Weak data and high-resolution cut-off

In these two functions, F is the structure factor and the subscripts o and c stand for observed and calculated this nomenclature is going to be used throughout the book. Each addend in this summation is multiplied by an individual weighting factor w, which reflects our confidence in this particular datum and is derived from the standard uncertainty a of that measurement. The only difference between these two minimization functions is that the left one corresponds to refinement against squared structure factors (F ), while the right equation describes refinement against f-values. 

A broader discussion of this matter is beyond the scope of this book, but the interested reader may turn to the articles by Hirshfeld and Rabinovich (1973) and Amberg et al. (1979) for more in-depth information. 

As mentioned above, it is important not to exclude weak data. However, there is no reason to use data from high-resolution shells when they are all very weak, since these reflections are in fact noise and contain no usable information. Generally, intensities are weaker at higher 2 angles and almost no crystal diffracts to the theoretical limit of dmax = /2. Some care must be taken in the determination of the effective maximum resolution of a dataset. 

Because of counting statistics (background higher than signal of the peak), sometimes slightly negative intensities 

---