Research On Black Series Futures Matching Trading Based On High Frequency Data

Paired trading strategy is widely used in relatively mature markets such as the United States and Europe,and used by institutional investors such as foreign hedge funds to obtain huge low-risk returns.For a long time,China's stock market has been a typical unilateral market,which can only be long but not short.The scope of financing and securities lending business is relatively limited,but whether short selling mechanism exists in the market is one of the prerequisites for realizing paired trading strategy.The futures market has a short-selling mechanism and a T+0 mechanism,which provides a good platform foundation for the practice and operation of my country's securities market pair trading strategy.To a certain extent,this strategy can suppress the sharp fluctuations in contract prices,which is conducive to further improving the price stability mechanism of the futures market.This paper selects the black series futures price as the research object,and the sample data is the minute line data of the contract.The main content can be divided into two parts: the black series futures trading demonstration under the traditional strategy and the empirical evidence of matching trading after using RBF neural network to optimize the trading signal,and the comparative analysis of the two strategies is carried out.The implementation of paired trading strategy mainly includes two important links: one is to choose a pairing target;the second is to establish a trading strategy.When selecting a matching combination,first of all,the contract price is analyzed and cointegrated,and the long-term equilibrium relationship between the NI2006 and NI2007 contract prices is selected and verified,and then the error correction model is used to estimate the short-term deviation of the price difference from its long-term equilibrium.The fixed multiple of the spread's standard deviation is used as the opening stop loss threshold,thereby establishing a trading strategy for hedging and arbitrage.Empirical results show that the strategy has achieved considerable benefits under the premise of low risk and low retracement.After verifying the feasibility of the paired trading strategy in the futures market,this paper selects a radial basis neural network model that can characterize the deviation of data from the center point to optimize trading signals.The neural network model uses the spread sequence of some contract pairs to train the original data.The K-Means clustering method is used to obtain the clustering center of the radial basis neural network,and the gradient weighting method is used to update the network weights,and the trained neural network model is used to predict the change trend of the spread.Based on the traditional strategy of spread standard deviation as the stop loss threshold,the forecast model makes a second judgment on the orders issued by the trading strategy,rejects the untimely opening or stop loss signals,and optimizes the trading signals.The optimized paired trading strategy is more cautious in opening positions and the stop loss is more reasonable.Under the conditions of more controllable risks and drawdowns,the profitability of the strategy is further improved.