The article discusses how to manage extreme risks in calm and turbulent markets using a type of statistical model called Extreme Value Theory (EVT). By focusing on rare events in the stock market, EVT models can accurately predict potential losses. The researchers found that conditional EVT models are especially good at estimating extreme risks, outperforming traditional methods like GARCH. This means that EVT models are a better choice for calculating both standard and extreme Value-at-Risk measures in financial risk management.