Indeed galaxies are bound by gravity
But what makes galaxies in the first place? Why are there many and not just one conglomeration of stars, all tightly packed together? If this is random (in fact, it is kinda a scientific fact), you have to accept that patterns are a distinctive feature or property of randomness. The Universe has innumerable billions of stars, and that's more than enough to evaluate its properties
Universe is expanding and has been doing so since the Big Bang. So it can't be in one big clump. It would probably be a very neatly organized sparse cloud of particles if not for randomness, which caused it to stick into various blobs
Should it be construed in the way you think that the distribution of galaxies in the Universe is not random?
Besides, you can't have it any other way from a purely mathematical point of view (the approach which you seem to be particularly fond of). How come? The reason is simple. If it were not for patterns, you would have a uniform distribution which is not random by definition, as simple as it gets. Stated differently, you can't have a random distribution without patterns given sufficient sample size
Uniform distribution is a very distinctive pattern
Indeed it is patterned, but it has only one pattern, while in a random distribution you can see plenty of them. They are random too, but it is exactly their randomness (which leads to them being many) that distinguishes the latter from the former. If you remove these patterns, you will get a uniform distribution, which violates the assumption of randomness (though the opposite is not necessarily true). So any way you twist the semantics of it, a random distribution without patterns (as in plural) turns into a uniform distribution (with just one repetitive pattern), and thus stops being random
So it's the other way round. Random doesn't stop being random because you spotted a pattern. It either wasn't random to begin with, or you're wrong and there is no pattern
There's not only one pattern in a random distribution as there are many, and their very existence makes a random distribution somewhat less random, from a practical point of view (superstitions or otherwise)