金融工程 证 券 研 究 报 告 市场形态周报（20250901-20250905） 本周指数普遍下跌 本周市场回顾与最新信号 从本周的指数表现来看，本周指数普遍下跌，其中沪深 300 下跌 0.81%，中证 500 下跌 1.85%，中证 1000 下跌 2.59%。 当前，上证 50 的隐含波动率为 19.73%，相对于上周下跌了 3.32%。上证 500 的隐含波动率为 23.4%，相对于上周下跌了 2.33%。中证 1000 的隐含波动率 为 25.94%，相对于上周下跌了 0.55%。沪深 300 的隐含波动率为 20.27%，相 对于上周下跌了 3.29%。 我们统计了最近信号的次数和胜率。2025 年 8 月 25 日到 2025 年 8 月 29 日正 面信号共出现了 2442 次，未来高点平均胜率为 28.76%，负面信号出现 3750 次，未来低点平均胜率为 70.79%。 从宽基择时策略来看，上证 50、恒生等权重、恒生指数、恒生可持续发展企业 指数、恒生香港 35 出现看多信号，其余宽基信号为中性。 出现连续看多信号股票、特殊看多信号股票和券商金股出现看多信号股票见正 文。 更多具体详 ...

金融工程 证 券 研 究 报 告 市场形态周报（20250818-20250822） 本周指数普遍上涨 ❖ 本周市场回顾与最新信号 从本周的指数表现来看，本周指数普遍上涨，其中沪深 300 上涨 4.18%，中证 500 上涨 3.87%，中证 1000 上涨 3.45%。 当前，上证 50 的隐含波动率为 20.3%，相对于上周上涨了 2.93%。上证 500 的 隐含波动率为 22.36%，相对于上周上涨了 2.82%。中证 1000 的隐含波动率为 25.91%，相对于上周上涨了 4.86%。沪深 300 的隐含波动率为 19.21%，相对 于上周上涨了 1.12%。 我们统计了最近信号的次数和胜率。2025 年 8 月 11 日到 2025 年 8 月 15 日正 面信号共出现了 3365 次，未来高点平均胜率为 70.33%，负面信号出现 3167 次，未来低点平均胜率为 27.82%。 从宽基择时策略来看，沪深 300、上证综指、万得微盘股指数、中证 2000、中 证 1000、国证 2000、中证 800、Wind 全 A、中证 500、上证 50、创业板指、 恒生可持续发展企业指数、恒生指数、 ...

Quantitative Models and Construction Methods 1. Model Name: Heston Model - **Model Construction Idea**: The Heston model is used to calculate the implied volatility of near-month at-the-money options, serving as a market fear index. Implied volatility reflects market participants' expectations of future volatility [8] - **Model Construction Process**: The Heston model is a stochastic volatility model where the variance of the asset price follows a mean-reverting square-root process. The model is defined by the following equations: $ dS_t = \mu S_t dt + \sqrt{v_t} S_t dW_t^1 $ $ dv_t = \kappa (\theta - v_t) dt + \sigma \sqrt{v_t} dW_t^2 $ Here: - \( S_t \): Asset price - \( v_t \): Variance process - \( \mu \): Drift rate of the asset price - \( \kappa \): Rate of mean reversion of variance - \( \theta \): Long-term variance - \( \sigma \): Volatility of variance - \( W_t^1, W_t^2 \): Two Wiener processes with correlation \( \rho \) [8] - **Model Evaluation**: The Heston model is widely recognized for its ability to capture the stochastic nature of volatility, making it suitable for modeling market fear indices [8] --- Quantitative Factors and Construction Methods 1. Factor Name: Multi-Long-Short Ratio Scissor Difference - **Factor Construction Idea**: This factor is based on the difference between the number of long and short signals within industry index constituent stocks. It is used to construct industry timing strategies [15] - **Factor Construction Process**: - Define the number of long and short signals for each industry index constituent stock on a given day - If no long signals are present, set the long signal count to 0; similarly, if no short signals are present, set the short signal count to 0 - Calculate the scissor difference as the difference between the long and short signal counts - Normalize the scissor difference to obtain the scissor difference ratio [15] - **Factor Evaluation**: The backtesting results show that the timing model based on this factor outperforms the respective industry indices in all cases, demonstrating excellent historical performance [15] --- Backtesting Results of Models 1. Heston Model - **Implied Volatility Results**: - SSE 50: 13.63% (down 2.78% WoW) - SSE 500: 15.75% (down 3.31% WoW) - CSI 1000: 17.15% (down 3.26% WoW) - CSI 300: 13.96% (down 2.31% WoW) [10] --- Backtesting Results of Factors 1. Multi-Long-Short Ratio Scissor Difference - **Performance Metrics**: - Timing models based on this factor outperformed their respective industry indices in all cases, achieving a 100% success rate in backtesting [15]

