''Self-organized'' formulation of standard percolation phenomena

We consider standard percolation processes such as epidemic processes with or without immunization. We show that their dynamics can be formulated so that they mimic self-organized critical phenomena: the wetting probability p needs not to be fine tuned to its critical value p(c) in order to arrive at criticality, but it rather emerges as a singularity in some time-dependent distribution. On the one hand, this casts doubts on the significance of self-organized as opposed to ordinary criticality. On the other hand, it suggests very efficient algorithms where percolation problems are studied at several values of p in a single run. As an example, we apply such an algorithm to directed percolation in 2 + 1 dimensions, where it allows a very precise determination of critical behavior.