Ask any banana in the bunch how many neighbors it touches, then ask its neighbors the same. On average, your neighbors have more neighbors than you do. It sounds impossible — how can almost everyone be below average? — but it’s true in nearly every social network, and it’s why your friends seem more popular than you. The reason: popular nodes get counted as someone’s friend far more often. Public-health teams exploit it to catch outbreaks early. Slide to compare.
The friendship paradox (Feld, 1991): in almost any network, the average degree of a randomly chosen NODE is less than the average degree of a randomly chosen node’s NEIGHBOR. Formally, the mean degree of a random endpoint of a random edge is ⟨k²⟩/⟨k⟩ ≥ ⟨k⟩, with equality only if every node has identical degree — the gap is Var(k)/⟨k⟩. High-degree nodes appear in many friends’ lists, biasing the ‘friend’ sample upward. It powers sensor strategies: monitoring random people’s friends detects epidemics earlier than monitoring random people. A fail-loud self-check throws unless the mean neighbor-degree is ≥ the mean node-degree on a test graph. ◆ real network science, node-verified.
The inequality is exact and holds for EVERY network with any degree variation; what varies is the SIZE of the gap (= variance/mean). It compares averages — it does not say every individual is below their friends.