We already observed it; it's too late to say it is too improbable. A particle from a star that exploded billions of years ago just went into my glass of water. But, if we ask the question before our sun blows up if a specific particle from it will land on the highest point of a planet 50 light years away, then we could say it wouldn't happen. So I don't know why problems like the Boltzmann brain are actually problems. Any starting configuration, ordered or not, from a quantum fluctuation has extremely improbable odds, 1/(the number of possible beginnings). But it already happened, so the probability is 1. What am I not understanding here?