Research On Portfolio Based On Multi-Varieties And Strategies Of China Commodity Futures Market

In recent years,with the rapid advancement of computer technology and data science,China's quantitative investment industry has developed rapidly,and the scale of domestic quantitative CTA funds has also expanded.In today's complex market environment,it is basically impossible for quantitative funds to rely on a single strategy to achieve stable profits.The combination of multiple strategies and multiple varieties is an inevitable choice for quantifying funds.This article starts with two different types of trading strategies,and slightly improves on the basis of classical strategies.It constructs a statistical arbitrage trading system and a quantitative timed trading system,respectively.The quantitative time-series trading system includes two strategies,Double Mean Strategy and Bollinger Channel Strategy,and the statistical arbitrage trading system mainly use cross-variety arbitrage.Then,this paper selects one-minute high-frequency data of nearly 40 futures varieties from China's commodity futures market from 2014 to 2018 to test the two trading systems.It shows that single-species and single-strategy cannot achieve stable profitability in China's futures market.Different strategies have different performances in different varieties.A single strategy cannot achieve full-species profitability.At the same time,under the single-strategy single-species,the fund curve clearly shows the characteristics of low yield and high volatility.This paper attempts to build a single-strategy multi-variety portfolio to improve the performance of the portfolio.The results showed that compared with single-species,multi-species portfolios can significantly improve the performance of the strategy.Under the simple equal weight combination method,the stability of the cross-varietal arbitrage strategy,the double-moving average trade strategy and the Bollinger Bands trading strategy has been greatly improved,and the Sharpe ratio has been significantly improved;and through the sample data backtesting,the winning ratios and profit-loss ratios of different varieties have been obtained.Information,the performance of multi-species portfolios after Kelly's formula optimization of funds has been further improved.The out-of-sample Sharpe ratio and maximum retracement are better than equal-weight strategies,especially the Bollinger Bands strategy.The out-of-sample Sharpe ratio is from 0.16.Increase to 1.85.Finally,this article attempts to superimpose different policy systems.Because of the alpha gains obtained by statistical arbitrage,and the quantitative ? gains obtained from the market,the different characteristics of the two strategies make the superposition of multi-strategy systems better able to reduce capital fluctuations and increase the risk-to-risk ratio.After empirical analysis,it is found that the multi-strategy multi-species trading system after superposition performs better,the maximum retracement further decreases,and the annualized volatility is only 4%.In addition,this paper finds that there is a certain positive correlation between market volatility and timed trading system.After the volatility timing,the strategy combination model can obtain higher benefits than simply superposition.