Coordination geometries of some cations

Of course it always depends on what restraints are used and how they are used. As described in Chapter 2 there are two kinds of restraints direct and relative ones. The former restrain a variable (e.g. an interatomic distance) to a certain target value, while the latter relate variables within the model without imposing any outside values. Relative restraints generally exert much milder influence on the model than direct ones, and even a very large number of justified SADI and delu restraints cannot jeopardize the ethical integrity of the model. This may be entirely different with direct restraints, and the use of many strong dfix and ISOR commands to make the model look the way the chemist wants it to, can indeed result in questionable crystal structures.  

In general, restraints must be applied with great care and only if justified. When appropriate, however, they should be employed without hesitation, and having more restraints than parameters in a refinement is nothing to be ashamed of. 

George Sheldrick says with the right restraints, you can fit an elephant to any data . 

---