Optimal forecasts are, under a squared error loss, conditional expectations of the unknown future values of interest. When stochastic demographic models are used in macroeconomic analyses, it becomes important to be able to handle updated forecasts. That is, when population development turns out to differ from the expected one, the decision makers in the macroeconomic models may change their behavior. To allow for this, numerical methods have been developed that allow us to approximate how future forecasts might look like, for any given observed path. Some technical details of how this can be done in the R environment are given.