The article states that the problem is that they looked at random galaxies, 1/3 were moving one way and 2/3 the other.

If the distribution was truly random, and the universe is isotropic, we should see roughly 50/50.