Several years ago Sharpe suggested a measure for the evaluation of portfolio performance. The measure was conceptually simple, easily calculated, and applicable to an entire investment portfolio, in contrast to the measures of Treynor and Jensen which measure only the undiversifiable risk in a portfolio. Sharpe's measure is still a frequently recommended tool for measuring portfolio performance. The measure is, however, biased. It is the purpose of this note to demonstrate the existence of the bias, indicate its size, and provide a means of correcting it.