- The report utilizes the Heston model to calculate implied volatility for near-month at-the-money options, serving as a market fear index. Implied volatility reflects market participants' expectations of future volatility[6] - Positive signals appeared 3052 times from July 14 to July 18, with an average future high-point success rate of 68.64%. Negative signals appeared 3426 times, with an average future low-point success rate of 27.06%[10] - The broad-based timing strategy signals indicate "bullish" for indices such as ChiNext Index, SSE 50, Wind Microcap Index, Hang Seng Equal Weight Index, Hang Seng Tech Index, Hang Seng Hong Kong 35, Hang Seng Index, Hang Seng Financials, Hang Seng Sustainable Development Enterprises Index, Hang Seng China Enterprises Index, and Hang Seng China (Hong Kong-listed) 100[12] - The industry timing strategy is constructed using the scissors difference ratio of long-short signals for industry index constituent stocks. If no bullish or bearish signals appear on a given day, the scissors difference value and ratio are set to zero. The model outperformed all respective industry indices historically, demonstrating excellent backtesting results[13] - Industry timing strategy signals show "bullish" for sectors such as building materials, light manufacturing, home appliances, textiles and apparel, non-ferrous metals, power and utilities, steel, consumer services, transportation, coal, pharmaceuticals, agriculture, forestry, animal husbandry, and fishing, and petrochemicals[15]

Quantitative Models and Construction Methods 1. Model Name: Heston Model - **Model Construction Idea**: The Heston model is used to calculate the implied volatility of near-month at-the-money options, serving as a market fear index. It reflects market participants' expectations of future volatility [7] - **Model Construction Process**: The Heston model is a stochastic volatility model where the variance of the asset price follows a mean-reverting square-root process. The model is defined by the following equations: $$ dS_t = \mu S_t dt + \sqrt{v_t} S_t dW_t^S $$ $$ dv_t = \kappa (\theta - v_t) dt + \sigma \sqrt{v_t} dW_t^v $$ where: - \( S_t \): Asset price - \( v_t \): Variance of the asset price - \( \mu \): Drift term - \( \kappa \): Speed of mean reversion - \( \theta \): Long-term variance - \( \sigma \): Volatility of variance - \( W_t^S, W_t^v \): Two Wiener processes with correlation \(\rho\) [7] - **Model Evaluation**: The Heston model is widely recognized for its ability to capture the stochastic nature of volatility, making it suitable for modeling market fear indices [7] --- Quantitative Factors and Construction Methods 1. Factor Name: Multi-Industry Timing Factor (Scissors Difference) - **Factor Construction Idea**: This factor is based on the difference in the number of stocks with bullish and bearish signals within an industry. It aims to identify timing opportunities by analyzing the divergence between bullish and bearish signals [14] - **Factor Construction Process**: - Define the number of stocks with bullish signals (\(N_{bullish}\)) and bearish signals (\(N_{bearish}\)) in an industry on a given day - If no bullish or bearish signals are present, set the respective count to 0 - Calculate the scissors difference as: $$ \text{Scissors Difference} = N_{bullish} - N_{bearish} $$ - Normalize the scissors difference to obtain a ratio: $$ \text{Scissors Ratio} = \frac{N_{bullish} - N_{bearish}}{N_{bullish} + N_{bearish}} $$ - Use this ratio to construct an industry timing strategy [14] - **Factor Evaluation**: The backtesting results show that the scissors difference timing model outperforms the respective industry indices in all cases, demonstrating excellent historical performance [14] --- Model Backtesting Results 1. Heston Model - Implied volatility for major indices: - **Shanghai 50**: 13.5% (down 0.91% from last week) - **Shanghai 500**: 15.29% (down 0.11% from last week) - **CSI 1000**: 16.79% (down 1.3% from last week) - **CSI 300**: 13.65% (down 0.83% from last week) [9] --- Factor Backtesting Results 1. Multi-Industry Timing Factor (Scissors Difference) - Backtesting results for selected industries: - **Real Estate**: Strategy annualized return 13.18%, maximum drawdown -34.3%; Index annualized return -1.21%, maximum drawdown -75.09% - **Light Manufacturing**: Strategy annualized return 21.84%, maximum drawdown -37.91%; Index annualized return 2.76%, maximum drawdown -67.79% - **Coal**: Strategy annualized return 28.73%, maximum drawdown -24.76%; Index annualized return -0.1%, maximum drawdown -69.7% - **Pharmaceuticals**: Strategy annualized return 19.22%, maximum drawdown -42.71%; Index annualized return 6.69%, maximum drawdown -55.37% [15][16]

