Core Insights - The report presents a systematic framework for automated factor research based on Large Language Models (LLM) and Prompt Engineering, aiming to explore the potential applications of AI in the entire quantitative investment chain [1] - The framework was first applied to low-frequency price-volume factors, optimizing the classic Alpha158 factor library and transitioning from an "optimization" paradigm to a "generation" paradigm [1] - AI demonstrated strong factor discovery capabilities in both fundamental and high-frequency data domains, successfully generating new factors and enhancing traditional factor libraries [1] - The report also explores AI's application in unstructured text analysis, utilizing the Gemini model to interpret sentiment from extensive research memos, creating unique sentiment indicators that effectively integrate into stock selection strategies [1] Group 1: Low-Frequency Price-Volume Factor Optimization - The framework was initially applied to the optimization of low-frequency price-volume factors, using the Alpha158 factor library as a foundation for optimization experiments [1] - AI identified logical flaws in original factors and proposed effective improvements, with optimization effects being consistent across multiple time windows from 5 to 60 days [1] - New factors generated by AI, with low correlation to sample factors, showed robust out-of-sample performance, with some factors achieving an Information Coefficient Information Ratio (ICIR) above 1.0 [1] Group 2: Fundamental and High-Frequency Factor Discovery - In the fundamental dimension, AI not only generated enhanced versions of classic factors but also innovatively expanded value, quality, and growth factors from novel perspectives [1] - In the high-frequency dimension, AI was empowered to directly generate Python code, uncovering a set of novel and high-performing high-frequency factors, with some strong signal factors achieving annualized returns exceeding 60% [1] - Integrating the AI-generated high-frequency factor library into the AGRU neural network model significantly improved annualized excess returns from 18.24% to 25.28% [1] Group 3: Alternative Data Processing and Sentiment Analysis - The report investigates AI's potential in processing alternative data, analyzing nearly one million words of research memos using the Gemini 2.5 Pro model [1] - A weekly sentiment factor was constructed, revealing unique asymmetric predictive capabilities, where negative sentiment strongly predicted future price declines, achieving annualized excess returns of 8.26% [1] - This sentiment factor exhibited low correlation with traditional price-volume and fundamental factors, serving as an independent and effective supplementary information source [1] Group 4: Comprehensive Strategy Development - A multi-dimensional information fusion strategy was developed, integrating AI-discovered high-frequency factors with low-frequency market data into the AGRU neural network to form a core Alpha [1] - The final strategy, enhanced by AI sentiment factors for risk adjustment, improved annualized excess returns from 11.15% to 11.81% while maintaining turnover rates [1] - The strategy demonstrated a significant increase in the information ratio from 2.18 to 2.31, validating AI's potential to empower quantitative research across multiple stages and achieve a "1+1>2" effect [1]

Group 1: Core Insights - The report emphasizes the classification of factors in commodity futures into six categories, including momentum, term structure, volume-price, positions, inventory, and volatility, forming a comprehensive analysis framework from market sentiment to fundamental supply-demand dynamics [2]. - The empirical results indicate that the term structure factor, particularly the roll yield, is the most robust predictor of alpha returns, while the inventory factor shows significant potential as a contrarian indicator, especially in extreme supply-demand conditions [3][4]. - The report highlights the frequency dependence of factor effectiveness, suggesting that different strategies should be employed based on the time dimension, with specific combinations of factors recommended for monthly and weekly strategies [5]. Group 2: Factor Selection - For monthly strategies, effective combinations include roll yield, contrarian inventory levels, and liquidity factors, along with some effective momentum factors [5]. - Weekly strategies should focus on roll yield and contrarian inventory levels as core components, supplemented by skewness and volatility estimation factors [5]. - The report notes that all momentum factors become ineffective at the weekly frequency, underscoring the necessity of dynamically adjusting the factor library based on trading frequency [5].

