Group 1: Market Sentiment Model Insights - The market sentiment score has slightly decreased to 1.1 as of December 21, down from 1.35 the previous week, indicating a neutral view from a sentiment perspective [7][11] - There is a notable improvement in the overall sentiment index score this week, with signs of a rebound in market trading activity [7][11] - The price-volume consistency indicator has shown improvement, suggesting a recovery in market sentiment, although structural differentiation remains [11][12] Group 2: Trading Activity and Volume - The total trading volume for the entire A-share market decreased by 9.86% week-on-week, with an average daily trading volume of 17,604.84 billion yuan [15] - The highest trading volume was recorded on December 17 at 18,343.65 billion yuan, indicating a peak in market activity [15] Group 3: Industry Performance and Trends - As of December 19, industries such as beauty care, pharmaceuticals, non-bank financials, agriculture, and retail have shown upward trends in short-term scores [39] - The communication sector has the highest short-term score of 79.66, indicating strong performance potential [39][40] - The industry crowding indicator shows a strong positive correlation with weekly price changes, with sectors like retail and light manufacturing leading in gains [44] Group 4: Leverage and Risk Appetite - The proportion of financing balance continues to rise, reaching a new high for the phase, indicating an increase in leveraged funds and a structural recovery in risk appetite [27][29] - The RSI indicator has shown a slight recovery, suggesting improved short-term upward momentum, although it remains in a low range [30][33] Group 5: Style and Growth Signals - The current model indicates a preference for small-cap and growth styles, with signals suggesting that growth style may strengthen further [39][49] - The short-term view for the growth style remains positive, while the small-cap style is also favored, although there are indications of potential weakening in future signals [49]

Market Overview - The market is currently in a weak oscillation state, with both sentiment model signals and moving average strength index indicating a difficult short-term breakout from this oscillation pattern [1][2] - The liquidity shock index for the CSI 300 was 0.41, lower than the previous week’s 0.51, indicating current market liquidity is above the average level of the past year by 0.41 standard deviations [2] - The PUT-CALL ratio for the SSE 50 ETF decreased to 0.83 from 1.08, suggesting a decline in investor caution regarding the short-term performance of the SSE 50 ETF [2] - The average turnover rates for the Shanghai Composite Index and Wind All A were 1.05% and 1.60%, respectively, indicating a decrease in trading activity [2] Economic Indicators - The onshore and offshore RMB exchange rates saw weekly increases of 0.2% and 0.28%, respectively [2] - In the U.S., the non-farm payrolls increased by 64,000 in November, exceeding the market expectation of 50,000, while the unemployment rate unexpectedly rose to 4.6%, the highest since September 2021 [2] - China's economic data for November showed a 4.8% year-on-year increase in industrial value added, a 4.2% increase in the service production index, and a 1.3% increase in retail sales [2] Market Sentiment and Technical Analysis - The SAR indicator showed that the Wind All A index broke above the reversal indicator on December 1 [2] - The market score based on the moving average strength index is currently at 170, which is at the 60.6% percentile for 2023 [2] - The sentiment model score is 0 (out of 5), indicating a negative trend model signal and a negative weighted model signal [2] Market Performance - For the week of December 15-19, the SSE 50 index rose by 0.32%, while the CSI 300 index fell by 0.28%, and the ChiNext index dropped by 2.26% [3] - The current market PE (TTM) is 21.8 times, which is at the 72.7% percentile since 2005 [3] Factor and Industry Analysis - The profitability factor crowding degree continues to rise, with small-cap factor crowding at 0.22, low valuation factor crowding at -0.51, high profitability factor crowding at 0.05, and high profitability growth factor crowding at 0.22 [3] - The industry crowding degree is relatively high in telecommunications, non-ferrous metals, comprehensive, power equipment, and basic chemicals, with significant increases in crowding for defense and military industry and commercial retail [4]

