Because that is the grounding assumption on which all existence is based. The point I'm trying to make is that we must trust our senses entirely, or not at all. As your example perfectly demonstrates, the latter is an unteneble position, which is what Schopenhauer meant when he said that
As a serious conviction, on the other hand, it could be found only in a madhouse; as such it would then need not so much a refutation as a cure.
By the "look out your window" argument a desert dweller who has never been out of the desert, must believe the rest of the world is a desert. Someone growing up on a tiny tropical island would have to believe the entire world is covered with water, and constantly warm. Both views are not only incompatible, but incorrect. Each could be presented with first hand accounts from others that there are other terrains and climates as well as pictures to indicate as such. Books, magazines, pictures, newspapers, would all contain accounts to show that there is more out there than what they have seen. Would it be more mad of them to assume that the evidence, overwhelming evidence, is in fact true, or that the presenters of the information are part of some giant conspiracy there to deceive them?
Science has allowed us to advance in ways people living now could have never imagined when they were young. Global commerce, international air travel, satellite radio, satellite tv, satellite phones, the space program, moon landings, Antarctic colonization, space exploration, close up studies of the sun and other planets in the solar system, even colonization of space. Is it more rational to trust one?s senses on things that we can see be it first hand evidence, through pictures and accounts or inferred based on logic, or to assume it is all a part of some giant conspiracy aimed at tricking us?
While the whole philosophical "you don't know if you're awake or sleeping" argument is out there, I trust that my senses have gathered enough evidence to refute that. If I had a clue where the paper I wrote on this was I'd scan it in. My closing argument was basically: Is it possible? Yes. It is probable? No