This paper argues for the following three theses: There is a clear reason to prefer physical theories with deterministic dynamical equations: such theories are both maximally simple and maximally rich in information, since given an initial configuration of matter and the dynamical equations, the whole evolution of the configuration of matter is fixed. There is a clear way how to introduce probabilities in a deterministic physical theory, namely as answer to the question of what evolution of a specific system we (...) can reasonably expect under ignorance of its exact initial conditions. This procedure works in the same manner for both classical and quantum physics. There is no cogent reason to subscribe to an ontological commitment to the parameters that enter the dynamical equations of physics. These parameters are defined in terms of their function for the evolution of the configuration of matter, which is defined in terms of relative particle positions and their change. Granting an ontological status to them does not lead to a gain in explanation, but only to artificial problems. Against this background, I argue that there is no conflict between determinism in physics and free will, and, in general, point out the limits of science when it comes to the central metaphysical issues. (shrink)