Abstract
In routine small-molecule structure refinements with SHELXL, the default weighting scheme relies on two adjustable parameters, a and b, in the WGHT command. For most refinements the a value is typically <0.1 while b is expected to be <1 or so. High values of a can be a consequence of poor counting statistics, i.e., weak data. Large values of b may result from uncorrected systematic errors and/or unresolved problems, such as missed symmetry, disorder, and twinning. Nevertheless, on occasion large b values persist, even when there are no apparent model deficiencies. In such cases, the available evidence suggests that the uncertainties derived during data reduction are too small. This short tutorial-style paper presents a straightforward means of achieving WGHT parameters within the expected ranges by fine-tuning parameters used in the Bruker data scaling, merging, and absorption correction program SADABS. Thus, it could also provide a basis for further exploration of weighting schemes in SHELXL. The general principles ought to be adaptable to data reduction software from other diffractometer manufacturers.