This paper uses the high-frequency data of 50ETF options to study the predictive power of net buying presure (NBP hereafter) for the underlying stock index return and volatility, and compares the difference between different weighting methods, call and put options, options with different degrees of moneyness and information asymmetry. It is found that 50ETF options' NBP conveys both the return and volatility information of the underlying stock index. The information of Delta-aggregated NBP has already been conveyed by simply aggregated NBP. Call, put options and options with different degrees of moneyness contain different information on underlying stocks' return and volatility. Simply aggregated, Gamma-and Vega-aggregated NBP convey the volatility information. The NBP of ATM and OTM options has better predictability for volatility than ITM options. Moreover, asymmetric effects of NBPs on index return and volatility are identified.