Quantitative Models and Construction 1. Model Name: Heston Model - **Model Construction Idea**: The Heston model is used to calculate the implied volatility of near-month at-the-money options, which serves as a market fear index. It reflects market participants' expectations of future volatility[7]. - **Model Construction Process**: The Heston model is a stochastic volatility model where the variance of the asset price follows a mean-reverting square-root process. The model is defined by the following equations: $ dS_t = \mu S_t dt + \sqrt{v_t} S_t dW_t^1 $ $ dv_t = \kappa (\theta - v_t) dt + \sigma \sqrt{v_t} dW_t^2 $ Here: - \( S_t \): Asset price - \( v_t \): Variance process - \( \mu \): Drift rate of the asset price - \( \kappa \): Rate of mean reversion of variance - \( \theta \): Long-term variance - \( \sigma \): Volatility of volatility - \( W_t^1, W_t^2 \): Two Wiener processes with correlation \( \rho \)[7]. - **Model Evaluation**: The Heston model is widely used in financial markets for its ability to capture the stochastic nature of volatility, making it a robust tool for implied volatility estimation[7]. --- Quantitative Factors and Construction 1. Factor Name: Multi-Industry Timing Factor (Scissor Difference) - **Factor Construction Idea**: This factor is based on the difference between the number of stocks with bullish and bearish signals within an industry. It is used to construct an industry timing strategy[15]. - **Factor Construction Process**: - Define the number of stocks with bullish signals as \( N_{bullish} \) and bearish signals as \( N_{bearish} \). - Compute the scissor difference as: $ \text{Scissor Difference} = N_{bullish} - N_{bearish} $ - Normalize the scissor difference by the total number of stocks in the industry to obtain the scissor difference ratio: $ \text{Scissor Difference Ratio} = \frac{N_{bullish} - N_{bearish}}{N_{total}} $ - Use this ratio to construct an industry timing strategy[15]. - **Factor Evaluation**: The backtesting results show that the timing model outperforms the respective industry indices in all cases, demonstrating excellent historical performance[15]. --- Backtesting Results of Models and Factors 1. Heston Model - **Implied Volatility Results**: - SSE 50: 14.41% (+2.91% WoW)[9] - SSE 500: 15.4% (+0.83% WoW)[9] - CSI 1000: 18.09% (+1.24% WoW)[9] - CSI 300: 14.48% (+3.15% WoW)[9] 2. Multi-Industry Timing Factor - **Performance Metrics**: - The timing model outperformed the respective industry indices in all cases, with a 100% success rate in backtesting[15]. - Specific industries with bullish signals include retail, light manufacturing, home appliances, and others[18].

