In finding parameters, we follow maximum likelihood approach that is estimate such values under which a model is most likely to generate the observed network. In the core of the estimation procedure lies the fact, that higher parameter values makes networks with higher corresponding statistics more probable (Koskinen & Snijders, 2014). For example, if a model involves reciprocity statistics, and a parameter for this statistics is positive, networks with higher degree of reciprocity will be more probable.