The main results from simulations of two different model systems, the soft‐sphere alloy and a polydisperse Yukawa fluid that has been studied with both Newtonian and Brownian dynamics, in the supercooled re´gime, are recalled. In particular it is emphasized that the only well established qualitative change resembling a ‘kinetic glass transision’ is a crossover in the microscopic self‐diffusion mechanism. One then briefly presents the similarities and differences with the predictions of mode‐coupling theory: although the agreement is fair for intermediate relaxation times, it seems that the critical laws are never clearly seen in numerical results, and that the theory provides a poor quantitative estimate of the thermodynamic coefficients of the glass.