To evaluate a probabilistic generalization we need to do more than to find one or two counterexamples. Which of the following questions should be asked when evaluating the logical strength of a probabilistic generalization?a) Can I think of at least three possible counterexamples?
b) is the generalization consistent with common sense?
c) were enough cases considered?
d) what are the implications of the generalizations, if it is true?

Probabilistic generalization is proposed in which possibilities on formal contexts is considered; probabilistic formal concepts that have projective force are distinct: nonclassified objects can be allocated to an initially found probabilistic formal concepts; the random features are eradicated from probabilistic formal concepts; probabilistic formal concepts are also very encompassing in terms to data noise. The outcome of experiment is presented in which the formal concepts (in their regular definition in FCA) are first altered by random noise and then improved by identifying probabilistic formal concepts.