The carried out analysis of stochastic identification models has allowed to find a rigorous mathematical proof of their inefficiency. Identification of object parameters and characteristics of noise using a uniform model is impossible. It leads to erroneous results which have abstract character and cannot be decoded by using any mathematical operations. These facts have been held back many years and practical identification was in impasse. Partly it is because finding analytical solution of system of difference equations is not possible using traditional computing methods. The obtained results show that procedures of identification should be complete and should include algorithms for decoding the coefficients of difference equations.