本文旨在探討次不同風險估計之Sharpe指標於投資組合選取與績效表現之優劣。先推導出公式型態相同之一般化Sharpe指標（GSR）與風險值形式Sharpe指標【GSR（VaR）】，再以此二模型進行模擬分析。經實證發現：（1）透過GSR或GSR（VaR）模型篩選後之新投資組合與未經篩選之原投資組合，三者間之投資績效或績效排序，均無顯著差異；（2）在股價波動頻繁的臺灣股市中，以「事前」Sharpe指標作為投資組合篩選決策，實務上有侷限性。

The purpose of this paper is to find out which Generalized Sharpe Ratio in VaR or variance form used for portfolio selection has better performance on Taiwan stock market. The empirical evidence based on the above two models derived in the same types of formula in this paper shows: (1) no matter whether the portfolios are screened and selected by two Sharpe Ratio models or not, the performance and the．rank of performance among the three portfolios are not significantly different; (2) the ex ante Sharpe Ratio models are not suitable for decision making in portfolio selection on Taiwan stock market where the stock prices fluctuated frequently.