The researchers studied how to predict stock market volatility by adding factors like implied volatility, overnight returns, and the volatility of realized volatility to a forecasting model. They found that including implied volatility improved predictions, but adding more factors made forecasts even better. Overnight returns helped in most markets, while the US benefited from considering the volatility of realized volatility. The leverage effect was useful in five markets. Their findings were supported by a value-at-risk analysis, showing the practical importance of their results.