The Barremian Helvetiafjellet Formation on Svalbard has been assigned two different stratigraphic/architectural models: (1) one in which an alluvial-paralic-marine succession overlies and transgressively onlaps an uplift-created unconformity, and (2) a layer-cake model with regressive, very low-gradient braided-stream deposits on the basal unconformity, passing upwards into coastal-plain, shoreline and shelf deposits without diachroneity of the constituent facies belts. The transgressive model has been criticized because its basal fluvial deposits necessarily (gradient considerations) become slightly younger sourcewards (northwest and northwards) and because the model appears to transgress landwards too steeply. The layer-cake model also has some flaws, as the synchronous basal braided-stream unit has a lateral extension landwards of some 150 km. The latter is problematic because a gradient of at least 1m/km would be expected in such a fluvial system, as seen from modern data, and the updip extent of such braided sheet sands is usually limited to 20-50 km. The layer-cake model also ignores a published, more extensive database for the Helvetiafjellet Formation, showing that there is, despite the overall vertical change from fluvial to shoreline deposits, a significant facies variability from place to place, with fluvial channels sometimes occurring high in the succession and coastal deposits near the base. The transgressive model meets this problem by having a diachronous, upward-deepening succession, as would be expected during transgression, but with significant autogenic environmental variability in time and space. The fluvial sandstone tongues in the transgressive model have less than 40 km of downdip extent and their overall diachronism predicts that timelines pass from fluvial through shoreline to shelf lithosomes. The layercake model therefore has some difficulty to explain the geometry, lateral facies relationships, gradient, grain-size variation and geomorphology of the system.