Rewiring a few links in a regular lattice collapses path length while leaving clustering intact, which is why real networks are both clustered and short-pathed.
Watts and Strogatz's 1998 model, and the formal counterpart to Milgram's empirical finding. A few long-range links do almost all the work of shortening paths.
In practice
- A firm's departments each talk mostly among themselves, and the three people who sit on two committees are why a question can cross the building in a day.
- A conference that flies in six outsiders changes more connections than a year of internal away-days, because the internal links were already dense.
- Standing
- The formal model behind Milgram's empirical small-world finding.
- Subject
- Networks & Collective Behavior
- Essay
- How connections shape what crowds do