Instead of working with z-scores, which correspond to weirdl...

Prompt

Instead of working with z-scores, which correspond to weirdly increasing values of x (1 in x0, 1 in x1, 1 in x2 for z=0, 1, 2, where "1 in x" is the rarity), why don't we define a k-score that goes like this: k=0 is 1 in y0 rarity k=1 is 1 in y1 rarity k=2 is 1 in y2 rarity y2/y1 should be y1/y0, shouldn't it? But for z-scores, x2/x1 > x1/x0 and keeps increasing—why? I am asking for the philosophy behind this architectural choice.