Quantitative Models and Construction Methods 1. Model Name: Heston Model - **Model Construction Idea**: The Heston model is used to calculate the implied volatility of near-month at-the-money options, serving as a market fear index. Implied volatility reflects market participants' expectations of future volatility [7]. - **Model Construction Process**: The Heston model is a stochastic volatility model where the variance of the asset price follows a mean-reverting square-root process. The model is defined by the following equations: $ dS_t = \mu S_t dt + \sqrt{v_t} S_t dW_t^1 $ $ dv_t = \kappa (\theta - v_t) dt + \sigma \sqrt{v_t} dW_t^2 $ Here: - \( S_t \): Asset price - \( v_t \): Variance process - \( \mu \): Drift rate of the asset price - \( \kappa \): Rate of mean reversion - \( \theta \): Long-term variance - \( \sigma \): Volatility of volatility - \( W_t^1, W_t^2 \): Two Wiener processes with correlation \( \rho \) [7]. --- Quantitative Factors and Construction Methods 1. Factor Name: Multi-Long-Short Ratio Scissor Difference - **Factor Construction Idea**: This factor is based on the number of stocks with bullish and bearish signals within an industry index. The scissor difference ratio is used to construct an industry timing strategy [15]. - **Factor Construction Process**: - Define the number of stocks with bullish signals as \( N_{bull} \) and bearish signals as \( N_{bear} \). - If \( N_{bull} = 0 \), set the bullish count to 0. Similarly, if \( N_{bear} = 0 \), set the bearish count to 0. - The scissor difference is calculated as \( N_{bull} - N_{bear} \), and the ratio is \( \frac{N_{bull} - N_{bear}}{N_{bull} + N_{bear}} \) [15]. - **Factor Evaluation**: The backtesting results show that this factor outperforms the respective industry indices in all cases, indicating excellent historical performance [15]. --- Backtesting Results of Models and Factors 1. Heston Model - **Implied Volatility Results**: - SSE 50: 13.48% (+1.11% WoW) - SSE 500: 16.97% (+0.72% WoW) - CSI 1000: 21.37% (+1.76% WoW) - CSI 300: 13.25% (-0.12% WoW) [9]. 2. Multi-Long-Short Ratio Scissor Difference - **Industry Timing Strategy Results**: - The timing model outperformed the respective industry indices in 100% of cases [15]. - **Specific Industry Examples**: - Retail Trade: Strategy Annualized Return 19.5%, Max Drawdown -43.39%, Index Annualized Return -1.49%, Max Drawdown -77.37% - Home Appliances: Strategy Annualized Return 16.27%, Max Drawdown -38.25%, Index Annualized Return 11.06%, Max Drawdown -48.96% - Comprehensive: Strategy Annualized Return 24.05%, Max Drawdown -40.81%, Index Annualized Return 0.11%, Max Drawdown -81.18% [16][18].

- The Heston model is used to calculate the implied volatility of near-month at-the-money options, serving as the market's fear index. Implied volatility reflects market participants' expectations of future volatility[9] - Positive signals appeared 3465 times between April 30, 2025, and May 9, 2025, with an average future high-point success rate of 59.95%. Negative signals appeared 2996 times, with an average future low-point success rate of 37.32%[14] - Broad-based timing strategy signals indicate bullish signals for the SSE 50 and Hang Seng Sustainable Development Enterprise Index, while other broad-based signals remain neutral[16] - Industry timing strategy signals are constructed using the scissors difference ratio of long and short positions in industry index constituent stocks. If no bullish or bearish signals appear on a given day, the scissors difference value and ratio are set to zero. This model outperforms respective industry indices in all cases, demonstrating excellent historical backtesting performance[17] - Industry timing strategy signals show bullish signals for sectors such as retail, home appliances, textiles and apparel, utilities, consumer services, transportation, and banking, while other industry signals remain neutral[20]

- The Heston model is used to calculate the implied volatility of near-month at-the-money options, serving as the market's fear index. Implied volatility reflects market participants' expectations of future volatility[7] - The shape-based timing strategy for broad-based indices shows a "bullish" signal for indices such as CSI 2000, Guozheng 2000, SSE 50, Wind Micro Cap Index, Wind All A, Hang Seng Financials, Hang Seng Hong Kong 35, Hang Seng China Enterprises Index, Hang Seng Index, Hang Seng Equal Weight, and Hang Seng Sustainable Development Enterprises Index. Other broad-based indices remain "neutral"[2][14] - The industry shape-based timing strategy is constructed using the scissors difference ratio of long and short positions in industry index constituent stocks. If no bullish or bearish stocks are present on a given day, the scissors difference value and ratio are set to zero. This model outperforms respective industry indices in all cases, demonstrating excellent historical backtesting performance[15] - Positive signals appeared 3,394 times between April 14, 2025, and April 18, 2025, with an average future high-point win rate of 65.46%. Negative signals appeared 4,047 times, with an average future low-point win rate of 35.78%[12] - Six technical stock patterns are summarized, including "Golden Needle Bottom," "Rocket Launch," "Full Red," "Hanging Line," "Heaven Line," and "Cloud Line." Positive patterns like "Golden Needle Bottom," "Rocket Launch," and "Full Red" show strong positive alert effects[26] - Stocks with special bullish shape signals this week include *ST Zhongcheng, Guozhong Water, *ST Haihua, *ST Xintong, ST Spring, and Zhongfu Shenying[27][28] - Broker gold stocks with a 70% win rate bullish shape signal this week include Aerospace Electric, Feilihua, and Qingda Environmental[31][32]