We show that the week after selling a large-prize Texas Lotto winning ticket, a retailer experiences a 12 to 38 percent relative increase in ticket sales. Some increase persists for up to 40 weeks. We document that the sales response increases with jackpot size and is larger in areas with more economically disadvantaged populations. Sales patterns across games and across retailers are not consistent with most advertising explanations. Furthermore, response patterns are not consistent with representativeness-based explanations for the hot hand or gambler's fallacy; we suggest an alternative explanation for the observed "lucky store" effect.