This trading strategy is based on the momentum paper published by AQR in the 1990's. The strategy file needs to run with modified pyalgotrade source code. Request to [email protected] if needed. I used "Tushare" as my datasource and pyalgotrade as the backtesting platform. After washing the raw financial data, I grouped daily k bars to monthly k bars for this strategy, which was devised to pick the three best performing stocks in one month with Outstanding_Share and EPS requirements to filter the entire stock population. Then hold the position from trading at market price in maximum volume for a month to measure the portfolio performance. Finally the PnL result turned out to be randomly distributed when running against year 2018-2019.
- Daily k bars are combined into monthly k bars
- All orders are sent using the market price, but some may not be filled due to volumes are messed up in monthly k bars
- The strategy picks up 3 best pct_chg stocks in one month with some outstanding share and EPS requirements to filter the population, and hold for a month to measure the portfolio performance
Month Best performing SH/SZ stocks with specified OS > 20 million and EPS > 1 win pct change Return after holding for 1 month 01/2019 601155 23.9% 110.7% 000858 18.5% 000001 18.3% 2019 March 600352 58.4% 99.4% 601155 33.5% 000858 32.9% 2019 May 603288 12.2% 96.0% 002142 1.3% 600036 -0.7% 2019 July 300498 12.9% 101.4% 600048 11.4% 600031 8.0% 2019 Sep 601155 14.5% 99.6% 600188 11.3% 002142 10.4% Total return: 101.4%