The article explores energy market data to predict future returns. By analyzing volatility measures, a Realized GARCH model is used to forecast returns and volatility. The study finds biases in existing measures, highlighting the need for accurate modeling. High-frequency data improves model accuracy compared to daily data. Out-of-sample results show limited benefits from new volatility measures, suggesting caution in their use. The study emphasizes the importance of careful modeling and forecasting in energy markets.