應用停損停利策略檢測台指期交易績效分析之研究

金融市場中，人們不斷嘗試追求更高、更穩定的報酬率且同時規避系統性風險之交易策略，因而使用了許多技術指標與類神經網路，但是目前還沒有一種人人皆可使用，而且能夠穩定獲得超越大盤報酬率的交易聖杯出現。但隨著人工智慧(AI)的發展，似乎出現了這種契機。但使用AI自動交易之後，人類經驗是否還有用處? 未來的金融交易市場，難道就不再需要人為判斷或介入了嗎? 人類交易員的績效是否永遠無法打敗機器? 因此本研究除了發現可供機器學習之資訊外，也嘗試使用一些簡單的基本策略，看是否能夠取得更好的成果。本研究針對當沖交易，並使用兩種熱門的技術指標-隨機指標(KD)與指數平滑異同移動平均線(MACD)，還有做為基本對照組買進並持有(Buy and Hold)策略，搭配兩種常用的交易輔助策略停損與停利。研究結果發現，將基本策略搭配適當停利模式後將有顯著的績效提升。

關鍵字

台指期；停損；停利； MACD ； KD

並列摘要

From Technical Analysis to Neural Network , people keep looking forward a trading strategy which is risk-averse and constantly profitable, this strategy also referred to as a “Holy Grill” for trading. However, with the development of artificial intelligence (AI), this seems to have emerged. But is there any use of human experience after automated trading using AI? Will future financial trading markets no longer require human judgment or involvement? Will the performance of human traders never defeat the machine? Therefore, this study has found In addition to machine learning information, try to use simple basic strategies to seek a chance to achieve better results. This study focuses on the use of instinct trading and uses two popular technical indicators - Stochastic (KD) and Indexed Moving Average Convergence and Similarity Moving Average (MACD), as well as a Buy and Hold strategy as a basic control group. , With two commonly used transaction assistance strategies, stop loss and take profit. According to the result of the study, Basic Strategy with proper use of take profit , a significant improvment on performance were achieved.