There is immense current interest in rare events in finance. The theory of "large deviations" is a mathematical theory which has been around at least since the 1980's, and is a circle of ideas which can be used to understand the origination and propagation of rare events. We outline how the theory of rare events can be used to understand investment-grade tranches of collateralized debt obligations (CDO's), financial instruments for which, by design, the losses are rare. Our focus will not be the abstract theory of large deviations, but some simple calculations which give us insight into CDO's, and which reflect the strength of the theory of large deviations as a useful tool in analysis of financial products.