Transaction Volume Percentile Difference Reversal Factor
factor.formula
Calculate the 13/16th percentile of the daily transaction amount distribution, denoted as $P_{13/16,t}$
Select the past N trading days, sort the daily $P_{13/16,t}$ values from high to low, and sum the rise and fall of the first M trading days, recorded as $M_{high}$
Select the past N trading days, sort the daily $P_{13/16,t}$ values from high to low, and sum the rise and fall of the next M trading days, denoted as $M_{low}$
Calculate the turnover percentile reversal factor M:
in:
- :
On day t, the 13/16 percentile of the daily transaction amount distribution. This percentile can be regarded as the representative amount level of large transactions on that day. A higher $P_{13/16,t}$ value means that transactions with large transaction amounts on that day are more frequent.
- :
The lookback period for calculating the reversal factor indicates the number of past trading days of transaction volume percentile data to be compared. The default value is 10, which can be adjusted according to the strategy.
- :
Select the number of trading days used to calculate the sum of the price increase and decrease, which represents the length of the lookback period corresponding to the high/low transaction volume percentile. The default value is 10, which can be adjusted according to the strategy.
- :
The sum of the price increases and decreases of the M trading days with the highest $P_{13/16,t}$ values in the past N trading days. This value represents the overall market price performance during periods of large-volume trading activity. A high value means that when large-volume trading is active, stock prices tend to rise.
- :
The sum of the price increases and decreases of the M trading days with the lowest $P_{13/16,t}$ value in the past N trading days. This value represents the overall market price performance during the period when large transactions are not active. A low value means that when large transactions are not active, the stock price tends to fall.
factor.explanation
This factor is based on microstructure theory and believes that the short-term reversal momentum of stocks comes from the behavioral characteristics of large-order transactions. The early version used "average daily single transaction amount" as the cutting standard, but this factor uses "13/16 percentile of daily intraday transaction amount" as the basis for dividing high/low transaction volume, which can more effectively capture the period when large-order transactions are active, thereby more accurately describing the reversal effect.
The logic of this factor is that when large-order transactions are active (i.e. $P_{13/16,t}$ is high), the market may be overly optimistic, causing the stock price to rise too fast in the short term; and when large-order transactions are inactive (i.e. $P_{13/16,t}$ is low), the stock price may be underestimated. By calculating the difference between $M_{high}$ and $M_{low}$, possible reversal opportunities can be identified.
This factor reflects the changes in the market microstructure and the potential impact of trading behavior on stock prices. It is a microstructure reversal factor. Its effectiveness may be affected by factors such as market liquidity, trading system and investor structure, so it needs to be comprehensively analyzed in combination with other factors and market conditions.