Strange Attractors

What you're looking at

Lorenz's discovery

In 1963 Edward Lorenz was running a crude weather simulation on a Royal McBee computer. To save time he restarted a run partway through, retyping the initial conditions from a printout. The printout showed three decimal places; the computer used six. That tiny rounding — 0.506 instead of 0.506127 — produced a completely different forecast. Same equations, different weather.

Deterministic but unpredictable

There's no randomness here. Every trajectory is fully determined by its starting point. But infinitesimal differences in initial conditions get amplified exponentially. After enough time, two nearby trajectories look completely unrelated. That's the core of chaos — deterministic yet practically unpredictable.

The butterfly effect

The pop version says "a butterfly flaps its wings and causes a tornado." The actual point is subtler: in chaotic systems, you can't separate "significant" causes from "insignificant" ones. Every perturbation matters. Long-range prediction isn't just hard — it's fundamentally limited, no matter how good your measurements get.

Strange means fractal

The "strange" in "strange attractor" means it has fractal structure — a non-integer dimension. The Lorenz attractor has a dimension of about 2.06. The trajectory never crosses itself, never repeats, but stays confined to this thin fractal sheet. It's bounded but infinite, structured but never periodic.