Two factors that can impact model run times are the number of portfolio locations and the size of the event set for a given peril/region model. A highly granular peril model, such as flood, will often have significantly larger event sets and, when coupled with a portfolio with over 1 million locations, it can take several days to run. Analysts often use workarounds to reduce run times, including splitting larger portfolios into smaller sections and then aggregating the results, but this can create a significant amount of time-consuming manual work to prepare the data.
To meet risk analysts’ and cat modelers’ complex needs at scale, Moody’s developed Risk Modeler — a next-generation, cloud-based modeling application on the Intelligent Risk Platform™. As a cloud-native solution, Risk Modeler has the power to quickly run large portfolios (over 1 million locations) against Moody’s RMS HD Models, the industry’s most detailed and complete probabilistic models. All RiskLink DLMs and ALMs, as well as HD Models, can be run through Risk Modeler, providing users the power and speed that cloud-native architecture allows. As an example, running 1 million locations against the North Atlantic Hurricane DLM historical event set was 36 times faster in Risk Modeler.