For amusement, one could calculate the fraction of the population that has a blended exclusivity score < x (using the article’s 0.425% or any other threshold strictly between 0 and 1), and one will, of course, obtain a result greater than x. So anyone following these dating rule is risking p-hacking their mates!

In the extreme case of using many independent traits, I think one might find that most of the population is in an exclusive minority :)

[0] Use logs. If there are n extra traits that aren’t in the original three, a random person scores some expected log exclusivity points for each extra trait. (Exactly 1 point, in fact, if you use natural logs and negate the result to get positive points.) Multiply that expectation by the number of traits and the average person scores n points. e^-n is very exclusive for even moderate n!