In my answer to this meta question I argued that the decision of whether or not a question should live on Physics SE is best approached by judging whether answers to that question would be likely to come from and be useful to other physicists. At the time, we were discussion questions with engineering or experimental flavour. However, I would argue that this approach should be used for all types of questions.
Recently, a question about moving poles around in the complex plane was migrated to Math SE, presumably because it was judged too mathematical for Physics SE. Interestingly, not only did my answer to that question use physics concepts to answer the question, but another Physics SE user commented that the question should be here rather than on Math.
I think this illustrates the point I was trying to make in the meta post linked above. Because the question was answered with physics, the nature of the information contained therein is relevant and interesting to other users of this site. Certainly one could have been given a purely mathematical answer, but a physics one was given and was deemed helpful.
Q: Should that question be migrated back to Physics SE? This suggests the further question: should we allow posts to accrue some answers before migrating/holding/closing them in order to see if physics will emerge where we may have expected pure math?
EDIT: Apparently the original reason for that question to be on Math SE was that the OP cross posted it on physics and math.