A three-dimensional parallel algorithm of the Monte Carlo method (ELSHOW) is developed for simulation of electron avalanches in gases. The parallel implementation is performed with the use of the wellproven PARMONC library, which accelerates the calculations of such integral characteristics as the number of particles in an avalanche, the coe cient of impact ionization, the drift velocity, and the others, as well as ways to select an appropriate size of the time step using the technique of dependent statistical tests. The algorithm consists of special methods of distribution modelling, a lexicographic implementation scheme for 'branching' of trajectories, a 'Russian roulette', justi ed construction of histograms and a polygon of frequencies, and the calculation of probabilistic error estimate of functionals. A comparison of the obtained results for nitrogen with previously published theoretical and experimental data is presented.The development of computer technology brings the demand for software working correctly in the sixdimensional phase space of coordinates and velocities (x, y, z, Vx, Vy, Vz), in particular, for simulation of breakdown in gases. The base for such calculations is a basic algorithm calculating not the proper eld of electrons, but taking into account only the external eld. This model is applicable in the beginning of breakdown development when the number of electrons is small and their own eld is small compared to the external one.Simulating the motion of electrons (electron avalanche) in gases, one has to solve Boltzmann's equation with a power component. To do that, we use the Monte Carlo method [1, 3]. There are algorithms solving Boltzmann's equation by other methods (see, e.g., [4,8]).The Monte Carlo method allows us to simulate in a natural way such complex and improbable processes as a permanent acceleration of electrons in a gas, which are di cult to consider in other approaches. With all advantages of this method, it is necessary to pay a special attention to the fact that one has to store in computer memory the coordinates of a six-dimensional phase space for all the electrons of an avalanche, while the number of these electrons grows exponentially in time [9]. This problem can be partly resolved by the well-known lexicographic scheme of trajectory 'branching' (see Section 8 below) and the 'Russian roulette' (Section 1). However, a su cient gain in computation time can be obtained in practice due to a parallelization technology.The Monte Carlo method was also used previously to solve such problems. The most interesting for comparison algorithm was presented in [10]. It di ers in another set of cross-sections and in the application of the maximum cross-section method for simulation of a run. A similar algorithm for argon and oxygen in combination with the PIC-method was proposed in [23].The Monte Carlo algorithm presented here was implemented in the software code ELSHOW (ELectron SHOWer). In contrast with other algorithms, it uses a lexicographic scheme, a parallelization technology, ...