We use (Bakshi, Kapadia, and Madan, 2003, Review of Financial Studies 16: 101-143) methodology to measure the option-implied ex ante skewness of the risk-neutral returns distribution for underlying stocks. We find a negative relation between option-implied skewness and subsequent stock returns, even after controlling for a myriad of firm-characteristic variables. Specifically, the cross-sectional stock return predictability of option-implied skewness is only significant during periods of low market return and high investor sentiment. Furthermore, we find that the predictive power of skewness can be attributed to market state rather than sentiment. Our findings suggest that investors should consider high option-implied skewness stocks as they would lottery-like stocks.