Course Page

The Art and Science of Prediction

Course Number:



Robert Blattberg,

Alan Montgomery,




Marketing, Business Technologies, Operations Research

Course Description

Most managers are ill equipped to predict key events or numbers, having neither the technical skills nor the right mindset.  All levels of an organization need to make predictions ranging from numerical forecasts of sales and earnings to predictions in which they have very limited data available to assist them such as predicting competitive response.  Perhaps the most common mistake is overconfidence.  The goal of this course is to teach students how to overcome their biases and make predictions not only in data rich environments but also ones that are data poor.


The course is designed to teach students how to make predictions for a wide variety of events and numbers.  To help develop their prediction skills, the course will provide 1) a structure and methodologies for making predictions and 2) the psychology of forecasting, and 3) incorporate them into a probabilistic framework.  Techniques we advocate are updating frequently, using “baselines”, learning from your mistakes, providing probabilistic not deterministic predictions, avoiding the use of ideological beliefs as the basis for predictions, and combining multiple and potentially disparate sources of information.  The students will be given different prediction assignments to help them master the art and science of predictions.


Lecture: 100min/wk and Recitation: 50min/wk