Large electric power systems are among the most complex systems created by humankind. One consequence of this complexity is that small unexpected disturbances trigger long chains of cascading component failures that can lead to massive power outages. Because of the many different mechanisms involved and the limited data available from historical cascades, modeling cascading failure is a challenging problem. Furthermore, among all the possible combinations of triggering events a relatively small number lead to large blackouts, which creates difficulties for conventional statistical approaches to risk analysis. In order to facilitate understanding about this complexity, this chapter describes several different approaches to modeling cascading blackouts and subsequently studying the risk of cascades, given a blackout model. Together these results suggest that when combined with good models, risk analysis can provide valuable and actionable insight into cascading failures in power systems.