戴维斯双击策略本周超额收益 2.39% Email：wuxx02@zts.com.cn 执业证书编号：S0740525060001 Email：wangpf@zts.com.cn 1、《量化择时周报：牛市格局仍在 延 续 ， 主 题 投 资 重 回 主 线 》 2026-01-25 2、《沪深 300 增强策略本周超额收 益 3.90%》2026-01-25 3、《净利润断层策略本周绝对收益 1.99%》2026-01-18 证券研究报告/金融工程定期报告 2026 年 02 月 01 日 执业证书编号：S0740525110003 分析师：吴先兴 报告摘要 相关报告 净利润断层策略 沪深 300 增强组合 请务必阅读正文之后的重要声明部分 戴维斯双击策略 分析师：王鹏飞 戴维斯双击即指以较低的市盈率买入具有成长潜力的股票，待成长性显现、市盈率相 应提高后卖出，获得乘数效应的收益，即 EPS 和 PE 的"双击"。策略在 2010-2017 年回测期内实现了 26.45%的年化收益，超额基准 21.08%。 策略在 2010-2017 回测期内实现了 26.45%的年化收益，超额基准 21.08%，且在回 测 ...

Quantitative Models and Construction Methods 1. **Model Name**: CSI 300 Enhanced Portfolio - **Model Construction Idea**: The model aims to achieve excess returns over the CSI 300 Index by leveraging quantitative strategies and factor-based stock selection - **Model Construction Process**: The model is constructed by selecting stocks from the CSI 300 Index based on specific quantitative factors and optimizing the portfolio to maximize excess returns while managing risk. The exact factors and optimization techniques are not detailed in the report - **Model Evaluation**: The model has shown consistent performance in generating excess returns over the CSI 300 Index in the year-to-date period[5][9][15] 2. **Model Name**: CSI 500 Enhanced Portfolio - **Model Construction Idea**: The model seeks to outperform the CSI 500 Index by utilizing quantitative strategies and factor-based stock selection - **Model Construction Process**: Stocks are selected from the CSI 500 Index based on quantitative factors, and the portfolio is optimized to achieve excess returns while controlling risk. Specific details of the factors and optimization are not provided in the report - **Model Evaluation**: The model's performance has been mixed, with negative excess returns in the year-to-date period[5][9][15] 3. **Model Name**: CSI 1000 Enhanced Portfolio - **Model Construction Idea**: The model aims to generate excess returns over the CSI 1000 Index through quantitative strategies and factor-based stock selection - **Model Construction Process**: Stocks are selected from the CSI 1000 Index using quantitative factors, and the portfolio is optimized to maximize excess returns while managing risk. Specific details of the factors and optimization are not provided in the report - **Model Evaluation**: The model has demonstrated positive excess returns in the year-to-date period[5][9][15] 4. **Model Name**: PB-Profit Combination Portfolio - **Model Construction Idea**: The portfolio combines price-to-book (PB) ratio and profitability factors to identify undervalued stocks with strong earnings potential - **Model Construction Process**: The portfolio is constructed by selecting stocks with low PB ratios and high profitability metrics. The exact methodology for combining these factors is not detailed in the report - **Model Evaluation**: The portfolio has shown strong performance, with significant positive excess returns over the CSI 300 Index in the year-to-date period[5][31][33] 5. **Model Name**: GARP Portfolio - **Model Construction Idea**: The portfolio follows the Growth at a Reasonable Price (GARP) strategy, focusing on stocks with a balance of growth and valuation metrics - **Model Construction Process**: Stocks are selected based on a combination of growth and valuation factors. The specific factors and their weights are not detailed in the report - **Model Evaluation**: The portfolio has achieved significant positive excess returns over the CSI 300 Index in the year-to-date period[5][35] 6. **Model Name**: Small-Cap Value Portfolio 1 - **Model Construction