This article introduces a liberal alternative to the well-known decision mo
dels under risk in the form of a rule-based fuzzy decision model. The struc
ture of the model corresponds to a simple, knowledge-based (expert-)system,
which is capable to justify its assumptions and conclusions. On the basis
of empirical data the computer program models the inference processes of hu
man decision makers in fuzzy-categories and rubs: of decision. The model cl
aims to reproduce decision processes with risky alternatives more adequate
than conventional models. In comparison with human test subjects the comput
er-based decisions produced strong correspondences with regard to the choic
e of the same alternatives.