Most futurological approaches that require a quantitative assessment of scenarios proceed in two stages: a qualitative exploration of the determinants to begin with, followed by the choice of a limited number of hypotheses, which in turn lead to a limited number of scenarios. Next, comes the application of mathematical models that by means of data reflect the hypotheses adopted to operationalize the various identified scenarios. Despite the near ubiquity of this model, the increasing complexity of contexts and possible interactions of the various factors involved in the fields of study are challenging its mode of operation. The paper proposes that futures foresight operate in a different way: by building databases of scenarios in order to identify areas of uncertainty. It explains the alternative method as one implemented by running models hundreds or thousands of times to ascertain possible scenarios and then conducting qualitative selection, in a second stage, in the target space of those scenarios.