Idea**: The portfolio targets small-cap stocks with value characteristics, aiming to outperform the micro-cap index - **Model Construction Process**: Stocks are selected based on small-cap and value factors. The exact methodology for combining these factors is not detailed in the report - **Model Evaluation**: The portfolio has underperformed the micro-cap index in the year-to-date period[5][37] 7. **Model Name**: Small-Cap Value Portfolio 2 - **Model Construction Idea**: Similar to Small-Cap Value Portfolio 1, this portfolio focuses on small-cap stocks with value characteristics - **Model Construction Process**: Stocks are selected based on small-cap and value factors. The exact methodology for combining these factors is not detailed in the report - **Model Evaluation**: The portfolio has outperformed the micro-cap index in the year-to-date period[5][39] 8. **Model Name**: Small-Cap Growth Portfolio - **Model Construction Idea**: The portfolio targets small-cap stocks with growth characteristics, aiming to outperform the micro-cap index - **Model Construction Process**: Stocks are selected based on small-cap and growth factors. The exact methodology for combining these factors is not detailed in the report - **Model Evaluation**: The portfolio has underperformed the micro-cap index in the year-to-date period[5][41] --- Model Backtesting Results 1. **CSI 300 Enhanced Portfolio** - Weekly return: -0.39% - Weekly excess return: -0.47% - Year-to-date return: 6.85% - Year-to-date excess return: 5.20%[9][15] 2. **CSI 500 Enhanced Portfolio** - Weekly return: -1.74% - Weekly excess return: 0.82% - Year-to-date return: 11.11% - Year-to-date excess return: -1.01%[9][15] 3. **CSI 1000 Enhanced Portfolio** - Weekly return: -0.97% - Weekly excess return: 1.58% - Year-to-date return: 11.99% - Year-to-date excess return: 3.31%[9][15] 4. **PB-Profit Combination Portfolio** - Weekly return: 0.92% - Weekly excess return: 0.84% - Year-to-date return: 6.17% - Year-to-date excess return: 4.52%[31][33] 5. **GARP Portfolio** - Weekly return: 0.95% - Weekly excess return: 0.87% - Year-to-date return: 11.43% - Year-to-date excess return: 9.78%[35] 6. **Small-Cap Value Portfolio 1** - Weekly return: -2.44% - Weekly excess return: -1.29% - Year-to-date return: 7.89% - Year-to-date excess return: -2.83%[37] 7. **Small-Cap Value Portfolio 2** - Weekly return: -1.64% - Weekly excess return: -0.48% - Year-to-date return: 12.37% - Year-to-date excess return: 1.66%[39] 8. **Small-Cap Growth Portfolio** - Weekly return: -2.07% - Weekly excess return: -0.92% - Year-to-date return: 9.13% - Year-to-date excess return: -1.59%[41] --- Quantitative Factors and Construction Methods 1. **Factor Name**: Market Capitalization (Size) Factor - **Construction Idea**: Small-cap stocks tend to outperform large-cap stocks over time - **Construction Process**: Stocks are ranked by market capitalization, and the top 10% (smallest) and bottom 10% (largest) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the factor's performance - **Evaluation**: The factor has shown mixed performance across different indices and time periods[43][44][45] 2. **Factor Name**: Price-to-Book (PB) Factor - **Construction Idea**: Low PB stocks are expected to outperform high PB stocks - **Construction Process**: Stocks are ranked by PB ratio, and the top 10% (lowest PB) and bottom 10% (highest PB) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the factor's performance - **Evaluation**: The factor has shown strong performance in the short term but mixed results in the year-to-date period[43][44][45] 3. **Factor