Scott likes to give "epistemic" ratings to his essays to solve this very problem. They range from "basically complete speculation" all the way to "pretty darn sure." It would be nice if he were more consistent with them, though.