While the standard deviation will tell a risk manager how much a security can vary over time, it doesn't tell them how much an actual position could lose to a certain degree of accuracy. This is where value at risk comes in. Value at risk is a statistical measure of the riskiness of a financial entity or portfolio of assets. It estimates the potential loss from adverse market moves. It's defined by the maximum dollar amount expected to be lost over a given time horizon at a predefined confidence level. For example, the 99% one-day VaR is 10,000. This means the value of the investment will not fall by more than 10,000 in one day with a 99% certainty. VaR can be calculated using different techniques. The historical method assumes the pattern of historic returns is indicative of future returns. Next, the parametric method assumes returns are normally distributed, and as such, requires expected returns, standard deviation, and correlations with a portfolio to be able to calculate the VaR. Finally, the Monte Carlo simulation is where many, many scenarios are run within a computer model, calculating future returns under each scenario and therefore, allowing the worst case outcome a certain percentage of the time to be calculated. This does not rely on historic patterns repeating themselves or returns being normally distributed but is the most complex of the three methods.