Name**: Price-to-Earnings (PE_TTM) Factor - **Construction Idea**: Low PE stocks are expected to outperform high PE stocks - **Construction Process**: Stocks are ranked by PE ratio, and the top 10% (lowest PE) and bottom 10% (highest PE) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the factor's performance - **Evaluation**: The factor has shown positive short-term performance but mixed year-to-date results[43][44][45] 4. **Factor Name**: Reversal Factor - **Construction Idea**: Stocks with recent underperformance are expected to outperform in the short term - **Construction Process**: Stocks are ranked by recent performance, and the top 10% (worst performers) and bottom 10% (best performers) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the factor's performance - **Evaluation**: The factor has shown positive short-term performance but negative year-to-date results[49][50] 5. **Factor Name**: Turnover Factor - **Construction Idea**: Stocks with lower turnover rates are expected to outperform those with higher turnover rates - **Construction Process**: Stocks are ranked by turnover rate, and the top 10% (lowest turnover) and bottom 10% (highest turnover) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the factor's performance - **Evaluation**: The factor has shown strong short-term performance but negative year-to-date results[49][50] 6. **Factor Name**: Volatility Factor - **Construction Idea**

Summary of Key Points Core Viewpoints - The report provides an analysis of dividend progress and index futures basis for major indices in January 2026, highlighting the current status and historical trends in dividends and basis rates [2][6][21]. Group 1: Dividend Statistics of Major Indices - In the year 2026, the dividend progress for major indices is as follows: - For the CSI 300, one company has passed the board proposal stage - For the SSE 50, one company has passed the board proposal stage - For the CSI 500, one company is in the implementation stage - For the CSI 1000, one company is in the proposal stage [11][13]. - The cumulative dividends for the CSI 500 index is 1.30 billion, while the other indices have not reported any cumulative dividends yet [13][15]. Group 2: Industry Dividend Statistics - The dividend progress for various industry indices in 2026 includes: - In the pharmaceutical and biological sector, one company is in the implementation stage and one has passed the board proposal stage - In the public utilities sector, one company has passed the board proposal stage - In the machinery equipment sector, one company is in the proposal stage and one is in the board proposal stage - In the coal sector, one company is in the implementation stage - In the oil and petrochemical sector, one company is in the implementation stage [12][15]. - The cumulative dividends reported for the pharmaceutical sector is 0.66 billion and for the coal sector is 1.30 billion [15]. Group 3: Index Futures Basis - The annualized basis rates considering dividends for various contracts are as follows: - For the CSI 300: near-month -0.25%, far-month 0.04%, near-quarter 0.44%, far-quarter 0.55% - For the SSE 50: near-month -0.89%, far-month -1.30%, near-quarter -1.67%, far-quarter -1.57% - For the CSI 500: near-month -4.14%, far-month -2.29%, near-quarter -0.53%, far-quarter 1.11% - For the CSI 1000: near-month -1.56%, far-month 1.25%, near-quarter 4.20%, far-quarter 5.11% [6][25]. - The report includes a detailed table of basis data for each contract, showing the latest closing prices, basis, dividend impact, and annualized basis rates [25][27].

