I do not think it matters how unlikely it is. As long as it is possible and there is no surefire culprit the theory needs to be tested, if only to rule it out.
"Probable" and "unlikely" are antonyms. It can't be unlikely and probable.
edit: The level of likeliness should affect the amount of effort put into testing the hypothesis. I'd like an explanation from them on why they think the Wow! signal is so rare if a comet is the explanation - NASA says there might be as many as a hundred billion comets in the solar system, so "where's the rest of their signals" is a legit criticism.
I think it's important to remember how empty space actually is, when taking this view point. For instance, an object randomly shot through the asteroid belt has near zero chance of actually hitting anything, and that's a relatively dense space. The Oort cloud, which is where a large number of the theoretically unknown comets live, is only about 5 Earth masses.
The bottom line is, pointing a telescope to a random spot in the sky is very, very likely to find nothing. Even with a billion comets and a billion asteroids and a billion stars.
Because the universe is expanding. Space itself expands the further away you are from a reference point. That means, everything is moving away from us (on a grand scale), and the greater the distance to the reference point, the bigger is the rate things in the universe are moving away from us. I think the rate is about 70 km/(Mpcs) - wich is an incredibly small number, scaled down to a meter it is ~10^-23 m / (ms). And as space expands over time, the density of matter in space get lower and lower.
Space isn't an empty vacuum, but that's tangential. The expansion in this context is metric explanation, which just means things are getting further and further apart between two points due to space "expanding" between non-gravitationally bound objects. The explanation for "why" is sort of "we don't know". AFAIK, the leading theory is a combination of momentum / inertia from the big bang and acceleration from dark energy. Which we also don't know a lot about.
Space was less empty when it was younger, so I suppose you could say that the answer to your question is "because it's over 13 billion years old."
But really, "why" questions like this come with too many hidden assumptions to be helpful. You need to clarify why you expect it to not be empty. The way I see it, you could just as easily ask why is space so full of stuff as it is.
#2) as far as we know, it is a lot less empty than it seems to be. We're missing about 85% of the mass in the universe (https://en.wikipedia.org/wiki/Dark_matter), and about 95% of its mass-energy.
So, the best theory we have is, in some sense, awfully inaccurate, at least until we invent some way to see the missing stuff.
The level of likeliness/probability for a given explanation depends crucially upon the level of likeliness/probability for all other explanations. This is the fatal flaw of research using p-values to assess evidence for a given theory. https://en.wikipedia.org/wiki/Bayes%27_theorem
