But how does it explain that? The situations are clearly very different on either end of the spectrum, so you can't directly use the intuition from one end on the other. The only thing you've really shown is that the categorization will be mildly imperfect.
There's rarely such a thing as a perfect category in real life. Arbitrary categories are still categories, and still work with moral intuition, they just take slightly more care around corner cases.
Extrapolating from one case is a bad idea when categories are fuzzy. You find such things as the sorites paradox.