Quantitative Models and Construction Methods 1. Model Name: Predatory Trading Game with Disinformation - **Model Construction Idea**: This model incorporates disinformation into a predatory trading game framework, where participants act based on distorted information, leading to deviations in equilibrium and market volatility[3][16][23] - **Model Construction Process**: 1. The model builds on the microstructure frameworks of Carlin et al. (2007) and Carmona & Yang (2011), introducing a victim (forced to adjust risky asset positions) and predators (seeking profit from the victim's constraints)[23] 2. The trading rate of participant \( n \) is defined as: $$ X^{n}(t) = X^{n}(0) + \int_{0}^{t}\alpha^{n}(s)\mathrm{d}s \tag{1} $$ where \( \alpha^{n} \) represents the trading rate, constrained by: $$ \alpha_{t}^{n} \in \mathbb{A}^{n} = \left\{\alpha_{t}^{n} \mid \mathcal{H}_{[0,T]}^{2}, X_{T}^{n} = 0 \right\} $$ 3. Temporary price impact is modeled as: $$ P_{t} - X_{t}^{0} = \lambda \sum_{i=1}^{N}\alpha_{t}^{i} \tag{4} $$ where \( \lambda \) is the elasticity factor[24] 4. Permanent price impact is expressed as: $$ \mathrm{d}X_{t}^{0} = \gamma \sum_{i=1}^{N}a_{t}^{i}\mathrm{d}t + \sigma\mathrm{d}W_{t} \tag{5} $$ where \( \gamma \) represents market plasticity, and \( \sigma \) is the volatility parameter[24] 5. Participants aim to maximize profits: $$ J^{n}(\mathbf{\alpha}) = \mathbb{E}\left(\int_{0}^{T}\alpha^{n}\left(X_{t}^{0} + \lambda\sum_{i=1}^{N}\alpha_{t}^{i}\right)\mathrm{d}t\right) \tag{8} $$ 6. Disinformation is introduced as a random distortion \( \tilde{x}_{0,1} = x_{0,1} + \epsilon \), where \( \epsilon \) represents the distortion[27] 7. The price process under disinformation is given by: $$ X_{t}^{0} = X^{0}(0) - \frac{1-e^{-\frac{N-1}{N+1}\frac{T_{T}}{\lambda}}}{1-e^{-\frac{N-1}{N+1}\frac{T_{T}}{\lambda}}}\gamma\left(\sum_{i=1}^{N}x_{0}^{i}+\bar{\nu}\right) + \frac{e^{\frac{T t}{\lambda}}-1}{e^{\frac{T t}{\lambda}}-1}\gamma\bar{\nu} + \sigma\left(W_{t}-W_{0}\right) $$ where \( \bar{\nu} \) is the error factor[30][31] - **Model Evaluation**: The model effectively captures the impact of disinformation on market dynamics, highlighting its role in amplifying volatility and disrupting equilibrium[16][30] --- Model Backtesting Results 1. Predatory Trading Game with Disinformation - **Maximum Price Fluctuation (MPF)**: $$ MPF_{\nu}(t_{*},t^{*}) := \operatorname*{max}_{t_{1},t_{2}\in[t_{*},t^{*}]}\left|\mathbb{E}\left(X_{t_{1}}^{0}-X_{t_{2}}^{0}\right)\right| $$ The model demonstrates that disinformation increases MPF, with a lower bound determined by: $$ MPF_{\tilde{\nu}^{*}}(0,T) \geq \operatorname*{min}_{\tilde{\nu}\in\mathbb{R}} MPF_{\tilde{\nu}}(0,T) = \gamma\sum_{i=1}^{N}x_{0}^{i} $$[34][37] - **Error Factor Impact**: The error factor \( \nu \) significantly influences price trajectories, with higher \( \nu \) leading to greater volatility[30][33] - **Tolerance Thresholds**: The system tolerates disinformation within specific boundaries \( b_{1} \) and \( b_{2} \), beyond which volatility escalates[38][40] --- Quantitative Factors and Construction Methods 1. Factor Name: Error Factor (\( \nu \)) - **Factor Construction Idea**: The error factor quantifies the degree and spread of disinformation in the market, serving as a key determinant of price volatility[30][33] - **Factor Construction Process**: 1. Defined as: $$ \tilde{\nu} := \frac{N_{w}}{N}\left(\tilde{x}_{0}^{1} - x_{0}^{1}\right) $$ where \( N_{w} \) is the number of misinformed participants, and \( \tilde{x}_{0}^{1} - x_{0}^{1} \) represents the distortion magnitude[30] 2. Generalized for multiple distortions: $$ \nu := \frac{1}{N}\,\sum_{l=1}^{\kappa}N_{w_{l}}\left(\bar{x}_{0,w_{l}}^{1} - x_{0}^{1}\right) $$ where \( \kappa \) is the number of distinct distortions[56] - **Factor Evaluation**: The error factor effectively captures the interplay between disinformation magnitude and its spread, providing insights into its impact on market dynamics[30][56] --- Factor