Design Of A Program Trading Strategy Based Cross-commodity Arbitrage In Futures

Program trading is a way of trading that gradually became popular in recent years in China. It uses computer language to describe the investment strategy or logic, with the help of computer efficient and fast computing ability, strong execution ability and sustainable operation ability of market analysis, automatic trading orders issued. It is systematic, disciplined, timeliness, etc. The futures market has the characteristics of flexible trade mechanism, transaction time, market liquidity is strong, very suitable for program trading. Therefore, based on the comprehensive description of program trading system, based on statistical analysis of historical data of the part of a futures contract, construct a suitable program of transactions across species arbitrage trading strategy.This strategy belongs to the basic principle of statistical arbitrage, cointegration theory and error correction model. Statistical arbitrage has a long history, has a very mature application in practice. Cointegration theory and error correction model is the analysis of construction method of data, trading strategy. The cointegration relation describes the long-term stable equilibrium relationship between the two time series, the error correction model considering the short-term non equilibrium cointegration, increase the accuracy of the regression model, help to improve risk control ability of trading strategies. In this paper, through mathematical derivation, the definition and test principle of the theory analysis, the trading strategy is more logical.The empirical analysis, this paper uses 13 cross species arbitrage portfolio as a sample. Through the analysis and comparison of the historical transaction data in sample statistics, select the best portfolio, based on the distribution of trading strategies to build its modeling error sequence. Finally, a pioneer program trading platform through the transaction, the back testing and optimization strategy, and the sample testing method to test the effect of strategy, its possible risks are briefly analyzed.