Following the 2007-08-credit crisis, investment banks have been focusing on how to increase the accuracy of risk measures. The parameters used for measuring risk have been increasing in relation with the evolution of calculation models. The originality of this work consists in providing, for the first time in literature, a quantitative research on the best method to use for selecting dates when using the popular Monte Carlo simulation. The object of this paper is to find the optimal distribution of simulation dates for measuring counterparty credit risk. We will first analyse a certain counterparty from a real-life portfolio to demonstrate how, when generating the regulatory measures of an Internal Model Method an optimal distribution of observation days is achieved by increasing the number of short- and medium-term time steps. In the second section we will analyse a test portfolio using a flat yield curve to better observe the differences when changing the distribution of simulation dates.