Sedimentary logs form the foundation of many studies of ancient and modern sedimentary successions. In siliciclastic settings in particular, vertical grain-size trends are important records of past depositional processes and environments, so that they are recorded with care and are often central to final interpretations. It is rare for the actual grain size to play a similar role. Yet there is significant value in (i) digitizing sedimentary logs to produce statistical grain-size data and (ii) the deliberate study of grain size, a process described here as “grain-size bookkeeping.” To illustrate this, over 5.9 km of sedimentary logs have been digitized from the Miocene to Pliocene paleo–Orinoco delta, the Cretaceous Ferron Last Chance and Notom deltas, and the Jurassic Ravenscar Group. The digital data reveal how grain size partitions into distinct sedimentary facies, proximal-to-distal changes, changes related to stratigraphy and base level, and the overall grain size of paralic systems. It emerges that fluvio-distributary channels are the coarsest-grained sediment bodies in each of the studied systems. The coarsest material does not reach the shoreline, though the grain sizes of fluvio-distributary channels and shoreline sand bodies overlap, in accordance with the concept that the former feeds the later. By contrast, overbank sediments are relatively fine-grained, suggesting that, with the exception of channel belts, coarse sediment can largely bypass the delta plain. Grain-size changes occur across some key stratigraphic surfaces, but not consistently so. Channels in valleys are, on average, coarser than similar channels in unconfined systems, but, in the presented datasets, valleys do not contain the coarsest channels. The data have also allowed the analysis of down-system fining rates in ancient, sandy fluvio-distributary systems, with grain size being measured to decrease at rates ranging from 0.7 to 7.7 μm/km—values that compare favorably with modern rivers. Such large-scale trends are ornamented by, and link to, smaller-scale spatial changes associated with, for example, channel bars, crevasses, and mouth bars, and an initial dataset of associated fining rates has been collected. In general, very large systems (rivers) have low fining rates because of their great size, whereas the converse is true for small systems, especially if the grain size range is large. Consideration of downstream fining rates has led to the insight that avulsion initiates an unequal race to the shoreline. Suspended very fine sand and silt is likely to reach the shoreline with the avulsion flood waters, but bedload will advance far more slowly, perhaps too slowly to reach the shoreline before the river avulses again. Some avulsions may lead to notable temporal variations in the caliber of sediment supplied to shorelines. As expected, the largest system, the paleo-Orinoco, is the finest grained. The Ferron deltas have catchment areas 12 and 22 times smaller than the Orinoco, and are the coarsest grained. Remarkably, though their catchment areas differ by a factor of two, they have almost identical sand grain-size distributions. The data have also proved powerful in refining paleogeographic reconstructions, in particular suggesting “missing” depositional elements needed to complete local sediment routing systems. Careful tracking of grain size is also beneficial in that it is a key control on permeability, the description of which is crucial to the prediction of subsurface fluid flow.