In spite of forty years of research, it is still unclear how the mechanical properties of paper, particularly strength, depend on the disordered geometry of the fibrous network. Most of our understanding of the fracture phenomena of paper is based on illustrative microscopic observations. Theoretical models have traditionally been focused on the behaviour of a typical element in the network. However, the failure process seldom starts from, or proceeds through, “typical” elements. Instead, the statistical distribution of local failures is crucial for the strength of paper. With the ever more powerful computers it is now possible to simulate numerically the behaviour of disordered systems such as paper. I believe that computer simulations, in combination with new measurements and effective data analysis will lead to a better understanding and more accurate microscopic characterization of the strength and fracture properties of paper.