Two-way price difference autocorrelation standardized synthetic factor
factor.formula
CDPDP:
in:
- :
The first-order difference of the price at the t-th time point is calculated as: $\Delta P_t = P_t - P_{t-1}$, where $P_t$ represents the price at time t.
- :
The price increase difference series, defined as the value of $\Delta P_t$ when $\Delta P_t > 0$, and 0 otherwise.
- :
The price decline difference series, defined as the value of $\Delta P_t$ when $\Delta P_t < 0$, and 0 otherwise.
- :
The correlation coefficient of the price increase difference series and its one-period lagged series is calculated as follows: first, screen out all time points where $\Delta P_t > 0$, and then calculate the 20-day rolling correlation coefficient of the price difference value $dP^+{t}$ at the corresponding time point and the price difference value $dP^+{t+1}$ lagged by one period.
- :
The correlation coefficient of the price decline difference series and its one-period lagged series is calculated as follows: first, filter out all time points where $\Delta P_t < 0$, and then calculate the 20-day rolling correlation coefficient between the price difference value $dP^-{t}$ at the corresponding time point and the price difference value $dP^-{t+1}$ lagged by one period.
- :
The mean of the correlation coefficients between the price increase difference series and its one-period lagged series.
- :
The mean of the correlation coefficients between the price decline difference series and its one-period lagged series.
- :
The standard deviation of the correlation coefficient between the price increase difference series and its one-period lagged series.
- :
The standard deviation of the correlation coefficient between the price decline difference series and its one-period lagged series.
factor.explanation
This factor is based on the autocorrelation of the price difference sequence and captures the inertial characteristics of stock price changes. Compared with the single sequence difference, this factor considers the autocorrelation of price increases and decreases respectively, so as to capture the continuity and reversal potential of price trends more finely. The lower the factor value, the more likely the price will reverse direction after rising or falling, which conforms to the logic of mean reversal, and is therefore usually negatively correlated with future earnings performance. In quantitative trading, this factor can be used as an effective tool for building reversal strategies, and can also be used to capture market sentiment and trading congestion.