This work offers a review of forecasting methodologies and empirical applications that are useful for macroeconomists. It provides an overview of econometric methods available for forecast evaluation, including both traditional methods as well as new methodologies that are robust to instabilities. It then analyses their usefulness, their assumptions as well as their implementation, to provide practical guidance to macroeconomists. The second part of this work addresses special issues of interest to forecasters, including forecasting output growth and inflation as well as the use of real-time data and structural models for forecasting.