Study On Dynamic Quantitative Timing System Based On Technical Analysis

In the field of investment,investors are most concerned about how to choose assets and when to invest.Around the two problems,many scholars and industry researchers carried out extensive research and analysis,researches has shown that no strategy can guarantee investors the stability of getting excess returns,As for the study of partial quantitative strategies back test,some strategies can get big benefits in specific market form but in another kind of market form has been disastrous,there are also some strategy returns more stable,but relatively low returns.A fixed strategy can't beat any market,but that doesn't mean there's no one way to get a higher return on a wide range of any market types.The purpose of this paper is to construct a dynamic quantitative timing system,which can automatically identify the current market is what kind of market form,and then to optimize combination weights so as to avoid the instability of a fixed strategy.This paper focuses on how to build a dynamic timing system.In a perfect dynamic timing system,there are three key elements,alternative basic strategy,how to select strategy combination and market judgment.First,a dynamic timing system must have a sufficient number of effective timing strategies,too few effective strategies can not guarantee that it can adapt to most market forms.It is easy to appear excellent in some market forms,but there is a loss in another market.Second,after the research on a number of strategies we can find the characteristics of different strategies,and then we can build the optimal timing strategies under different market form combination,and give the strategies different weights under different market form,and trie to find the optimal strategy combinations in different market form in order to ensure high and stable return.Third,after the completion of the optimal timing strategy of portfolio construction,the most important link is the market judgment,all the preparatory work is based on the correct judgement of current market.This paper constructs several indicators to the probability of correct classification on the market,and then use the combination of different selection in different markets,to obtain high returns stable continued.The first step is to analyze the basic quantitative timing strategy.In order to complete the construction of dynamic selection system,this paper selected the closing price from April 16,2010 to December 11,2017 of the main stock index futures contracts as the research object,the research range can be divided into four periods,with 17 basic timing strategies of annualized return,rate of maximum retracement,long times,short times,winning rate to analyse.This paper summarizes the characteristics of different strategies and the suitable market patterns through the research and analysis of the basic quantitative strategies.The second step is the robustness test.Each strategy has its index parameters and policy parameters.Through the study,it is found that most strategy effects are sensitive to parameter setting,and the change of parameters has important influence on the timing system.In this paper,we examine the robustness of the above strategies,and analyze the robustness of different strategies in different markets,and several alternative strategies under different markets by traversing the parameters.The third step is the construction of the optimal timing combination.In this paper,three optimal timing combinations are constructed through two approaches.The first approach is the portfolio theory applied to the quantitative timing combination,we look on the timing strategies as assets,to analyze the rate of return and variance using the portfolio theory and then to calculate the effective frontier,to select the different timing combination weight.The fourth step is the market judgment index construction.In this paper,we find that the cross frequency of the oscillating market are relatively high and the unilateral market is low,based on which,we set up the index based on the cross frequency market.In addition,this part also uses the more mature but not widely used ADX index and CMI index to make a trend judgment.The fifth step is constructing dynamic selection system and out of sample testing.By judging the current market type,we can determine the optimal timimg combination in the current market form to conduct automatic transaction.In the end,this paper carries out an out of sample testing of the dynamic timing system by simulating the daily transaction data of 1200 price sequences and 1000 trading days by machine.It is found that in the oscillating market,KD strategy,J strategy,and BIAS strategy have significant timing effect,and the 5-30 average line strategy,average line segmentation strategy and long-short alignment strategy have obvious advantages in the unilateral market.The 5-30 strategy has maintained a high return and stability.MACD strategy is the most stable but the return is not high.Through the study above,it is found that there is no high return and low risk strategy that can adapt to any market.However,different strategies can be selected through identifying different market forms,which will maintain a steady and sustained earnings in different market conditions.The timing of combination system of earnings retracement ratio show the best effect,combination with the average and the MACD cross frequency achieved average annual return of more than 18%,but the bes robustness method is the portfolio theory method.The equal weight method performs worse in both annual return rate and robustness.In the market judgement index,the ADX system shows a very high robustness in each timing combination.Correspondingly,the CMI system's volatility of the yield is large and the robustness is poor.