The article discusses how using imperfect measures of volatility can lead to incorrect evaluations of forecast accuracy. When comparing different volatility forecasts, it's important to use efficient proxies like intra-daily range or realized volatility. The researchers found that the choice of volatility proxy can impact which forecast is deemed the best. They also identified conditions for maintaining accurate rankings of forecast performance when using a proxy. Tests of predictive ability can help when limited information about the user's preferences is available.