I don't think you can separate it that way, though - not really. There are vast numbers of potential ways in which history would have worked out differently if Hitler had been killed; as has been pointed out, the possibilities are even quite different depending on when exactly you kill him. But although we can propose various scenarios, I don't believe anyone can make an honest claim that any one of them is the most likely of all, at any one point. And even if someone could, it matters little: history is replete with unlikely eventualities and purely chance circumstances.
As a result, nobody can really answer the question of "would it be ethical to go back in time and kill Hitler" by appealing to any individual proposed scenario: "No, it would not be ethical, because [specific situation X] could possibly result", because that could simply be sidestepped with "what if additional things could be done to mitigate that specific situation", which leads to additional lines of calculation and turtles all the way down. Rather, the answer would seem to be "No, it would not be ethical, because an indeterminate number of equally-negative or more-negative situations could result, and there's no way to predict the likelihood of any of them versus potential positive outcomes". And that IS a conclusion that could be applied equally to any proposed changing-history-via-time-travel scenario, no matter the particulars.