Across most rifted margins, the extension measured from fault geometries is far less than that required to explain whole crustal and lithospheric thinning. This ‘extension discrepancy’ is commonly explained through crustal depth-dependent stretching (DDS – perhaps better termed depth-dependent thinning, DDT). However, several independent lines of evidence (velocity structure, rheological modelling, reconstructions, ODP drilling results) show that the amount of DDT required to explain the extension discrepancy cannot have occurred. Instead, it is suggested that as extension increases, complex geometries arise which are not completely interpreted, leading to a massive underestimation of the amount of extension. The implications are that pre-rift and early syn-rift reservoir and source rocks are likely to be widely scattered across deep margins. In the absence of massive crustal DDT, the deficit in syn-rift subsidence observed at some margins can be explained by thermal and dynamic uplift, igneous addition or mantle serpentinization during rifting. But it is also possible that syn-rift subsidence has been systematically underestimated if local water level was substantially below global sea-level, as indicated at some margins by the formation of thick evaporites at the end of rifting.