金工定期报告20251107：换手率变化率的稳定GTR选股因子绩效月报20251031-20251107

Quantitative Models and Construction Methods 1. Model Name: Stability of the Growth Rate of Turnover Rate (GTR) - Model Construction Idea: The model identifies stocks with high turnover rate volatility but stable growth or decline in turnover rate, which may indicate potential future price increases. The model introduces a new metric, "stability of the growth rate of turnover rate," to capture this trend[6][15]. - Model Construction Process: The GTR factor is constructed based on the acceleration property of turnover rate changes, combined with the stability of the new factor. The specific formula or mathematical representation of the GTR factor is not provided in the report[6][15]. - Model Evaluation: The GTR factor demonstrates low correlation (less than 0.1) with other turnover rate factors in the same series, indicating its uniqueness and potential to enhance the performance of related factors[6]. 2. Model Name: Purely Enhanced TPS_Turbo Factor - Model Construction Idea: The TPS_Turbo factor is derived by applying a "purely enhanced" method to combine the GTR factor with Turn20 and STR factors, aiming to improve stock selection capabilities[6]. - Model Construction Process: The TPS_Turbo factor is constructed by integrating the GTR factor with Turn20 and STR factors using a purely enhanced method. The exact mathematical formulation of the combination is not provided in the report[6]. - Model Evaluation: The TPS_Turbo factor demonstrates superior stock selection ability compared to its components, as evidenced by its performance metrics[6]. 3. Model Name: Purely Enhanced SPS_Turbo Factor - Model Construction Idea: Similar to TPS_Turbo, the SPS_Turbo factor is created by applying the "purely enhanced" method to combine the GTR factor with other turnover rate factors, aiming to achieve better performance in stock selection[6]. - Model Construction Process: The SPS_Turbo factor is constructed by combining the GTR factor with other turnover rate factors using the purely enhanced method. The specific formula or mathematical representation is not provided in the report[6]. - Model Evaluation: The SPS_Turbo factor exhibits excellent stock selection performance, surpassing the other factors in the series[6]. --- Model Backtesting Results 1. GTR Factor - Annualized Return: 13.11% - Annualized Volatility: 10.22% - IR: 1.28 - Monthly Win Rate: 66.53% - Maximum Drawdown: 10.81%[7][11] 2. TPS_Turbo Factor - Annualized Return: 36.11% - Annualized Volatility: 13.22% - IR: 2.73 - Monthly Win Rate: 78.39% - Maximum Drawdown: 9.86%[7][11] 3. SPS_Turbo Factor - Annualized Return: 37.21% - Annualized Volatility: 10.87% - IR: 3.42 - Monthly Win Rate: 81.36% - Maximum Drawdown: 7.22%[7][11] --- Factor Backtesting Results (October 2025) 1. GTR Factor - Long Portfolio Return: 2.27% - Short Portfolio Return: 0.96% - Long-Short Portfolio Return: 1.31%[15] 2. TPS_Turbo Factor - Long Portfolio Return: 3.19% - Short Portfolio Return: -0.33% - Long-Short Portfolio Return: 3.52%[16] 3. SPS_Turbo Factor - Long Portfolio Return: 3.65% - Short Portfolio Return: 0.09% - Long-Short Portfolio Return: 3.56%[19]