Then, for theists, #2 would be "theist" and #1 would match #3; for atheists #2 would be "atheist" and #1 may or may not match #3.

Then my view is that those who start with creationist leanings (in #1 and #3, #2 being "theist") but pursue a scientific career find, in the course of their studies, that #1 (the observable features of the universe) are not as they had thought; then #1 and #3 no longer match. #3 is complex and difficult to change without feeling like it's being changed arbitrarily (and will probably need to be changed repeatedly as #1 changes with further study); but #2 is a switch, far easier to flick, and therefore far more commonly flicked.