@neologist,
You and I have been over that and found no common ground. The way that you interpret the "evidence" of seeing stars in the sky, for example, isn't falsifiable. You look at the stars and infer a god. Others don't. By itself, it's not more evidence for than against, so it's not evidence at all.
If you go further and do experiments to measure the distances to the stars, their composition, velocity, mass, etc, you accumulate a body of evidence that still doens't point to a supernatural explanation. The more data you have, and the more complete your mundane cosmological model, the less probable the supramundane explanation becomes. Being inferential, it's an asymptotic approach to zero, so it's out of line to declare that it's 100% sure that there's no god. However, it's also out of line to try to wedge a god into that ever-decreasing gap.