Quantitative Models and Construction Methods 1. Model Name: Short-term Timing Strategy - **Model Construction Idea**: The model aims to provide short-term market timing signals by analyzing macroeconomic fundamentals, valuation metrics, sentiment indicators, and liquidity conditions. It integrates these factors into a comprehensive signal for market timing decisions [16][17] - **Model Construction Process**: 1. **Macroeconomic Fundamentals**: - Manufacturing PMI: A value above 50 indicates expansion, while below 50 indicates contraction. The latest PMI is 49.20, giving a cautious signal [16][19] - Credit Impulse: The current long-term loan pulse growth rate is at the 54.24% percentile over the past 5 years, providing a neutral signal [16][19] - M1 Growth Rate: The filtered M1 growth rate is at the 86.44% percentile, indicating strong growth and a positive signal [16][19] 2. **Valuation Metrics**: - PE Median: The current PE median is at the 93.55% percentile over the past 5 years, signaling caution [16][19] - PB Median: The current PB median is at the 90.57% percentile, also signaling caution [16][19] 3. **Sentiment Indicators**: - Beta Dispersion: At the 44.07% percentile, providing a neutral signal [17][19] - Volume Sentiment Score: At the 24.15% percentile, indicating weak sentiment and a cautious signal [17][19] - Volatility: At the 32.01% percentile, providing a neutral signal [17][19] 4. **Liquidity Conditions**: - Money Market Rate: At the 30.51% percentile, indicating relatively loose liquidity and a positive signal [17][19] - Exchange Rate Expectation: At the 30.51% percentile, indicating a strong RMB and a positive signal [17][19] - Average 5-day Financing Inflows: At the 51.70% percentile, providing a neutral signal [17][19] - **Model Evaluation**: The model demonstrates strong performance with significant annualized returns and reduced drawdowns compared to the benchmark [18][21] 2. Model Name: Growth-Value Style Rotation Model - **Model Construction Idea**: This model identifies rotation opportunities between growth and value styles based on macroeconomic cycles, valuation spreads, and sentiment differences [27][28] - **Model Construction Process**: 1. **Macroeconomic Fundamentals**: - Profit Cycle Slope: A steep slope benefits growth styles [27][29] - Interest Rate Cycle: High levels favor value styles [27][29] - Credit Cycle: Strengthening credit cycles benefit growth styles [27][29] 2. **Valuation Metrics**: - PE Spread: The 5-year percentile of the growth-value PE spread is 30.29%, favoring growth styles [27][29] - PB Spread: The 5-year percentile of the growth-value PB spread is 47.25%, also favoring growth styles [27][29] 3. **Sentiment Indicators**: - Turnover Spread: At the 40.86% percentile, indicating a neutral signal [28][29] - Volatility Spread: At the 65.89% percentile, indicating a neutral signal [28][29] - **Model Evaluation**: The model has delivered consistent annualized returns and outperformed the benchmark in most years, though it underperformed slightly in 2025 [28][30] 3. Model Name: Small-Cap vs. Large-Cap Rotation Model - **Model Construction Idea**: This model evaluates the relative performance of small-cap and large-cap stocks using 11 effective rotation indicators, including liquidity, sentiment, and valuation metrics [31][33] - **Model Construction Process**: - Key Indicators: - **Liquidity Metrics**: R007 and financing balance changes, both favoring large-cap stocks [31][33] - **Sentiment Metrics**: Theme trading sentiment and beta dispersion, both favoring large-cap stocks [31][33] - **Valuation Metrics**: PB dispersion and MACD signals, favoring large-cap stocks [31][33] - Comprehensive Signal: The model aggregates these indicators into a composite signal, which currently suggests a 100% allocation to large-cap stocks [31][33] - **Model Evaluation**: The model has consistently generated positive annualized excess returns since 2014, with strong performance in 2025 [32][33] --- Model Backtesting Results 1. Short-term Timing Strategy - **Annualized Return**: 16.37% (Benchmark: 4.76%) [18][21] - **Annualized Volatility**: 14.79% (Benchmark: 11.59%) [18][21] - **Maximum Drawdown**: 14.07% (Benchmark: 31.41%) [18][21] - **Sharpe Ratio**: 0.9641 (Benchmark: 0.2865) [18][21] - **2025 Performance**: Strategy Return: 23.60%, Benchmark Return: 13.41%, Excess Return: 10.19% [18][21] 2. Growth-Value Style Rotation Model - **Annualized Return**: 12.71% (Benchmark: 7.96%) [28][30] - **Annualized Volatility**: 20.79% (Benchmark: 20.64%) [28][30] - **Maximum Drawdown**: 43.07% (Benchmark: 44.13%) [28][30] - **Sharpe Ratio**: 0.5842 (Benchmark: 0.3782) [28][30] - **2025 Performance**: Strategy Return: 25.36%, Benchmark Return: 26.19%, Excess Return: -0.84% [28][30] 3. Small-Cap vs. Large-Cap Rotation Model - **Annualized Return**: 33.64% (Benchmark: 22.11%) [32][33] - **Annualized Excess Return**: 11.53% [32][33] - **Maximum Drawdown**: 40.70% [32][33] - **2025 Performance**: Strategy Return: 33.64%, Benchmark Return: 22.11%, Excess Return: 11.53% [32][33]

