Research On The Quantitative Trading Strategies

With the growing in the stock market, how to invest has become a key issue of discussion, how to choose the suitable varieties from a number of futures varieties, how to choose the right time to buy or sell, how to exit the market at the right position in order to maximize profits or minimize losses, how to choose a different trading strategies in different market conditions, how to dig out from piles of historical data so that we could find the regular pattern, how to combine different varieties or investment strategy. These problems have become impossible for the traditional investors. It is this time quantitative trading began flashed in the minds of many investors, investors can realize their ideas by means of a computer, Indeed, the computer let investors become more rational, could get more stable earnings.In this paper,author construct two trading strategies, one is a trend trading strategy, our perpose is seeking the optimal trading strategies with improved genetic algorithm, the optimal trading strategy could get stable gains and smaller retracement. Another is the arbitrage strategy which is based on the use of GARCH models to find arbitrage opportunities. This paper selected stock index futures and other contracts, due to the relatively familiar with the stock index futures and stock index futures is active.The innovation of this paper is:(1)the paper adopt improved genetic algorithm, crossover probability and mutation probability took the idea of dynamic change,which is proved irrational.because of outstanding individuals should have a greater probability into the next generation, the weak individual should have relatively small variation probability.(2)the paper constructed a complete transaction process, rather than simply selecting several technical indicators,and consider the short lots and T + 0 conditions.The stock index futures contracts, rather than stocks,is considered. Firstly,the paper adopt CCI and MACD to conduct admission strategy.Secondly,the paper use K-line to filter.Finally the strategy has fixed loss percent and dynamic profit percent to ensure that the profit realize maximization and retracement is small as soon as possible; genetic algorithm’s fitness function use a two-dimensional vector-profit and history retracement,the value of retracement will reflect investors’ risk preferences.However,a strategy should consider more aspects,rather than only two aspects,which is task of next study.(3)Construction of arbitrage strategy use statistical models to capture arbitrage opportunities, instead of using historical transaction data to calculate arbitrage interval. The results show that strategy can capture a very effective profit opportunities.