This paper documents that median holding period in structured products based on market index is less than a day from initial purchase to liquidation even for retail investors. Less than 6% of all series ever traded by retail investors are held until maturity. More importantly, buy-and-hold strategies perform worse than frequent trading. Based on a unique proprietary dataset that provides the details of all transactions - including account identifier and direction of the trade - in the Korean ELW (equity linked warrant) market between 2009 and 2011, we find that both HFT (high frequency trader) accounts and non-HFT accounts perform worse when either average holding period is long or average end-of-the-day position is large. Such failure of buy-and-hold strategy likely reflects time decaying properties, i.e. theta, of option-like products. Our findings suggest that measuring expected returns for options simply assuming that they are held until maturity may underestimate the true expected return.