Huh. There's bound to be some weird constraint there. E: Looking at the sibling comments, I'm disregarding negative numbers here. Probably that breaks the line

Huh. There's bound to be some weird constraint there. E: Looking at the sibling comments, I'm disregarding negative numbers here. Probably that breaks the line of thought here. For example, a sequence made up of all numbers from 1 to n is unsolvable. A sequence of [1, 2, ..., n-1, n+1] for n >= 3 also seems to be unsolvable, since no number up to n-1 can split into a smaller number (they are all in the list) and n+1 also cannot split, because 1 is in there. So it seems that there would be an unsolvable class of lists along the lines of [1, 2, ..., k, ..., n-1, n+k] by that argument, because the n+k cannot split ever, because all numbers up to k are in that list. And I guess for any unsolvable list with a max value of M, you can add any number below M into the unsolvable list, because that just removes moves. Ah. Nerdsniped :)