xixi0222 / find-alpha-volatility-factor Goto Github PK

1. Use total, upper, down, relative volatility factors to find Alpha.
2. Define own functions for trading direction & volume.
3. Implement whole trading process & back-test with daily frequency.
4. Calculate and record important financial data.
5. Plot charts & Output trading financial data into files for visualization.

1. Volatility factors:

• Total Volatility (VOL):
  It measures the total volatility of the instrument. With higher volatility, it demands for higher return.
• Upper/Down Volatility (UVOL/DVOL):
  It measures the volatility by trend (go up or down) of the instrument. With higher unilateral volatility, the price volatility is higher in this side.
• Relative Volatility (RSJ): 𝐑𝐒𝐉 = 𝐔𝐕𝐎𝐋 − 𝐃𝐕𝐎𝐋
  It measures the relative volatility of the instrument. With higher absolute value of relative volatility, it shows short-term pressure (sharp long and steady short power), and lower future return.

2. Financial data: Holding, Position, Total assets, Available cash, Profit and Loss, Return, Sharp's retio, volatility...
3. Trading records: Trading direction, Trading volume, Turnover rate, Trading times, Maximum drawdown...

1. RSJ has the best result in most years.
2. Volatility Factors Windows: Shorter Time is better. 5 days, 10 days, 15 days windows perform better than long-time windows.
3. Macro Factors Windows: Longer Time is better. 24 months provides a better PnL, return and maximum drawdown.
4. Period Analysis:

• This model follows the trend of the market.
• It has a conservative approach. When the market is very volatile, it does not go as high or as low as CSI500.
• It needs improvement in a market with high volatility.
• It performs good in stable conditions. It overperforms the Chinese market by 30% in last two years.

Data: CSI500 Index
Data Source: Wind