An abstract description of the RichardsonKalitkin method is given for obtaining a posteriori estimates for the proximity of the exact and found approximate solution of initial problems for ordinary differential equations (ODE). The problem Ρ{{\Rho}} is considered, the solution of which results in a real number uu. To solve this problem, a numerical method is used, that is, the set Hℝ{H\subset \mathbb{R}} and the mapping uh:Hℝ{u_h:H\to\mathbb{R}} are given, the values of which can be calculated constructively. It is assumed that 0 is a limit point of the set HH and uh{u_h} can be expanded in a convergent series in powers of h:uh=u+c1hk+...{h:u_h=u+c_1h^k+...}. In this very general situation, the RichardsonKalitkin method is formulated for obtaining estimates for uu and cc from two values of uh{u_h}. The question of using a larger number of uh{u_h} values to obtain such estimates is considered. Examples are given to illustrate the theory. It is shown that the RichardsonKalitkin approach can be successfully applied to problems that are solved not only by the finite difference method.