Systems composed of many particles involve a very large number of degrees of freedom, and it is most often uninteresting—not to say hopeless—to try to describe in a detailed way the microscopic state of the system. The aim of statistical physics is rather to restrict the description of the system to a few relevant macroscopic observables, and to predict the average values of these observables, or the relations between them. A standard formalism, called “equilibrium statistical physics”, has been developed for systems of physical particles having reached a statistical steady state in the absence of external drive (like heat flux or shearing forces for instance).