本文以海龜交易法則核心交易策略為主軸,運用系統交易分析系統,在台股期貨做模擬交易實證。係以台灣期貨交易所上市之台股期貨(TX)為研究標的,研究期間為1998年7月21日(台股期貨開始交易)至2011年11月7日止的日線資料及分時資料為主。研究期間共設計8個實驗系統,並區隔不同期間樣本資料,將全樣本區間劃分為樣本內配適期及樣本外檢驗期,以樣本內配適期做進、出場參數最佳化,並將其模擬交易於樣本外檢驗期做績效實證,以驗證前後兩期交易績效的一致性。並透過與原始海龜核心交易法則短線30分線系統做同期間比較分析,以歷史資料做進、出場參數最佳化回測驗證是否有助於未來交易之獲利。實驗結果顯示,採海龜核心交易法則做為台股期貨交易規則,在全樣本期間確實可以從中獲利。至於進、出場參數最佳化實驗結果顯示,參數最佳化不一定能提高未來交易獲利之穩定度,這可能與配適期與檢驗期行情波動與幅度是否相似有關。最後本文嘗試在策略中加入停損機制,實驗結果發現採用停損策略確實可以提升獲利,這隱含透過某種程度做適當風險控制,亦能達到類似參數最佳化之效果。
The core of the study is done based on the principle of “Way of the turtle” to do buying and selling futures by using specially programmed trading system and target to study TAIFEX futures (TX) in Taiwan Future Exchange. Trading period is from July 21st, 1998 to Nov. 7th, 2011. We make 8 different trading strategies and have the test data separated to in-sample and out-of-sample by different trading period. And by using the best testing result of trading during the period of in-sample, we simulate the same trading strategies and do buying and selling futures during the out-of-sample to verify whether the performance will be the same or not.Also we compare and analyze the performance with the best testing result of buying and selling futures from the strategy of 30-minute trading period under model of “Way of the turtle”. With above findings, we want to very whether that strategy will gain positive returns from future trading or not. The study identifies that buying and selling futures in TAIFEX (TX) under “way of turtles” strategy do get returns in both in-and out- of sample period. However, the best testing result do not equal to having stable returns from buying and selling futures. That maybe related to the changes and high/low ratios in Taiwan futures market. At the final stages of the study, we have “Loss Stop” added into our trading strategies. It comes out that strategy “Loss Stop” not only significantly increases returns and minimize the risk, but also achieve the best testing result.