斜率趨勢指標在台灣期貨市場的應用

Abstract

本研究應用蘇玄啟(2005) 碩士論文中所使用的股價技術指標－斜率趨勢技術指標，使用此指標並測試此指標的操作績效。 本研究以1999年1月5日至2008年5月6日的台指期貨日內價格資做為研究樣本，並將研究期間分為樣本期間與預測期間，採用「只做多」、「多空兼做」，「只做多配合停損」、「多空兼做配合停損」操作策略，並在不同的交易成本下，測試198種不同時間週期的斜率趨勢指標策略是否具備有效性(Effectiveness)、持續性(persistence)及可預測性(Predictability)。

我們使用t分配來檢定斜率趨勢技術指標的有效性；等級相關係數檢定來測試績效持續性，最後使用二次分配來測試此指標是否具有績效可預測性。 實證研究結果如下： (1)斜率趨勢技術指標具有有效性，在測試期間的22組資料中，必存在一個指標可以可打敗買進持有策略。 (2)使用指標在只做多，或是多空兼做，其指標間績效的排序並沒什麼改變。 (3)沒有一種操作方式(只做多，多空兼做，只做多配合停損、多空兼做配合停損) 具有顯著的成功率可以持續打敗大盤。 (4)斜率趨勢技術指標的績效具有持續性，但如果結合停損策略，則會失去持續性。 (5) 我們建議投資人可以在期貨市場中使用 50 or 53 30-min 時間週期的斜率趨勢技術指標。而超過 83 30-min 的斜率趨勢技術指標績效並不顯著。 (6)在空頭市場中，斜率趨勢技術指標明顯有效打敗大盤，配合停損策略後則更加的明顯。

This study aims to apply the methods devised by Su (2005). By adopting the quantitative technical trading strategy Slope Trend Indicator (referred to herein as “STI”), we analyzed the performance of the Taiwan futures market. The sample data consisted of 30-minute intraday data on the prices of Taiwan futures recorded from January 5, 1999 to May 6, 2008. To measure the performance of 198 STI strategies under different trading costs, we examined four kinds of trading decisions: long-only, long-and-short, long-only with stop-loss, and long-and-short with stop-loss. Standard t-statistic analysis has been adopted to estimate STI’s effectiveness. Spearman Rank Correlation Test has been used to test the performance of persistence. Moreover, the Binomial Proportionality Test has been appointed to assess STI’s predictability. The empirical results are as follows: (1) STI was effective. In a test period comprising 22 groups, there existed at least one STI strategy to beat the buy-and-hold strategy in each group. (2) The return rank orders of the long-only and long-and-short STI strategies were roughly the same. (3) None of the four trading decisions yielded significant success rates. This test result differed from Su’s (2005). (4) STI strategies possessed significant persistence; however, when combined with stop-loss, they lost their persistence. (5) We recommended the 50 or 53 30 min STI strategies in the Taiwan futures market. In addition, the returns of STI strategies longer than 83 30-minute were insignificant. (6) STI strategies performed well in the bear market, especially when combined with the stop-loss strategy.