Backtesting Results 1. Error Factor (\( \nu \)) - **Maximum Price Fluctuation (MPF)**: Higher \( \nu \) values correspond to increased MPF, with a minimum threshold determined by: $$ MPF_{\nu}(0,T) \geq \gamma\sum_{i=1}^{N}x_{0}^{i} $$[34][37] - **Tolerance Thresholds**: The system tolerates \( \nu \) within boundaries \( b_{1} \) and \( b_{2} \), with specific dependencies on market parameters and game duration[38][40] - **Dynamic Evolution**: The tolerance for \( \nu \) increases over time, reducing the potential for disinformation to amplify volatility in the long term[90][91] --- Additional Insights - **Information Updates**: New information can mitigate the impact of disinformation by adjusting the error factor \( \nu \), with the timing of updates being critical to minimizing volatility[84][92][95] - **Randomness and Misjudgment**: Random price movements can lead even informed participants to misjudge their information, complicating the detection and correction of disinformation[100][101][103] - **Profit Implications**: Disinformation affects profit expectations, with informed participants benefiting under certain conditions, while widespread disinformation can erode these advantages[49][51][56]

Quantitative Models and Construction Methods National Enterprise Fundamental Factor Stock Selection Strategy - **Model Name**: National Enterprise Fundamental Factor Stock Selection Strategy [3] - **Model Construction Idea**: The strategy is based on fundamental factors tailored to national enterprises, considering both general and industry-specific factors [5][6] - **Model Construction Process**: - Define the sample pool using the CSI National Enterprise Index (000955.CSI) and stocks listed on the Beijing Stock Exchange for over six months with central or local state-owned enterprise attributes [3] - Classify industries into dividend-oriented and growth-oriented categories based on ZX third-level industry logic [3][4] - Select general factors such as ROE (TTM), operating cash ratio, labor productivity, asset-liability ratio, and dividend yield [5][6] - Incorporate industry-specific factors like ROIC, prepayment growth rate, inventory turnover rate, and capital expenditure/depreciation ratio for different industries [6][8] - Adjust factor weights based on industry characteristics, emphasizing dividend yield for dividend-oriented industries and reducing the weight of asset-liability ratio for growth-oriented industries [9] - Calculate scores using weighted averages of general and industry-specific factors, normalize the scores, and assign weights to stocks based on their scores [11] - Formula for stock weight: $$w_{i}={\frac{s c o r e_{i}^{3}}{\sum_{i=1}^{N}s c o r e_{i}^{3}}}$$ [11] - **Model Evaluation**: The strategy effectively captures the characteristics of national enterprises, balancing dividend stability and growth potential [5][6] Technology Theme Fundamental Factor Stock Selection Strategy - **Model Name**: Technology Theme Fundamental Factor Stock Selection Strategy [19] - **Model Construction Idea**: Focus on technology stocks with high R&D investment and strong growth potential, using fundamental factors to identify stocks in their growth and mature stages [20][23] - **Model Construction Process**: - Define the sample pool based on SW third-level industries and R&D investment criteria (R&D expenses > 5% of revenue or R&D personnel > 10% of total employees) [19][20] - Exclude stocks in the shock and decline stages based on cash flow lifecycle analysis [22][23] - Select general factors such as profitability, growth ability, technical level, supply chain concentration, and alpha factors [24][28] - Incorporate specific factors for growth and mature stages, such as management expense ratio, R&D expense ratio, accounts receivable turnover rate, and PB-ROE [24][28] - Adjust scores using R&D expense multipliers to emphasize high R&D industries [28][29] - Formula for stock weight: $$w e i g h t_{i}={\frac{s c o r