This article analyzes the implications of basic lottery tests for the proba
bility weighting function w(p). We first show that the w(p) function with o
ne argument cannot accommodate three basic tests of lottery choice. We also
discuss in detail the links between the w(p) function with one argument an
d the preference reversal paradox. In the last section we propose an evalua
tion function that accommodates the restrictions imposed by the tests and w
e provide an example of such function.