JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
Back to the library
Network Science 1960

On the Evolution of Random Graphs

Paul Erdős & Alfréd Rényi

Scatter links at random and, at one precise instant, a giant web snaps into being.

Choose your version
In depth · the introduction

Keep adding random connections to a crowd of strangers and, at one sharp moment, the scattered little groups fuse into a single web that holds almost everyone.

The idea, unpacked

Take a large number of dots and start joining random pairs with lines, one after another. At first you get many tiny, separate clusters. For a long while nothing dramatic happens — the clusters grow only slowly. Then, at a precise tipping point, something sudden occurs: one cluster abruptly outgrows all the others and links up a large fraction of every dot in the whole picture.

Erdős and Rényi found exactly where that tipping point is. The magic number is one connection per dot, on average. Below an average of one link each, the network is a dust of small fragments. Cross that line and a single 'giant' web snaps into being. The change is not gradual; it is a threshold, like water turning to ice at a single temperature.

Where it came from

Paul Erdős — the wandering Hungarian who owned almost nothing, slept on colleagues' sofas, and wrote papers with more co-authors than anyone in history — teamed up with Alfréd Rényi, who liked to say that a mathematician is a machine for turning coffee into theorems. In a burst of papers around 1959–1960 they asked a deceptively simple question: what does a typical graph look like? By imagining the graph growing edge by edge, they discovered that its character changes in sudden jumps, and wrote the first natural history of these random networks.

Why it mattered

Before them, graphs were studied one at a time, by hand. Erdős and Rényi showed you could reason about a whole random ensemble at once and predict its features with near-certainty for large sizes. And their tipping-point idea turned out to be everywhere: the same mathematics describes when a disease tips into an epidemic, when a rumour goes viral, when a power grid or the internet becomes robustly connected, and when a material suddenly starts to conduct. They had found a law of connection itself.

A way to picture it

Think of popcorn kernels scattered across a pan, each able to stick to a neighbour. Sprinkle sticky links at random and you get a few small clumps. Keep going and the clumps stay small — until, at one critical amount of stickiness, the next few links bridge the clumps into a single sheet that spans the whole pan. One moment: loose crumbs. The next: one connected mass. The crossing happens just when each kernel has, on average, a single link.

Seventy dots in a disk joined by random lines. A slider sets the average number of links per dot. Below one link each, only small separate clusters form. As the average crosses one, a single giant connected group lights up in colour and rapidly takes in most of the dots.

Where it sits

This is the founding paper of network science. It stands beside percolation theory in physics and epidemic models in biology, which describe the very same sudden tipping point. Later researchers — Watts and Strogatz on 'small worlds', Barabási and Albert on hubs — built richer models, because real networks (friendships, the web, neurons) are not perfectly random. But every one of them measures its network against the Erdős–Rényi random graph: the baseline of pure chance.

The original document
Original source text
P. Erdős & A. Rényi · Publ. Math. Inst. Hungar. Acad. Sci. 5 (1960) · pp. 17–61 · dedicated to Professor P. Turán on his 50th birthday
The model
For n labelled vertices there are C(n,2) = n(n−1)/2 possible edges. Erdős and Rényi study the graph Γ(n,N) formed by choosing N of those edges at random, every such graph being equally likely. They let N grow with n and ask how a typical graph looks at each stage — treating the steadily rising N as a kind of time, so that the graph 'evolves'.
Threshold functions
Their central discovery is that almost every property appears not gradually but at a sharp threshold. For many properties A there is a threshold function: if N grows much more slowly than it, almost no graph has A; if N grows much faster, almost every graph has A. The switch is abrupt, and most of the paper computes these thresholds, property by property.
The double jump — the giant component
The most famous threshold governs the largest connected component, near N ≈ n/2 — that is, an average of about one edge per vertex. Below it (N = cn/2 with c < 1) every component is small, of order log n. Exactly at c = 1 the largest is of order n^(2/3). Above it (c > 1) a single 'giant' component of order n emerges and absorbs a finite fraction of all vertices, while every other component stays small. Erdős and Rényi named this sudden change the 'double jump'.
[ … ]
Connectivity and a whole natural history
Pushing N higher, isolated vertices are the last obstacle to connectivity: around N ≈ (n/2)·ln n the last isolated vertex is absorbed and the graph becomes connected — almost exactly when its minimum degree first reaches 1. Isolated trees appear first (the larger the tree, the later it shows up), then the first cycles arrive together with the giant near mean degree one. The paper charts this whole sequence of thresholds, giving the first detailed natural history of the random graph.
Budapest · 1960