By statistical reliability of models (SRM) of random properties of a medium based on the Monte Carlo method we mean the attainment of a state of affairs in which: a) real tests would furnish us with a statistic of the distribution of properties adequate to the statistic given by the process of simulated tests ("play') ; b) a calculation based on a Monte Carlo model of the final (or if necessary the intermediate) stages of a process in which randomness of certain properties of the medium is manifested will give, with the desired precision, the real characteristics of the stages of the process which are realized in experiment. This division of requirements has a definite foundation.The point is that some characteristics of a continuous medium (modulus of elasticity, strength, density, etc.) must be assessed on the basis of the results of tests on specimens representing the material of the medium. For this reason, the frequency of attributing to the material of the medium in the model cell any particular value of a sampied quantity [1] must not only reflect the frequency of realization of this value in tests on specimens (which might be a sufficient requirement to ensure statistical reliability of modeling), but must also take account of the fact that the statistics of the results of tests of the properties of specimens drawn from the medium may depend both on the size and shape of the specimens and on the fact that the specimen test conditions will usually not correspond to the conditions in which the element of the medium occurs in the real process. In the latter case we are faced with the problem of the scale factor, as well as with the problem of getting the specimen test conditions the same as those realized in the process under investigation.Both these problems leave their mark on the requirements imposed on the size of the model cell and on the process of sampling ('play') itself. This is taken into account in the formulation of the determinant condition for which we get reliability of our chosen Monte Carlo model of the random properties of a continuous medium. The question arises of whether we can satisfy these conditions. In fact, on the one hand, we acknowledge the insufficiency of the first condition; and on the other, uncertainty appears in the possibility of obtaining information, on the bases of the results of tests of the properties of a continuous medium (either on specimens or in situ), suitable for use in calculations on various processes occurring in it. Consequently, the question of the size of the model ceil, regarded by us as crucial for satisfaction of the conditions for SRM, must be decided separately for each class of problem. As an illustration of one possible approach to the determination of the size of the model ceil, let us consider the process of wave loading of a continuous medium with formation of cracks. For convenience we shall denote this problem by the letter S. As a random property of the medium in this process we shall take its tensile strength o ~ [2]. We shall regard the mate...