Some have argued that low-risk investments grow more attractive at longer investment horizons as their Sharpe ratios improve relative to high-risk investments. This article rejects this view by showing that the long-term standard deviation of returns can be misleading as an indicator of risk: Because of compounding, investments with identical standard deviations can have entirely different long-term risk characteristics even if all instantaneous investment returns are normally distributed. Furthermore, even if an individual investor accepts standard deviation of long-term returns as a measure of long-term risk, the relative attractiveness of high-risk investments can be kept constant by proper portfolio adjustments.