The Quaternary potentially provides the most ideal sedimentary record for accurate and precise radiometric dating of sequence boundaries. Three up to 60-m-deep cores from the paleovalley succession of the late Quaternary Rhine-Meuse system were obtained in a transect oriented parallel to the present Dutch coast. The succession consists predominantly of stacked units of sandy to gravelly fluvial channel-belt deposits from the last two glacials (oxygen-isotope stages (OIS) 2-4 and 6-8). Although, in theory, these units are expected to be separated by a sequence boundary, the homogeneous nature of the deposits commonly precludes unambiguous detection; such sequence boundaries have therefore been referred to as ‘cryptic’. Perhaps the most viable technique for numerical dating of such strata is Optically Stimulated Luminescence (OSL) dating that allows direct age measurement of the time of burial of quartz grains up to 150 ka, and sometimes more. A database of ~30 quartz OSL ages demonstrates that identification of the sequence boundary, associated with the relative sea-level fall of the last glacial, is far from straightforward. Only one of the cores provides an unconformity of >50 kyr that unequivocally separates strata from OIS 2-4 and OIS 6-8, as might be expected in a setting dominated by 100-kyr glacio-eustatic cycles. In the other two cores, resolving the last sequence boundary chronologically is problematic. This is most likely due to a fluvial falling-stage systems tract (OIS 4) immediately overlying transgressive estuarine channel deposits from OIS 5. Our findings suggest (1) that even the most recent glacio-eustatically controlled sequence boundary may sometimes be impossible to pinpoint despite temporal resolution of <10 kyr; (2) that tidal ravinement surfaces may be temporally much more distinct than sequence boundaries; and (3) that age relationships across sequence boundaries may change rapidly over short (km-scale or less) distances.