It seems like a lot of attention in the rasterization section is spent on improving the last 5% of image quality. After all, most of the structure of the image is already there, it just doesn't look quite as good as it could.
"Good enough" is a pretty subjective criteria, though as visual creatures we have surprising amounts of acuity. Simple geometry may not be the best demonstration of the need for that "5%" - jaggies on human faces, for example, are very obvious to human eyes, and static jaggies don't even take into account video aliasing and dynamic jaggy effects.
Of course, this question boils down to an engineering one, and if we can get away with 60fps, or if 95% is really good enough, then so be it! Good to point out the engineering constraints we need to consider, though.