SmartRisk uses a 99.5% expected tail loss (ETL) in a proprietary heavy-tail model. Loosely speaking, it's the average of the worst 0.5% returns of the portfolio.
Historically, calculations of risk have used a Gaussian model (bell curve, normal distribution) to calculate risk. Largely because of its relative simplicity and modest computational requirements, it was used as the primary risk model from the late 1950's through the early 2000's. As inexpensive, massive computing power has become widely available, sophisticated market participants have switched to more realistic and robust models known as heavy-tail models.
Heavy-tail models capture a broader spectrum of extreme market events that happen far more often than traditional models would predict. These are also the events that cause clients to react with emotion and to make bad decisions. By managing risk and expectations with SmartRisk, clients are better prepared to make rational decisions regarding their portfolios and the potential for major market events.
More on the subject can be found in our white paper: The Risk Evolution.