Somewhere on Reddit there's a great Mike Hamburg (I think?) comment about his approach to quickly analyzing new cipher designs --- is it linear? Can a SAT solver break it? There's more to it, but not much more, and it's one of the denser crypto comments I've read.