- The liquidity shock indicator based on the CSI 300 index was 0.41 on Friday, lower than the previous week (0.51), indicating that the current market liquidity is higher than the average level of the past year by 0.41 standard deviations[2] - The PUT-CALL ratio of SSE 50ETF options trading volume fluctuated downward, being 0.83 on Friday, lower than the previous week (1.08), indicating a decrease in investors' caution about the short-term trend of SSE 50ETF[2] - The five-day average turnover rates of the SSE Composite Index and Wind All A Index were 1.05% and 1.60%, respectively, at the 68.85% and 73.75% percentiles since 2005, indicating a decline in trading activity[2] - The RMB exchange rate fluctuated last week, with the onshore and offshore exchange rates rising by 0.2% and 0.28%, respectively[2] - The SAR indicator showed that the Wind All A Index broke through the reversal indicator on December 1[12] - The moving average strength index, calculated using Wind secondary industry indices, scored 170, at the 60.6% percentile since 2023[12] - The sentiment model score was 0 out of 5, with the trend model signal being negative and the weighted model signal also being negative[12] - The congestion degree of the profitability factor continued to rise, with the small-cap factor congestion degree at 0.22, low-valuation factor at -0.51, high-profitability factor at 0.05, and high-growth factor at 0.22[4][19] - The congestion degree of industries such as communication, non-ferrous metals, comprehensive, electrical equipment, and basic chemicals was relatively high, with the congestion degree of the national defense and military industry and retail industry rising significantly[4][21]

- The report introduces a quantitative model based on Convolutional Neural Networks (CNN) to analyze price-volume chart data and predict future prices. The learned features are then mapped to industry theme indices, such as the CSI 300 Non-Bank Financial Index, the CNI Commercial Satellite Communication Industry Index, and the CSI Sub-Chemical Industry Theme Index[79][81] - The construction process involves standardizing price-volume data into chart formats and applying CNNs to extract features. These features are subsequently used to allocate weights to specific industry themes[79][81] - The model's evaluation highlights its ability to capture complex patterns in price-volume data and its application in identifying promising industry themes like non-bank financials, satellites, and chemicals[79][81]

Market Performance - The Sci-Tech 50 Index decreased by 2.99% over the last five trading days, while the ChiNext Index fell by 2.26%. In contrast, the large-cap value stocks rose by 1.52%, and large-cap growth stocks declined by 1.39%. The Shanghai Stock Exchange 50 Index increased by 0.32%, and the small-cap index represented by the CSI 2000 dropped by 0.37%. The retail and non-bank financial sectors performed well, while electronics and power equipment lagged behind [1]. Risk Premium and Valuation Levels - As of December 19, 2025, the static PE of the CSI All Share Index indicates a risk premium of 2.79%, calculated as the inverse of the PE minus the yield of ten-year government bonds. The two-standard deviation boundary is at 4.71%. The valuation levels show that the CSI All Share Index's PE TTM is at the 80th percentile, with the Shanghai 50 and CSI 300 at 74% and 73%, respectively. The ChiNext Index is close to 55%, while the CSI 500 and CSI 1000 are at 59% and 60%, respectively, indicating that the ChiNext Index's valuation is relatively at the historical median level [1]. ETF Fund Flows - In the last five trading days, ETF inflows amounted to 72.1 billion yuan, while the financing balance decreased by approximately 7.6 billion yuan. The average daily trading volume across the two markets was 1.738 trillion yuan [2]. Convolutional Neural Network Observations - The analysis utilized convolutional neural networks to model charted price and volume data, mapping learned features to industry themes. The latest thematic allocations include non-bank financials, satellite communications, and chemicals, specifically focusing on the CSI 300 Non-Bank Financial Index, the CSI Commercial Satellite Communication Industry Index, and the CSI Sub-Segmented Chemical Industry Theme Index [2][11]. Index Information - The report includes specific index codes and names, such as the CSI 300 Non-Bank Financial Index, the CSI Commercial Satellite Communication Industry Index, and the CSI Sub-Segmented Chemical Industry Theme Index, among others [3][12]. Market Sentiment and Risk Preference - The report tracks the proportion of market sentiment above the 200-day long-term moving average and monitors the risk preferences between equity and bond assets [14]. Financing Balance - The financing balance data is provided, indicating trends in market leverage and investor sentiment [16]. Individual Stock Performance - Statistics on individual stock performance year-to-date based on return ranges are included, providing insights into market dynamics [18]. Oversold Indices - The report notes instances of oversold conditions in certain indices, which may indicate potential buying opportunities [20].

