Climate models are complicated things. Their construction often begins with off-the-shelf physical equations, but these equations are always approximated, and typically the approximation is rough. For us to trust a climate model and its results, it must earn its credentials. There are at least two ways in which a model can earn these credentials: fit with the data and theoretical pedigree. For Eric Winsberg, although fit is preferred, we may need to rely on pedigree when data is scarce. In this paper, I argue that theoretical pedigree is a bad way for climate models to earn their credentials and that we should be somewhat more optimistic about the prospects of fit.