An experimentally well-supported model, which realistically describes particle behaviour in HGMS, was developed more than a decade ago. The model assumed that HGMS is essentially a deep bed filter onto which an external magnetic field is imposed. In this present contribution, the model is revisited. It is shown that under suitable conditions the collision efficiency of particles with the matrix approaches unity, even in the absence of the magnetic traction force. It is further shown that the efficiency of capture is determined by the interplay of the static dipolar magnetic interaction and hydrodynamic shear stress. The threshold magnetic induction required for particles to be retained by the matrix is derived. Experimental data are used to demonstrate that there is a limited incentive to operate HGMS at a field higher than 2 T Moreover, it is shown that undue emphasis on the magnetic field strength usually results in poor metallurgical performance. Excessively high capital and operating costs are an additional consequence of such an approach.