2014 Palisade @Risk conference, New Orleans

摘要

Huybert Groenendaal (EpiX Analytics) speaking at the Palisade Risk Conference in New Orleans late last year offered some insights into building 'decision-focused' @Risk models based on his experience with a large multinational client. First, at the design stage, a model needs to be appropriately sized for computability and still offer the desired output resolution. Also, every iteration must be a possible future scenario. Often, modelers fail to take account of relationships between inputs that may bias the Monte Carlo results. To help keep modelers on-track, EpiX offers 'Risk Optimizer' a free @RISK add-on that helps modelers chose the best method for characterizing relationships between variables. Failure to do so typically results in the underestimation of risk.