When performing classification and regression predictive analytics, it important to do multiple levels of analyses. For example, in predicting bankruptcy, it's important to understand how often companies actually go bankrupt. A model that predicted 50% of firms would go bankrupt would not be very useful if the actual rate of bankruptcy was about 2%.
Base rates refer to the probability that some event will happen based on some historical average. However, it important to be aware of the base rate fallacy: A prediction that places too little weight on the base rates of the past and instead uses different or new, recently received, information.