Buried thrust systems, commonly analysed in terms of duplex structures, are often interpreted as forming in a particular sequence. This model considers that only one thrust is active at a time and the resulting structural geometry is well ordered. However, a reappraisal of structural geometries from the Moine Thrust Belt in NW Scotland has shown that many apparent duplex structures are more complex. Although the upper (roof) thrust of some duplexes may be folded and bulged by underlying imbricate slices (in accordance with simple duplex models), the footwall to these roof thrusts cut up and down stratigraphic section (in conflict with the simple duplex models). Examples include the Glencoul Thrust with its underlying imbricate system, and the Foinaven ‘duplex’. Simultaneous slip on an array of imbricate thrusts can bulge duplex roofs (relationships classically used to infer piggy-back sequences of thrust ‘propagation’). However, activity on more hinterland-ward imbricate thrusts can cause the roof onto which they branch to truncate more foreland-ward structures. In this fashion, the ‘duplex roof’ can cut up and down stratigraphic section, mimicking an extensional fault. By varying thrust activity in three dimensions within an imbricate system, local parts of ‘roof thrusts’ can generate domains of overstep behaviour that have only local significance. Existing interpretations of overstep relationships in the Moine Thrust Belt as representing late-orogenic low-angle extensional faults may need revision. The new interpretations of deformation activity, where displacements in thrust arrays happen synchronously rather than sequentially, are consistent with emergent thrust and fold belts, with analogue experiments and with mechanical models of orogenic wedges.