e_{i}}{\sum_{i=1}^{50}s c o r e_{i}}}$$ [30] - **Model Evaluation**: The strategy highlights technology stocks with strong R&D capabilities and growth potential, effectively capturing industry-specific dynamics [24][28] Consumer Theme Fundamental Factor Stock Selection Strategy - **Model Name**: Consumer Theme Fundamental Factor Stock Selection Strategy [38] - **Model Construction Idea**: Focus on consumer stocks with direct-to-consumer business models, using fundamental factors to identify stocks with strong growth, profitability, and governance [38][39] - **Model Construction Process**: - Define the sample pool based on SW third-level industries, categorizing stocks into daily manufacturing, optional manufacturing, daily services, and optional services [38][39] - Select general factors such as growth-profitability-cash flow composite factor, operating cash flow ratio, ESG management score, and economic sensitivity [40][41] - Incorporate specific factors like market share, R&D expense ratio, accounts receivable turnover rate, and marketing expense ratio [40][41] - Adjust scores using PS (TTM) multipliers to emphasize stocks with lower price-to-sales ratios [46][47] - Formula for stock weight: $$w e l g h t_{i}={\frac{S c o r e_{i}^{a d j}}{\sum_{i=1}^{50}S c o r e_{i}^{a d j}}}$$ [48] - **Model Evaluation**: The strategy effectively identifies consumer stocks with strong fundamentals and growth potential, balancing profitability and governance [40][41] --- Model Backtesting Results National Enterprise Fundamental Factor Stock Selection Strategy - **Annualized Return**: 22.93% [12][15] - **Annualized Volatility**: 20.85% [15] - **Sharpe Ratio**: 1.0961 [15] - **Calmar Ratio**: 0.9963 [15] - **Maximum Drawdown**: -23.01% [15] Technology Theme Fundamental Factor Stock Selection Strategy - **Annualized Return**: 30.61% [31][34] - **Annualized Volatility**: 27.61% [34] - **Sharpe Ratio**: 1.1070 [34] - **Calmar Ratio**: 0.8962 [34] - **Maximum Drawdown**: -34.16% [34] Consumer Theme Fundamental Factor Stock Selection Strategy - **Annualized Return**: 24.86% [49][52] - **Annualized Volatility**: 22.99% [52] - **Sharpe Ratio**: 1.0825 [52] - **Calmar Ratio**: 1.0197 [52] - **Maximum Drawdown**: -24.38% [52]

- The DecompGRU model was used to construct a weekly long-only stock selection portfolio, holding the top 200 stocks with the highest integrated scores based on the model. The portfolio is rebalanced weekly on the first trading day, using factor values updated after the previous Friday's close. Stocks from the CSI All Share Index are selected, excluding stocks with trading halts or price limits, and transaction costs are not considered. The benchmark for comparison is the CSI All Share equal-weighted index[7][9] - The DecompGRU model's stock scores were aggregated to construct an ETF rotation portfolio. The ETF pool is limited to industry and thematic ETFs, retaining only the ETF with the highest 5-day average trading volume if multiple ETFs track the same index. ETFs must meet minimum trading volume criteria (5-day average > 20 million and 20-day average > 10 million). The portfolio holds 2-6 ETFs per period and is rebalanced weekly without a fixed schedule. The benchmark for comparison is the Wind ETF Index[10][12] - A sentiment factor was constructed using user behavior data from Tonghuashun, aggregating stock-level heat metrics (browsing, watchlist, and click counts) normalized by market share on the same day and multiplied by 10,000. The sentiment factor serves as a proxy for "emotional heat" at the broader index, industry, and concept levels[14] - A simple rotation strategy was built based on weekly heat change rates (MA2 smoothed) for broad-based indices. On the last trading day of each week, the strategy buys the index with the highest heat change rate. If the "Others" group has the highest rate, the strategy remains in cash. The strategy's annualized return since 2017 is 8.74%, with a maximum drawdown of 23.5%. In 2025, the strategy