請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/15568
標題: | 程式交易策略與投資組合之研究 Research on Program Trading Strategy and Investment Portfolio |
作者: | Chinq-shiun Chiu 屈慶勳 |
指導教授: | 李存修 |
關鍵字: | 移動平均線,包寧傑帶狀,程式交易,交易策略,最佳化,績效衡量,投資組合,最大績效拉回, Moving average,Bollinger band,Program trading,Optimization,Performance evaluation,Investment portfolio,Maximum draw down( MDD), |
出版年 : | 2013 |
學位: | 碩士 |
摘要: | 本論文研究作多及作空型之程式交易策略,個人投資者可藉由一般電腦設備及MultiCharts交易軟體,用較少的初始本金,即可建立個人化之交易系統及創造穩定之績效。
本研究設計之作多及作空策略簡述如下: 1. 作多型:捕捉波浪式上漲之盤態,以收盤價回測Bollinger Band之下限(3.5~4.5天的均線及1.3~2.4%標準差)再往上突破時買進;加上進場濾網:收盤價站上35~50天均線且此均線往上;收盤價跌破7.5~9.5天均線出場;停損設買入價的1~2.5%損失。 2. 作空型:捕捉急跌盤態,以收盤價跌破Bollinger Band之下限(2.2~4天的均線及2.3~3.1%標準差)作空;加上進場濾網:收盤價跌破22~45天均線且此均線往下;收盤價站上3.6~7.5天均線出場;停損設放空價的1~1.5%損失。 此作多及作空策略應用於四種商品:台指期(TXF1)、Mini -NASDAQ 100(NQ1)、小型輕原油(QM1)和日圓(JY1),在經過每個商品參數最佳化後之整體投資組合之交易策略績效之實證如下(2002年至2012年): 在交易分析層面: (1)勝率:43.0%;(2)平均獲利/平均虧損比率:2.3;(3)交易總次數:870 至於策略分析層面: (1)淨利:6,026,425元(初始本金200萬元);(2)最大策略虧損:458,181元;(3)策略在市場之時間:70% 最後則是週期性分析層面: (1)平均年報酬率:14.3%;(2)年報酬率的標準差:13.6%;(3)夏普比率(年):0.9 以此多時序、多空策略、多商品建立之投資組合,配合適當之資金管理以提高資金效率,能減少最大績效拉回,減少交易心理壓力,進而戰勝人性弱點。本研究更以推進分析評估2013第一季,其實證績效亦可持續創新高,本研究證實了可持續性的個人交易系統。 This thesis studies the long and short trading strategies. Individual investors with less initial principal by PC and MultiCharts trading software can create a personalized trading system with stable performance. The design of long and short trading strategies is summarized below: 1.Long trading strategy: capturing the rising wave pattern. When the closing price goes down to the lower limit of the Bollinger Band (3.5 to 4.5daysMA, and 1.3 to 2.4% standard deviation), buy one contract on the lower limit breakthrough withthe filter (on the closing price above 35 to 50 days MAwith positiveslope). When closing price falls below the 7.5 to 9.5 days of MA, sell one contract. Stop-loss sets the bid price for a 1~2.5% loss. 2. Short trading strategy: capturingthe plunge pattern. When the closing price falls below the lower limit of the Bollinger Band (2.2 to 4daysMA, and 2.3 to 3.1% standard deviation), sell one contract withthe filter (closing prices below 22 to 45 daysMAwith negativeslope).When closing price is above 3.6 to 7.5 days MA, buy-to-cover one contract.Stop-loss sets the bid price for a 1 to 1.5% loss. These long and short trading strategies are applied to four commodities: TAIEX, Mini-NASDAQ 100, Minilight sweet crude oil, and Japanese yen. After parameter optimization for each commodity, the overall portfolio performance summaries are shown below (2002-2012): for the trading analysis: winning percentage: 43.0%, average profit /average loss ratio: 2.3, and total number of trading: 870. For strategic analysis: net income: NTD 6,026,425 (initial principal amount of NTD 2.0 million),Maximum draw down: NTD 458,181, and strategy at the time of the market: 70%. Finally, the cyclical analysis: the average annual rate of return: 14.3%, the annual rate of return standard deviation: 13.6% and Sharpe ratio (year): 0.9 The multi-timing, long and short strategies, multi-commodity portfolio, with the management of the funds to improve capital efficiency, can reduce the MDD and trading psychological pressure, and thus to overcome the weaknesses of human nature. This study further evaluates the performance in 2013 Q1 and the trading system can create the new high on net profit. The research demonstrates the sustainable personalized trading system. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/15568 |
全文授權: | 未授權 |
顯示於系所單位: | 財務金融組 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-102-1.pdf 目前未授權公開取用 | 3.29 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。