In order not to have to determine a probability distribution for each single situation we make use of models. These are, on the one hand, probability distributions that are suggested by important characteristics of the respective observation process. On the other hand, there are distributions that, as experience has shown, yield a good adaptation to empirical distributions for many applications. For this it is necessary that they depend on one or various parameters. These are individual numerical values that appear in the formula of the probability or distribution function. They permit the fine tuning (Feinjustierung) of distributions for the respective application situation.