In the last few decades optically stimulated luminescence (OSL) dating has become an important tool in geochronological studies. The great advantage of the method, i.e. dating the depositional age of sediments directly, can be impaired by incomplete bleaching of grains. This can result in a scattered distribution of equivalent doses (DE), leading to incorrect estimation of the depositional age. Thoroughly tested protocols as well as good data analysis with adequate statistical methods are important to overcome this problem. In this study, samples from young fluvial sand and flood plain deposits from the Elbe River in northern Germany were investigated to compare its depositional ages from different age models with well-known historical dates. Coarse grain quartz (100–200 μm and 150–250 μm) and polymineral fine grains (4–11 μm) were dated using the single aliquot regenerative (SAR) dose protocol. The paleodose (DP) was calculated from the DE data set using different approaches. Results were compared with the development of the Elbe River, which is well-documented by historical records and maps covering the last 1,000 years. Depending on the statistical approach it can be demonstrated that depositional ages significantly differ from the most likely depositional age. For the investigated coarse grain quartz samples all ages calculated from the MAM-3UL, including their uncertainties, are within the historical documented age. Results of the polymineral fine grain samples are overestimating the historically documented depositional age, indicating undetectable incomplete bleaching. This study shows the importance of using an adequate statistical approach to calculate reliable OSL ages from fluvial sediments.