Small-World Networks: Six Degrees of Separation Explained

simulator intermediate ~7 min
Loading simulation...
Small World: N=60, K=4, p=0.1

With 60 nodes, 4 neighbors, and rewiring probability 0.1, the Watts-Strogatz model produces a network in the 'small-world' regime: average path length drops dramatically while clustering remains high.

Formula

L(p) drops sharply for small p
C(p) stays high until p approaches 1

Six Degrees of Separation

In 1967, Stanley Milgram conducted his famous experiment: he asked people in Nebraska to forward a letter to a target person in Boston, passing it only through personal acquaintances. The letters that arrived took an average of just six steps. This 'six degrees of separation' phenomenon puzzled scientists for decades — how can a world of billions be so small?

The Watts-Strogatz Breakthrough

In 1998, Duncan Watts and Steven Strogatz cracked the puzzle. They showed that you can start with a perfectly regular network — a ring where everyone knows only their nearest neighbors — and by randomly rewiring just a tiny fraction of connections, you get a network with both short path lengths (like a random network) and high clustering (like a regular lattice). This 'small-world' regime exists for a broad range of rewiring probabilities, roughly 0.01 < p < 0.5.

Why It Works

The key insight is that long-range shortcuts are disproportionately valuable. In a ring lattice, reaching the opposite side requires traversing half the network. A single random shortcut to the other side cuts the path length in half. A few dozen such shortcuts reduce the average path length from O(N) to O(log N), while local clustering barely changes because most edges remain local.

Try It Yourself

Start with p=0 (regular ring lattice) and slowly increase it. Watch how the cyan rewired edges create shortcuts across the circle. Enable 'Show shortest path' to see how the path between opposite nodes shrinks dramatically. The inset chart shows how L(p) drops while C(p) stays high — the hallmark of the small-world effect.

FAQ

What is a small-world network?

A small-world network is a network that combines two properties: high clustering (your friends tend to know each other) and short average path lengths (any two nodes can be reached in surprisingly few steps). This mirrors the 'six degrees of separation' phenomenon observed in real social networks.

What is the Watts-Strogatz model?

Proposed by Duncan Watts and Steven Strogatz in 1998, this model starts with a regular ring lattice where each node connects to its K nearest neighbors. Each edge is then randomly rewired with probability p. Even very small p (around 0.01-0.1) dramatically reduces path lengths while preserving high clustering.

What is the clustering coefficient?

The clustering coefficient measures how interconnected a node's neighbors are. If all your friends know each other, your clustering coefficient is 1. In real social networks, clustering is typically 10-100x higher than in random networks of the same size.

Why do small-world networks matter?

Small-world properties appear in neural networks, power grids, social networks, and metabolic networks. They enable efficient information transfer (short paths) while maintaining local redundancy (high clustering). This architecture is optimal for many biological and social systems.

Sources

Embed

<iframe src="https://homo-deus.com/lab/network-science/small-world/embed" width="100%" height="400" frameborder="0"></iframe>
View source on GitHub