Statistical models have two main components. First, a mathematical formula that expresses a deterministic, predictable component, for example the fitted straight line that enables us to make a prediction of a son’s height from his father’s. But the deterministic part of a model is not going to be a perfect representation of the observed world.
