A method of calculating value-at-risk (VaR) that is based on historical data in order to assess the impact of market moves (i.e., market risk) on a portfolio/ a fund/ an investment, etc. This method, which assumes that history will repeat itself, uses re-organized, actual historical returns, ordered from worst to best or lowest to highest. In other words, under this method, VaR is directly calculated from past or historical returns.

For example, if a firm wants to calculate the 1-day 95% VaR for an investment that covers 100-day of data, the 95th percentile would correspond to the best of the worst 5% of returns. As the firm is using 100-day data, the VaR would simply correspond to the 5th worst day over the period.

This method of value at risk (VaR) calculation is also known as historical VaR (historical value at risk) or the historical simulation.