We find that, on average, for the full sample of stocks comprising the Samp;P 100 index, theoretically superior implied volatilities are less accurate forecasts than historically based forecasts of future volatilities. However, when we split our sample based on proxies for liquidity, we find...