Group 1 - Market sentiment score continued to decline, reaching 1.35 as of December 12, down from 2.4 the previous week, indicating a bearish outlook from a sentiment perspective [2][8] - The overall trading volume in the market increased significantly, with total trading volume for the week rising by 15.14% compared to the previous week, averaging 19,530.44 billion yuan per day, with a peak of 21,190.10 billion yuan on December 12 [14][16] - The industry score model indicates that sectors such as non-bank financials, communication, defense, and automotive are showing upward trends in short-term scores, with communication having the highest short-term score of 77.97 [40][41] Group 2 - The correlation between industry congestion and weekly price changes is strong, with a coefficient of 0.33, indicating that sectors with high congestion like communication and defense are leading in gains, while sectors with low congestion like steel and environmental protection are lagging [45][46] - The current model suggests a preference for large-cap and growth styles, with signals indicating that growth style may strengthen further in the future [40][51] - The financing balance ratio continues to rise, reaching a new high for the phase, indicating an increase in leveraged funds and a structural recovery in risk appetite [26][28]

- The report indicates that the market is on the edge of an upward trend signal, with the core observation indicator being whether the profitability effect is positive. The current trend line of the WIND All A Index is around 6262 points, and the closing price is at 6264 points, just on the verge of turning positive[2][5][7] - The timing system signal shows that the distance between the moving averages is 4.03%, significantly greater than the absolute value of 3%, indicating that the market has returned to an upward trend pattern[2][5][6] - The industry trend allocation model shows that the mid-term distress reversal expectation model signals attention to liquor and real estate; the TWO BETA model continues to recommend the technology sector, focusing on consumer electronics and domestic computing power. The industry trend model shows that the engineering machinery/industrial metals/energy storage sectors continue their upward trend[2][5][7] - From the valuation indicators, the PE of the WIND All A Index is around the 80th percentile, which is a medium level, and the PB is around the 50th percentile, which is a relatively low level. Based on the short-term trend judgment and the position management model, it is recommended that absolute return products with the WIND All A as the main stock allocation should have a position of 60%[5][7][12]

- The report introduces a convolutional neural network (CNN) model for analyzing price-volume data and predicting future stock prices. The model maps learned features to industry theme indices such as communication, artificial intelligence, and growth momentum indices[4][83][85] - The CNN model constructs standardized graphical representations of price-volume data for individual stocks within specific time windows. These graphical representations are then used for deep learning-based modeling[83][84] - The latest thematic configurations derived from the CNN model include indices such as CSI Communication Equipment Theme Index, ChiNext Artificial Intelligence Index, CSI 5G Communication Theme Index, and ChiNext Growth Momentum Index[4][85][86] - The report evaluates the CNN model as a promising approach for integrating AI into quantitative analysis, particularly for thematic investment strategies[83][86] - Backtesting results and specific performance metrics for the CNN model are not explicitly provided in the report[83][86]

