本文以無母數之典型評價法來配適台指選擇權之市場資料,在模型中加入限制條件,並以台指期貨來作為台指選擇權之標的資產,來與傳統B-S模型以及使用相同資訊之隱含波動率B-S模型相比較,藉以了解典型評價法在台灣市場之評價與避險績效。實證結果顯示台指期貨報酬率顯著不服從常態分配,即不符合B-S模型之基本假設,因而使得加入條件限制之典型評價法比B-S模型有較佳之評價績效。避險測試結果顯示未加入條件限制之典型評價有最佳之避險績效,可能因為條件限制的加入,降低了Delta值估計的精確度所致。最後,在評價誤差之因素分析中,亦證實了增加限制條件,有效地降低評價誤差。 This study applies nonparametric Canonical Valuation to TAIEX options, adds constraint to the model, and uses futures as the underlying of TAIEX options, in order to investigate the performance of option pricing and hedging. We find the returns of futures violent the assumptions of normality B-S model. Thus the constrained canonical model outperforms B-S model. The result of hedging shows that the unconstrained canonical model is the most efficient in hedging. This may be that adding constraint to the model reduces the accuracy in delta estimating. Moreover, the regression of pricing errors shows that the additional constraint in these two models decreases pricing errors.