Quantitative Models and Construction Methods 1. Model Name: Multi-Cycle Timing Model for Interest Rates - **Model Construction Idea**: The model uses kernel regression to identify support and resistance lines in interest rate trends. It evaluates the breakthrough patterns of interest rate movements across different investment cycles (long, medium, and short) to generate composite timing signals[11][24]. - **Model Construction Process**: - **Data Input**: Yield-to-Maturity (YTM) data for 5-year, 10-year, and 30-year government bonds[11][24]. - **Kernel Regression**: Applied to capture the support and resistance lines of interest rate trends[11]. - **Cycle Classification**: - Long cycle: Monthly frequency - Medium cycle: Bi-weekly frequency - Short cycle: Weekly frequency[11][24]. - **Signal Generation**: - If at least two cycles show downward breakthroughs of support lines, the signal is "bullish" (e.g., 5-year and 10-year YTM signals are bullish) - If at least two cycles show upward breakthroughs of resistance lines, the signal is "bearish" (e.g., 30-year YTM signal is bearish)[11][24]. - **Model Evaluation**: The model effectively captures multi-cycle resonance in interest rate trends, providing actionable timing signals for different bond maturities[11][24]. 2. Model Name: Multi-Cycle Trading Strategy - **Model Construction Idea**: The strategy is based on the multi-cycle timing signals generated by the above model. It allocates bond portfolios dynamically based on the direction of interest rate trends and cycle breakthroughs[24][29]. - **Model Construction Process**: - **Portfolio Allocation Rules**: - If at least two cycles show downward breakthroughs and the trend is not upward, allocate fully to long-duration bonds. - If at least two cycles show downward breakthroughs but the trend is upward, allocate 50% to medium-duration bonds and 50% to long-duration bonds. - If at least two cycles show upward breakthroughs and the trend is not downward, allocate fully to short-duration bonds. - If at least two cycles show upward breakthroughs but the trend is downward, allocate 50% to medium-duration bonds and 50% to short-duration bonds. - In other cases, allocate equally across short, medium, and long durations[24][29]. - **Stop-Loss Mechanism**: If the daily excess return of the portfolio falls below -0.5%, adjust holdings to equal-weighted allocation[29]. - **Performance Benchmark**: Equal-weighted allocation across short, medium, and long durations serves as the benchmark[24][29]. - **Model Evaluation**: The strategy demonstrates robust performance across different market conditions, with high win rates for both absolute and excess returns over the past 18 years[24][29]. --- Model Backtesting Results 1. Multi-Cycle Timing Model for Interest Rates - **5-Year YTM**: - Long-term annualized return: 5.46% - Maximum drawdown: 2.88% - Return-to-drawdown ratio: 1.9 - Short-term annualized return (since 2024): 2.09% - Maximum drawdown: 0.59% - Return-to-drawdown ratio: 3.55 - Long-term excess return: 1.06% - Short-term excess return: 0.64%[25][28]. - **10-Year YTM**: - Long-term annualized return: 6.03% - Maximum drawdown: 2.74% - Return-to-drawdown ratio: 2.2 - Short-term annualized return (since 2024): 2.34% - Maximum drawdown: 0.58% - Return-to-drawdown ratio: 4.05 - Long-term excess return: 1.63% - Short-term excess return: 1.06%[28][33]. - **30-Year YTM**: - Long-term annualized return: 7.28% - Maximum drawdown: 4.27% - Return-to-drawdown ratio: 1.7 - Short-term annualized return (since 2024): 2.47% - Maximum drawdown: 0.92% - Return-to-drawdown ratio: 2.7 - Long-term excess return: 2.39% - Short-term excess return: 2.16%[33][37]. 2. Multi-Cycle Trading Strategy - **5-Year YTM**: - Annualized return (2008-2025): 2.10%-14.83% - Excess return (2008-2025): 0.29%-2.77%[37]. - **10-Year YTM**: - Annualized return (2008-2025): 0.11%-17.08% - Excess return (2008-2025): -0.08%-4.41%[37]. - **30-Year YTM**: - Annualized return (2008-2025): -0.36%-19.93% - Excess return (2008-2025): -0.39%-5.48%[37].