achieved a return of 36.8%, compared to the benchmark's 35%[20][23] - A concept-level sentiment strategy was constructed by selecting the top 5 concepts with the highest weekly heat change rates. Stocks within these concepts were filtered to exclude the bottom 20% by market capitalization. Two portfolios were created: a "TOP" portfolio holding the top 10 stocks by total heat within each concept, and a "BOTTOM" portfolio holding the bottom 10 stocks. The BOTTOM portfolio achieved an annualized return of 15.71% with a maximum drawdown of 28.89%. In 2025, the BOTTOM portfolio returned 41.8%[40][41] - The DecompGRU TOP200 portfolio achieved a cumulative absolute return of 60.48% and an excess return of 34.62% relative to the CSI All Share equal-weighted index since its inception on March 31, 2025. The portfolio's maximum drawdown was 10.08%, with weekly and monthly win rates of 67.50% and 100%, respectively. In December 2025, the portfolio's absolute return was 7.57%, with an excess return of 5.46%[9] - The ETF rotation portfolio achieved a cumulative absolute return of 26.23% and an excess return of 1.56% relative to the Wind ETF Index since its inception on March 18, 2025. The portfolio's maximum drawdown was 7.82%, with weekly and monthly win rates of 60.98% and 66.67%, respectively. In December 2025, the portfolio's absolute return was 2.35%, with an excess return of -1.51%[12][13]

Quantitative Models and Construction Methods - **Model Name**: Timing Radar Hexagon **Model Construction Idea**: The model evaluates equity market performance through a multi-dimensional framework, incorporating liquidity, economic fundamentals, valuation, capital flows, technical trends, and crowding indicators. These dimensions are summarized into four categories: "Valuation Cost-Effectiveness," "Macroeconomic Fundamentals," "Capital & Trend," and "Crowding & Reversal," generating a composite timing score within the range of [-1, 1][1][6][8] **Model Construction Process**: 1. Select 21 indicators across six dimensions (liquidity, economic fundamentals, valuation, capital flows, technical trends, and crowding)[1][6] 2. Aggregate these indicators into four categories: - Valuation Cost-Effectiveness - Macroeconomic Fundamentals - Capital & Trend - Crowding & Reversal 3. Normalize the composite score to fall within the range of [-1, 1][6][8] **Model Evaluation**: The model provides a comprehensive view of market conditions, offering a balanced perspective across multiple dimensions[6][8] Model Backtesting Results - **Timing Radar Hexagon**: - Current composite score: -0.11 (down from -0.01 last week)[6][8] - Liquidity score: 0.25 (neutral to slightly positive)[6][8] - Economic fundamentals score: -0.50 (neutral to slightly negative)[6][8] - Valuation score: -0.51 (neutral to slightly negative)[6][8] - Capital flows score: 1.00 (positive)[6][8] - Technical trends score: 0.00 (neutral)[6][8] - Crowding score: -0.75 (neutral to slightly negative)[6][8] Quantitative Factors and Construction Methods Liquidity Factors 1. **Factor Name**: Monetary Direction Factor **Construction Idea**: Measures the direction of monetary policy using central bank policy rates and short-term market rates. A positive factor value indicates monetary easing, while a negative value indicates tightening[10] **Construction Process**: - Calculate the average change in policy rates and short-term rates over the past 90 days - Assign a score of 1 if the factor > 0 (easing), and -1 if < 0 (tightening)[10] **Evaluation**: Effectively captures monetary policy direction[10] 2. **Factor Name**: Monetary Intensity Factor **Construction Idea**: Based on the "interest rate corridor" concept, measures the deviation of short-term market rates from policy rates[12] **Construction Process**: - Compute deviation = DR007/7-year reverse repo rate - 1 - Smooth and z-score the deviation - Assign a score of 1 if the factor < -1.5 standard deviations (easing), and -1 if > 1.5 standard deviations (tightening)[12] **Evaluation**: Captures the intensity of monetary policy changes[12] 3. **Factor Name**: Credit Direction Factor **Construction Idea**: Reflects the trend in credit transmission to the real economy using long-term loan data[15] **Construction Process**: - Calculate the year-over-year growth of long-term loans over the past 12 months - Assign a score of 1 if the factor shows an upward trend compared to three months ago, and -1 if downward[15] **Evaluation**: Tracks credit trends effectively[15] 4. **Factor Name**: Credit Intensity Factor **Construction Idea**: Measures whether credit data significantly exceeds or falls short of expectations[19] **Construction Process**: - Compute (new RMB loans - median forecast) / forecast standard deviation - Assign a score of 1 if the factor > 1.5 standard deviations (positive surprise), and -1 if < -1.5 standard deviations (negative surprise)[19] **Evaluation**: Captures credit surprises effectively[19] Economic Factors 1. **Factor Name**: Growth Direction Factor **Construction Idea**: Based on PMI data, measures the trend in economic growth[23] **Construction Process**: - Calculate the 12-month average and year-over-year change of PMI data - Assign a score of 1 if the factor shows an upward trend compared to three months ago, and -1 if downward[23] **Evaluation**: Tracks economic growth trends effectively[23] 2. **Factor Name**: Growth Intensity Factor **Construction Idea**: Measures whether economic growth data significantly exceeds or falls short of expectations[26] **Construction Process**: - Compute (PMI - median forecast) / forecast standard deviation - Assign a score of 1 if the factor > 1.5 standard deviations (positive surprise), and -1 if < -1.5 standard deviations (negative surprise)[26] **Evaluation**: Captures growth surprises effectively[26] 3. **Factor Name**: Inflation Direction Factor **Construction Idea**: Measures the trend in inflation using CPI and PPI data[29] **Construction Process**: - Compute 0.5 × smoothed CPI year-over-year + 0.5 × raw PPI year-over-year - Assign a score of 1 if the factor shows a downward trend compared to three months ago, and -1 if upward[29] **Evaluation**: Tracks inflation trends effectively[29] 4. **Factor Name**: Inflation Intensity Factor **Construction Idea**: Measures whether inflation data significantly exceeds or falls short of expectations[32] **Construction Process**: - Compute (CPI or PPI - median forecast) / forecast standard deviation - Assign a score of 1 if the factor < -1.5 standard deviations (negative surprise), and -1 if > 1.5 standard deviations (positive surprise)[32] **Evaluation**: Captures inflation surprises effectively[32] Valuation Factors 1. **Factor Name**: Shiller ERP **Construction Idea**: Adjusts earnings for inflation and calculates the equity risk premium (ERP) relative to 10-year government bond yields[35] **Construction Process**: - Compute Shiller PE = inflation-adjusted average earnings over the past 6 years - Calculate ERP = 1/Shiller PE - 10-year bond yield - Normalize using a 6-year z-score[35] **Evaluation**: Provides a robust measure of equity valuation[35] 2. **Factor Name**: PB **Construction Idea**: Measures valuation using the price-to-book ratio[38] **Construction Process**: - Compute PB × (-1) - Normalize using a 6-year z-score, truncating at ±1.5 standard deviations[38] **Evaluation**: Tracks valuation effectively[38] 3. **Factor Name**: AIAE **Construction Idea**: Measures aggregate investor allocation to equities, reflecting market risk appetite[41] **Construction Process**: - Compute AIAE = total market cap of CSI All Share Index / (total market cap + total debt) - Normalize using a 6-year z-score[41] **Evaluation**: Captures market risk appetite effectively[41] Capital Flow Factors 1. **Factor Name**: Margin Trading Increment **Construction Idea**: Measures the trend in leveraged funds using margin trading data[44] **Construction Process**: - Compute the 120-day average increment