The contributions into the disruption radial force on the tokamak vacuum vessel wall are calculated and analyzed. One is due to the induced toroidal current in the wall, and another is due to the poloidal current. The latter is not accounted for in the models that represent the wall as a set of isolated toroidal filaments. It is shown that such modeling must lead to significant errors in the evaluation of the force during either thermal or current quench. The analytical derivations are performed here for an arbitrary tokamak configuration with final estimates for a circular large-aspect-ratio plasma and a coaxial wall reacting on perturbations as a perfect conductor. The results are compared with those recently obtained numerically by the codes DINA, MAXFEA and CarMa0NL. The discrepancies between the DINA simulations (Khayrutdinov et al 2016 Plasma Phys. Control. Fusion58 115012) and earlier analytical predictions are explained. The recent conclusion (Villone et al 2015 Fusion Eng. Des.93 57) on the role of the disruption-induced poloidal current in the wall is confirmed and extended to a wider area.