Group 1: Outdoor Apparel Industry - The outdoor footwear and apparel industry has maintained rapid growth since 2021, with a CAGR of 25.3% for outdoor apparel and 18.4% for outdoor footwear, projected to grow by 24.5% and 16.3% respectively in 2025 [24][26] - Online sales of outdoor footwear are growing at over 40%, while growth in outdoor apparel has slowed to low single digits since Q2 2025; specific categories like down jackets and quick-dry clothing are experiencing strong growth, with some quarterly YoY growth nearing 100% [24][26] - The industry is seeing increased competition among brands, with top brands like Kailas and Berghaus maintaining high growth through specialized product lines, while others like The North Face are underperforming; pricing trends are weakening overall, but some high-demand brands are still able to increase prices [24][26] Group 2: AI Application in Computing Industry - Major international companies are focusing on AI application in vertical scenarios, with OpenAI launching ChatGPT Health and Amazon optimizing cross-border e-commerce operations through AI [28] - Domestic companies are also advancing in AI applications, with Alibaba upgrading health services and ByteDance's Volcano Engine becoming a key AI cloud partner for major events [28] - The market for AI applications is expected to see significant growth, with predictions indicating that the GEO market will reach $24 billion globally by 2026, driven by high consumer trust in AI applications in China [30][32] Group 3: Public Utilities Industry - The public utilities sector, including electricity, gas, and water, is characterized by its "essential" nature, with stable long-term growth prospects; the transition to low-carbon energy sources is expected to increase the share of clean energy consumption to 28.6% by 2024 [32][33] - The global electricity shortage is becoming more pronounced, leading to increased electricity prices and making the sector an attractive investment area, particularly as AI development accelerates [33]

- The report discusses high-frequency stock selection factors, deep learning factors, and AI-enhanced portfolios, summarizing their historical and 2026 performance in terms of IC, RankMAE, long-short returns, long-only excess returns, and monthly win rates[9][10][11] - High-frequency factors include intraday skewness, downside volatility proportion, post-opening buying intention proportion, post-opening buying intensity, net large-order buying proportion, net large-order buying intensity, improved reversal, end-of-day trading proportion, average single-order outflow proportion, and large-order-driven price increase[7][9][10] - Deep learning factors include GRU(50,2)+NN(10), residual attention LSTM(48,2)+NN(10), multi-granularity models with 5-day and 10-day labels, which are trained using advanced machine learning techniques like AGRU[7][9][10] - AI-enhanced portfolios are constructed based on deep learning factors, specifically the multi-granularity 10-day label model, and include four combinations: CSI 500 AI-enhanced wide constraint, CSI 500 AI-enhanced strict constraint, CSI 1000 AI-enhanced wide constraint, and CSI 1000 AI-enhanced strict constraint. These portfolios aim to maximize expected returns under specific constraints such as turnover, industry, and market cap limits[73][74][75] - The optimization objective for AI-enhanced portfolios is defined as maximizing the expected excess return, represented by the formula: $$\operatorname*{max}_{w_{i}}\sum\mu_{i}w_{i}$$ where \(w_i\) is the weight of stock \(i\) in the portfolio, and \(\mu_i\) is the expected excess return of stock \(i\)[74][75] - Performance results for high-frequency factors show positive long-short returns for most factors in January and 2026, with notable results for factors like intraday skewness (1.55%), downside volatility proportion (1.65%), and post-opening buying intensity (2.86%)[5][9][10] - Deep learning factors also demonstrate strong performance, with GRU(50,2)+NN(10) achieving a long-short return of 2.79% in 2026, and the multi-granularity 5-day label model achieving 2.13%[5][9][10] - AI-enhanced portfolios show mixed results, with the CSI 500 AI-enhanced wide constraint portfolio recording a -4.47% excess return in 2026, while the CSI 1000 AI-enhanced strict constraint portfolio achieved a relatively better performance of -1.57%[5][14][73]

