Book excerpt: Despite its haphazard growth, the Web hides powerful underlying regularities - from the organization of its links to the patterns found in its use by millions of users. Probabilistic modelling allows many of these regularities to be predicted on the basis of theoretical models based on statistical methodology.