of margin trading balances - Assign a score of 1 if the 120-day increment > 240-day increment, and -1 otherwise[44] **Evaluation**: Tracks leveraged fund trends effectively[44] 2. **Factor Name**: Turnover Trend **Construction Idea**: Measures market activity using turnover data[47] **Construction Process**: - Compute log turnover moving average distance = ma120/ma240 - 1 - Assign a score of 1 if max(10, 30, 60-day) > 0, and -1 otherwise[47] **Evaluation**: Captures market activity effectively[47] 3. **Factor Name**: China Sovereign CDS Spread **Construction Idea**: Reflects foreign investors' perception of China's credit risk[50] **Construction Process**: - Compute the 20-day difference of smoothed CDS spreads - Assign a score of 1 if the difference < 0, and -1 otherwise[50] **Evaluation**: Tracks foreign investor sentiment effectively[50] 4. **Factor Name**: Overseas Risk Aversion Index **Construction Idea**: Captures global risk appetite using the Citi RAI Index[53] **Construction Process**: - Compute the 20-day difference of smoothed RAI - Assign a score of 1 if the difference < 0, and -1 otherwise[53] **Evaluation**: Tracks global risk appetite effectively[53] Technical Factors 1. **Factor Name**: Price Trend **Construction Idea**: Measures market trends using moving average distances[56] **Construction Process**:

Core Insights - The report highlights the "Davis Double Click Strategy," which involves buying stocks with low price-to-earnings (PE) ratios that have growth potential, and selling them once their growth is realized, leading to a multiplier effect on returns [3] - The "Net Profit Discontinuity Strategy" has achieved an annualized return of 29.42% since 2010, with a year-to-date absolute return of 67.17% and an excess return of 36.79% over the benchmark index [9][10] - The "Enhanced CSI 300 Portfolio" is constructed based on investor preference factors, showing a stable historical excess return of 20.92% relative to the CSI 300 index this year [12][16] Group 1: Davis Double Click Strategy - The strategy has generated a back-tested annualized return of 26.45% from 2010 to 2017, outperforming the benchmark by 21.08% [3][7] - In 2025, the strategy's cumulative absolute return is reported at 55.89%, exceeding the CSI 500 index by 25.50% [7][10] - The strategy's performance is characterized by stability, with excess returns exceeding 11% in each of the seven complete years during the back-test period [3][7] Group 2: Net Profit Discontinuity Strategy - This strategy focuses on stocks that show a significant upward price gap on the first trading day after earnings announcements, indicating market approval of the earnings report [9][10] - The strategy has achieved a cumulative absolute return of 67.17% in the current year, with an excess return of 36.79% over the benchmark index [10][11] - Historical performance shows that the strategy has an annualized return of 29.42% and a stable excess return of 26.22% over the benchmark since its inception [9][10] Group 3: Enhanced CSI 300 Portfolio - The portfolio is designed based on investor preferences, including GARP, growth, and value investors, aiming to identify undervalued stocks with strong profitability [12][16] - The portfolio has shown a relative excess return of 20.92% against the CSI 300 index this year, with a monthly excess return of 2.58% [12][16] - The strategy's performance is supported by a historical back-test that indicates stable excess returns [12][16]

01 主题：茶饮板块更新&茶饮龙头新动作背后的思考 时间： 12月29日（周一）15 : 00 主讲：郭梦婕 食饮首席分析师 林叙希 食饮分析师 参会密码：882583 02 主 题：新能源龙头的新机会系列之5 时间：12月30日（周二）15 : 30 主讲： 刘强 院长助理&电新首席分析师 参会密码： 781535 03 主 题：电子行业1月投资观点更新 时间： 12月31日（周三）15 : 00 主讲： 张世杰 电子首席分析师 参会密码： 181319 04 主题： 行业配置模型回顾与更新系列（十九） 时间： 1月4日（周日）9 : 00 主讲： 刘晓锋 金工首席分析师 参会密码： 951676 05 主题：人形机器人飞轮启动，2026爆款时刻元年 时间： 1月7日（周三）15 : 30 主讲：刘虹辰 汽车首席分析师 参会密码：713893 ER 17 梦 設 nx 婷 行业首席分析 上分析 IIT 识别二维码立即参会 06 主题： 社服行业近况更新 时间：1月9日（周五）15:00 主讲： 王湛 社服分析师 参会密码： 981559 会议号码: +86-4001888938 (中国) +86-01053 ...