Quantitative Models and Construction Quantitative Models and Construction Process 1. **Model Name**: Linear Multi-Factor Model for Index Enhancement - **Construction Idea**: The model is based on a linear multi-factor framework, incorporating style, price-volume, and fundamental factors to construct monthly rebalanced index enhancement portfolios for major indices like CSI 300, CSI 500, CSI 1000, and CSI A500[41][42] - **Construction Process**: - Factors used include size, mid-cap, reversal, volatility, turnover, PB, ROE, SUE, R&D ratio, adjusted net profit expectations, analyst coverage, and others[41] - Risk control constraints include limits on size, valuation, individual stock, and industry deviations[42] - Two weighting methods are tested: IC mean weighting and ICIR weighting. ICIR weighting considers factor volatility, aiming for more stable performance[42][57] - **Evaluation**: ICIR weighting outperforms IC mean weighting, especially in recent years when factor returns have declined, and volatility has increased[57][44] 2. **Model Name**: Composite Strategy for CSI 300 Index Enhancement - **Construction Idea**: Combines multiple strategies to improve performance by allocating weights to different sub-strategies[60][53] - **Construction Process**: - The composite strategy consists of three components: 1. **Base Index Enhancement Strategy** (60% weight) 2. **In-Scope Satellite Strategy** (30% weight), focusing on momentum and fundamental factors 3. **Out-of-Scope Satellite Strategy** (10% weight), targeting small-cap, high-growth stocks[60][53] - Monthly rebalancing is applied to the portfolio[60] - **Evaluation**: The composite strategy improves annualized returns by 3.6% compared to the base strategy, with higher stability across years. However, relative drawdowns may increase in certain years[55][60] --- Model Backtesting Results Linear Multi-Factor Model for Index Enhancement 1. **CSI 300 Index**: - IC Mean Weighting: Annualized excess return 10.0%, tracking error 5.1%, IR 1.85[45] - ICIR Weighting: Annualized excess return 11.1%, tracking error 5.2%, IR 2.01[45] - 2025 Results: IC Mean Weighting excess return 6.8%, ICIR Weighting excess return 10.7%[57][45] 2. **CSI 500 Index**: - IC Mean Weighting: Annualized excess return 11.0%, tracking error 5.1%, IR 2.08[46] - ICIR Weighting: Annualized excess return 12.3%, tracking error 4.7%, IR 2.53[46] - 2025 Results: IC Mean Weighting excess return 3.1%, ICIR Weighting excess return 9.5%[57][46] 3. **CSI 1000 Index**: - IC Mean Weighting: Annualized excess return 14.8%, tracking error 5.4%, IR 2.67[47] - ICIR Weighting: Annualized excess return 17.4%, tracking error 5.0%, IR 3.39[47] - 2025 Results: IC Mean Weighting excess return 5.1%, ICIR Weighting excess return 10.2%[57][47] 4. **CSI A500 Index**: - IC Mean Weighting: Annualized excess return 7.7%, tracking error 4.5%, IR 1.67[49] - ICIR Weighting: Annualized excess return 10.3%, tracking error 4.5%, IR 2.21[49] - 2025 Results: IC Mean Weighting excess return 4.8%, ICIR Weighting excess return 13.2%[57][49] Composite Strategy for CSI 300 Index Enhancement 1. Annualized excess return: 12.2%, compared to 8.6% for the base strategy[55] 2. Information ratio: Improved from 1.56 (base strategy) to 1.93 (composite strategy)[55] 3. 2025 Results: The composite strategy mitigated drawdowns during periods of small-cap and low-valuation factor underperformance, outperforming the base strategy[56][60] --- Quantitative Factors and Construction Quantitative Factors and Construction Process 1. **Factor Name**: Small-Cap Factor - **Construction Idea**: Captures the performance of small-cap stocks relative to the market[50] - **Construction Process**: - Exposure to small-cap stocks is measured and incorporated into the portfolio construction process - The factor contributed 3.7% to the excess return of the CSI 300 enhancement strategy in 2025[50][57] 2. **Factor Name**: SUE (Standardized Unexpected Earnings) - **Construction Idea**: Measures earnings surprises to identify stocks with positive earnings momentum[50] - **Construction Process**: - SUE is calculated and used as a factor in the multi-factor model - Higher exposure to SUE contributed positively to the ICIR-weighted portfolio in 2025[50][57] 3. **Factor Name**: R&D Ratio - **Construction Idea**: Reflects the intensity of research and development investment as a proxy for innovation[50] - **Construction Process**: - R&D ratio is calculated and included in the factor set - The factor contributed positively to the ICIR-weighted portfolio in 2025[50][57] --- Factor Backtesting Results 1. **Small-Cap Factor**: Contributed 3.7% to the excess return of the CSI 300 enhancement strategy in 2025[50][57] 2. **SUE Factor**: Contributed 2.75% to the excess return of the ICIR-weighted portfolio in 2025[50][57] 3. **R&D Ratio Factor**: Contributed 0.88% to the excess return of the ICIR-weighted portfolio in 2025[50][57]

