To interpret upcoming results from the LHC it will be important to discriminate different models of new physics in a statistically meaningful manner. In this talk I will introduce Bayesian statistics as a tool for this. The comparison of different models in Bayesian statistics usually involves the calculation of a likelihood function and a prior distribution of the theoretical model. The likelihood function is typically calculated numerically and I will show this by using Monte-Carlo techniques. The talk looks at a toy example using only a few CKM observables.