Core Insights - The report highlights that the barbell strategy is superior, with a significant excess return from the selected electronic sector combination [1][5] - The A-share market experienced fluctuations, with micro-cap stocks leading gains and the CSI Dividend Index performing strongly, while the ChiNext index showed a notable decline [1][6] Strategy Tracking Dividend Strategy - The A-share market showed volatility, with micro-cap stocks leading the gains and the CSI Dividend Index performing well, while the ChiNext index faced a significant pullback [6][13] - Within the dividend sub-sectors, the dividend value category outperformed pure dividend assets [6][13] - The Central State-Owned Enterprises High Dividend 30 combination and the Balanced Dividend 50 combination slightly underperformed against the CSI Dividend Total Return Index this week [6][19] Electronic Sector - The A-share market displayed clear differentiation in returns among sectors, with essential consumer and financial sectors rebounding significantly, achieving over 2% excess returns relative to the entire A-share market [6][22] - The electronic sector's internal performance showed that display panels outperformed other sub-sectors [6][22] - The selected electronic sector enhancement combination outperformed the electronic total return index, with a weekly excess return of approximately 1.57%, placing it in the top 33% of active technology products [6][29]

- The report introduces a "Total Heat Index" indicator, which aggregates the browsing, self-selection, and click counts of individual stocks, normalized by their market share on the same day, and then multiplied by 10,000, with a value range of [0,10000][7] - The "Total Heat Index" is used as a proxy variable for "emotional heat" to track the overall market sentiment for broad-based indices, industries, and concepts[7] - The report constructs a simple rotation strategy based on the weekly heat change rate (MA2) of broad-based indices, buying the index with the highest heat change rate at the end of each week, and staying out of the market if the highest change rate belongs to the "others" group[13][15] - The rotation strategy based on the heat change rate (MA2) has an annualized return of 8.74% since 2017, with a maximum drawdown of 23.5%, and a return of 31.42% in 2025[16] - The report also constructs two simple portfolios: one selecting the top 10 stocks with the highest total heat from the top 5 concepts with the largest heat change rate, and another selecting the bottom 10 stocks with the lowest total heat from the same concepts[32][33] - The "BOTTOM" portfolio, which selects low-heat stocks from high-heat concepts, has historically achieved an annualized return of 15.71% with a maximum drawdown of 28.89%, and a return of 41% in 2025[34] - The "Total Heat Index" indicator is defined as the sum of browsing, self-selection, and click counts of individual stocks, normalized by their market share on the same day, and then multiplied by 10,000, with a value range of [0,10000][7] - The report calculates the weekly heat change rate (MA2) for different groups of stocks, including broad-based indices, industries, and concepts, and uses this to construct a rotation strategy[13][15] - The rotation strategy involves buying the index with the highest heat change rate at the end of each week, and staying out of the market if the highest change rate belongs to the "others" group[13][15] - The report constructs two simple portfolios: one selecting the top 10 stocks with the highest total heat from the top 5 concepts with the largest heat change rate, and another selecting the bottom 10 stocks with the lowest total heat from the same concepts[32][33] - The rotation strategy based on the heat change rate (MA2) has an annualized return of 8.74% since 2017, with a maximum drawdown of 23.5%, and a return of 31.42% in 2025[16] - The "BOTTOM" portfolio, which selects low-heat stocks from high-heat concepts, has historically achieved an annualized return of 15.71% with a maximum drawdown of 28.89%, and a return of 41% in 2025[34]