It's easy to change my mind on this: just provide me a decision rule that'll put rural decay data points on the "failure" side of the classification boundary while leaving urban decay on the other.

That's just the igon value.