The expected return of a portfolio of two stocks is calculated as the weighted average of the expected returns of these stocks. Shouldn't the standard deviation of returns of a portfolio of two stocks also just be calculated as the weighted average of the standard deviations in returns of these stocks?