Classical force fields are not only sets of equations with which to scare students. They are normally a model for the physics of a system like a protein molecule. Sometimes, however, one is only interested in some aspects of a protein, such as its structure or composition. In that case, it may be better to build force fields or score functions optimised for that goal. If one is willing to forget physics completely, one can treat force field construction as an optimisation problem rather than a question of real physics. This philosophy can be taken further and even applied to ideas from protein sequence analysis. That field is rooted in ideas from evolution, but by treating the parameterisation as an optimisation problem, one can completely avoid the use of evolutionary models.