An indicator which is calculated after the data reduction of aΒ test data set may be used to estimate the (systematic) instrument error at a macromolecular X-ray source. The numerical value of the indicator is the highest signal-to-noise [I/Ο(I)] value that the experimental setup can produce and its reciprocal is related to the lower limit of the merging__R__factor. In the context of this study, the stability of the experimental setup is influenced and characterized by the properties of the X-ray beam, shutter, goniometer, cryostream and detector, and also by the exposure time and spindle speed. Typical values of the indicator are given for data sets from the JCSG archive. Some sources of error are explored with the help ofΒ test calculations using__SIM_MX__[Diederichs (2009),__Acta Cryst.__D65, 535β542]. One conclusion is that the accuracy of data at low resolution is usually limited by the experimental setup rather than by the crystal. It is also shown that the influence of vibrations and fluctuations may be mitigated by a reduction in spindle speed accompanied by stronger attenuation.