Quantitative Models and Construction Methods 1. Model Name: Short-term Timing Strategy - **Model Construction Idea**: The model integrates macroeconomic fundamentals, valuation, sentiment, and liquidity signals to generate short-term timing recommendations for the A-share market[16][18][19] - **Model Construction Process**: - **Macroeconomic Fundamentals**: - Manufacturing PMI: A PMI value above 50 indicates economic expansion, while below 50 indicates contraction. The latest PMI is 49.20, signaling caution[16][19] - Credit Impulse: The long-term loan pulse growth rate is at the 54.24% percentile over the past 5 years, indicating a neutral signal[16][19] - M1 Growth Rate: The filtered M1 growth rate is at the 86.44% percentile over the past 5 years, signaling optimism[16][19] - **Valuation**: - PE Median: The A-share PE median is at the 93.47% percentile over the past 5 years, signaling caution[17][19] - PB Median: The A-share PB median is at the 88.92% percentile over the past 5 years, signaling caution[17][19] - **Sentiment**: - Beta Dispersion: At the 44.07% percentile over the past 5 years, indicating a neutral signal[17][19] - Volume Sentiment Score: At the 39.12% percentile over the past 5 years, signaling caution[17][19] - Volatility: At the 56.00% percentile over the past 5 years, indicating a neutral signal[17][19] - **Liquidity**: - Money Market Rate: At the 30.51% percentile over the past 5 years, indicating optimism[18][19] - Exchange Rate Expectation: At the 30.51% percentile over the past 5 years, signaling optimism[18][19] - Average 5-day Financing Amount: At the 47.15% percentile over the past 5 years, indicating a neutral signal[18][19] - **Model Evaluation**: The model demonstrates strong performance with significant annualized returns and reduced drawdowns compared to the benchmark, showcasing its robustness in short-term market timing[18][23] 2. Model Name: Growth-Value Style Rotation Model - **Model Construction Idea**: The model evaluates macroeconomic cycles, valuation spreads, and sentiment differences to determine the optimal allocation between growth and value styles[27][28] - **Model Construction Process**: - **Macroeconomic Fundamentals**: - Profit Cycle Slope: A steep slope favors growth[28][29] - Interest Rate Cycle: High levels favor value[28][29] - Credit Cycle: Strengthening credit cycles favor growth[28][29] - **Valuation**: - PE Spread: The growth-value PE spread is at the 34.76% percentile, favoring growth[29] - PB Spread: The growth-value PB spread is at the 41.12% percentile, favoring growth[29] - **Sentiment**: - Turnover Spread: At the 75.52% percentile, favoring growth[29] - Volatility Spread: At the 67.92% percentile, favoring a balanced allocation[29] - **Model Evaluation**: The model has delivered consistent annualized returns and reduced drawdowns compared to the benchmark, though recent performance has shown slight underperformance[28][30] 3. Model Name: Small-Cap vs. Large-Cap Style Rotation Model - **Model Construction Idea**: The model uses 11 effective rotation indicators, including liquidity, sentiment, and valuation metrics, to determine the optimal allocation between small-cap and large-cap stocks[31][33] - **Model Construction Process**: - **Key Indicators**: - Indicators such as R007, financing balance changes, and thematic trading sentiment currently favor large-cap stocks[31][33] - **Comprehensive Signal**: The model aggregates individual signals to generate a composite recommendation, which currently suggests overweighting large-cap stocks[31][33] - **Model Evaluation**: The model has consistently generated positive annualized excess returns since 2014, demonstrating its effectiveness in capturing style rotation opportunities[32][33] --- Model Backtesting Results 1. Short-term Timing Strategy - **Annualized Return**: 16.40% (benchmark: 4.77%)[18][23] - **Annualized Volatility**: 14.80% (benchmark: 11.59%)[23] - **Maximum Drawdown**: 14.07% (benchmark: 31.41%)[23] - **Sharpe Ratio**: 0.9651 (benchmark: 0.2876)[23] - **2025 YTD Return**: 23.60% (benchmark: 13.49%)[18][23] 2. Growth-Value Style Rotation Model - **Annualized Return**: 12.74% (benchmark: 7.97%)[28][30] - **Annualized Volatility**: 20.80% (benchmark: 20.66%)[30] - **Maximum Drawdown**: 43.07% (benchmark: 44.13%)[30] - **Sharpe Ratio**: 0.5853 (benchmark: 0.3785)[30] - **2025 YTD Return**: 25.13% (benchmark: 25.96%)[28][30] 3. Small-Cap vs. Large-Cap Style Rotation Model - **Annualized Return**: 19.73% (benchmark: 12.67%)[33] - **Maximum Drawdown**: 40.70% (benchmark: 44.32%)[33] - **2025 YTD Return**: 33.83% (benchmark: 22.54%)[32][33]