There exist a variety of methods for extracting weights and values in multi-criteria decisions based on different rankings. However, it is difficult to determine which the useful ones are and how they correspond to the decision makers’ perceptions, if at all. How do we know that what we are using is really significant, especially in situations when the decision bases are vague? One category of methods that has proved relatively successful is to use so-called generated surrogate weights that are, in some sense, meant to represent rankings and there are various suggestions as to how best to distil them from input information. In this paper, using a number of simulations, we review some leading algorithms for automatic weight generation without external parameters besides cardinal and ordinal rankings and provide some guidelines for selecting a surrogate weight-generating function for MCDM applications, in ordinal as well as cardinal information settings. We also propose an alternative with some attractive properties compared with the existing ones.