Geometry for the selfish herd

Every animal in a herd is trying to get to the center. Not for warmth, not for social bonding, not for the good of the group—but to put other animals between itself and the predator. The herd looks cooperative. It's actually mutual exploitation.

Hamilton formalized this insight mathematically using 'domains of danger'—Voronoi diagrams where every point within a polygon is closer to one individual than to any other. Animals on the edge have larger domains; predators attack the closest prey. The geometry is unforgiving: edge animals die first. Natural selection ruthlessly favors individuals who minimize their domain by moving toward the center, even if this pushes others outward to their deaths.

The paper demolished the prevailing view that herding evolved for group benefit. Instead, Hamilton showed herding emerges from purely selfish individual optimization. No cooperation required. No group selection needed. Just geometry and self-interest producing what looks like coordination.

For organizations, this reframes clustering behavior—companies co-locating in Silicon Valley, banks concentrating on Wall Street, law firms sharing buildings. The explanation isn't 'ecosystem synergy.' It's domain of danger minimization. Being in the cluster reduces individual risk, even if the cluster itself creates no value. The herd protects not because it cooperates, but because each member uses others as cover.