Sandpile — Self-Organized Criticality
Add grains of sand to a square grid. When any cell reaches 4 grains, it topples: 1 grain to each of 4 neighbors. Repeat. The system self-organizes to a critical state where avalanches of every size occur — power-law distributed. Bak-Tang-Wiesenfeld 1987. The same dynamics describe earthquakes, neuronal cascades, forest fires, and financial crashes.
Sandpile grid (drop sand anywhere)
Avalanche size distribution (log-log)
why power laws emerge without tuning
Standard physics has tuned critical points: water boils at exactly 100°C, magnets demagnetize at exactly the Curie temperature. Tiny parameter changes destroy the criticality.
Self-organized criticality (Bak, Tang, Wiesenfeld 1987): some systems find the critical state by themselves, without external tuning. The sandpile is the canonical example:
• Add a grain → cell may topple → neighbors may topple → cascading avalanche
• Open boundaries → sand falls off edge, total can stay finite
• System self-organizes to a state where avalanche size P(s) ∝ s^(-τ) with τ ≈ 1.0 for 2D
Power-law distribution: most avalanches involve 1-2 cells. Rare ones span the entire grid. There is no characteristic scale. The histogram on a log-log plot is a straight line.
Where SOC shows up:
• Gutenberg-Richter law: earthquake magnitudes follow a power law (with τ ≈ 1.0) — same as the sandpile
• Neuronal avalanches: cortical activity in rats shows power-law-distributed cascade sizes
• Forest fires: fire size frequencies match the same scaling
• Stock market crashes: returns have power-law tails (fat-tailed risk)
• Solar flares, extinction events, traffic jams, biological evolution (punctuated equilibrium)
The unifying claim: many complex systems are poised at the edge of stability. Small perturbations sometimes cascade catastrophically — and the distribution of catastrophes is universal, independent of microscopic details. Watching the BTW sandpile is watching this universality in action.