In many fields of engineering, frequency (statistical) interpretation of the concept of probability is widely used. At the same time, experts often forget that such an interpretation contains appropriate restrictions, which can not lead to distorted results. An important problem that arises when using frequency probability interpretation is to check the correctness of the results. In the general case, neither the verification procedure nor the procedure of falsification can not lead to an answer to this question. If we consider the classes of events as sets of finite power, then the probabilistic conclusion would directly indicate the arithmetic ratio of powers. This conclusion could be checked by simple representations of all elements of both classes. If we consider endless classes of events, or an infinite number of finite statistical sequences, then no finite number of observations can fully confirm or completely discard the probabilistic conclusion. It can not be ruled out that this final sequence of observations is a fluctuation, a deviation of the relative frequency in a given sequence from the relative frequency in the entire infinite class of the analyzed events. In this regard, it should be noted that the practice adopted by specialists in all circumstances to exclude from consideration the results of individual observations that significantly differ from the results of the bulk of observations carried out within the framework of this study can not be considered flawless. This exception procedure should be applied with caution, applying it only after a careful analysis of the situation, after identifying the reasons for the specified distinction.