The street network's angular properties were found more suitable than metric properties for capturing the observed pedestrian and vehicle movement flows in space syntax modeling. Some studies relate this state to the underlying street network structure that create the potential for movement across the network. The aim of this paper is to clarify why the angular structure of the network has superiority over the metric structure. The investigation entailed analysis of street network' centralities and movement flows obtained through agent-based simulations conducted for two cities that differ in the pattern and size of street network. The findings indicate that the superiority of the angular structure can be explained by two structural properties: (i) a multi-scale correlation between to-movement and through-movement potentials (centrality measures) of the same distance type; and (ii) an overlap between movement potentials of different distance types across scales of the network. These structural properties create coherent and dominant angular foreground structures that fit movement flows in both study cities.