> #2 is _much_ closer to #1 than #3 (let alone #4)

Even with linear scaling, being one third of the way between two numbers is not what I would call underlined-much closer. But zero punches above its weight here. Those extra orders of magnitude should make some impact on the scale.