But getting down to the scale of a snowball cooling without anything *bigger* being affected is difficult. Hard to say it's impossible, but seems very tricky. So in the case that dark matter is a bunch of macro-scale objects, you'd expect to see deviations in larger gravitationally-bound structures.