1. If the chance for a huge colony to go extinct, raising the coefficient to galactic scale does not change the relative probability significantly unless you add constraints to the population dynamics. Basically you could get the same numbers for people in a colony and colonies in a galaxy. "only ONE" kinda obscures the fact that this one would be a mighty huge one.

2. On straight lines perhaps with ages long acceleration. Planning such a course might well take more than nuclear reactor, no?