I think that could be possible. If sexuality were multi-dimensional and “gayness” was just a 1-D collapse of a higher dimensional space then you could pick a vector in the higher dimensional space to represent gayness, such that a few points at the extreme happen to have the same dot-product with that vector.
But then you would be defining gayness around the gymnastics of setting that up instead of something you are actually trying to estimate about people on that spectrum.
You have ignored the possibility that even though there are finite gay people, among those finite gays one of them might be infini-gay.