In any case, it doesn't actually matter because (as I said) the universe is infinitely complex and therefore its creation is infinitely unlikely.
I won't look into the details of Dawkins' specific argument, but your bit here is just weird. Even granting some arbitrary evaluation of complexity (you might be compounding with the infinite nature of ignorance), how do you evaluate the likelihood of creation (i.e. "Creation")? Should such a thing even have the property of likelihood? And in principle, why would "likelihood" be a linear function of "complexity"?