- The report analyzes the performance of industries with continuous upward or downward trends over multiple years, using the Changjiang Level-1 Industry Index as the statistical object, covering data from 2005 to 2025[8][15] - Industries that experienced continuous growth for 2-4 years showed negative average returns in the following year, ranging from -1.02% to -6.30%, with a probability of positive returns between 28% and 40%, indicating difficulty in maintaining positive returns after prolonged growth[12][15][20] - Industries that experienced continuous declines for 2-4 years showed positive average returns in the following year, ranging from 7.02% to 18.03%, with a probability of further declines between 26% and 44%, suggesting a recovery opportunity after significant declines[12][15][20] - Industries with continuous positive excess returns relative to the Wind All-A Index for 2-4 years had a low probability (11%-35%) of achieving positive excess returns in the following year, with negative average excess returns, indicating that strong relative performance is difficult to sustain[12][17][21] - Industries with continuous negative excess returns relative to the Wind All-A Index for 2-4 years had a high probability (61%-65%) of continuing negative excess returns in the following year, with negative average excess returns, suggesting that weak relative performance is hard to reverse[12][17][21] - The report highlights a mean-reversion characteristic in absolute returns, where industries tend to weaken after prolonged growth and recover after prolonged declines, aligning with the market principle of "what goes up must come down, and what goes down must come up"[12][16][22]

- The "Buy Intention Factor" showed strong performance at the beginning of the year, with intraday high-frequency skewness factor, intraday downside volatility proportion factor, post-opening buy intention proportion factor, post-opening buy intention strength factor, post-opening large order net buy proportion factor, post-opening large order net buy strength factor, intraday return factor, end-of-day trading proportion factor, average single outflow amount proportion factor, and large order push-up factor all being evaluated[5][6][9] - The "Multi-Granularity Factor" showed average performance, with GRU(10,2)+NN(10) factor, GRU(50,2)+NN(10) factor, multi-granularity model (5-day label) factor, and multi-granularity model (10-day label) factor being evaluated[5][6][9] - The "AI Enhanced Portfolio" had a poor start to the year, with significant drawdowns observed in the weekly rebalanced CSI 500 AI enhanced wide constraint portfolio, CSI 500 AI enhanced strict constraint portfolio, CSI 1000 AI enhanced wide constraint portfolio, and CSI 1000 AI enhanced strict constraint portfolio[5][6][9] Quantitative Factors and Construction Methods 1. **Factor Name: Intraday High-Frequency Skewness Factor** - **Construction Idea**: Measures the skewness of intraday returns to capture the asymmetry in return distribution[5][6] - **Construction Process**: Calculated using high-frequency data to determine the skewness of returns within a trading day[5][6] - **Evaluation**: Demonstrated strong performance at the beginning of the year[5][6] 2. **Factor Name: Intraday Downside Volatility Proportion Factor** - **Construction Idea**: Measures the proportion of downside volatility in intraday returns[5][6] - **Construction Process**: Calculated using high-frequency data to determine the proportion of downside volatility within a trading day[5][6] - **Evaluation**: Showed moderate performance[5][6] 3. **Factor Name: Post-Opening Buy Intention Proportion Factor** - **Construction Idea**: Measures the proportion of buy intentions after market opening[5][6] - **Construction Process**: Calculated using high-frequency data to determine the proportion of buy intentions after the market opens[5][6] - **Evaluation**: Demonstrated strong performance at the beginning of the year[5][6] 4. **Factor Name: Post-Opening Buy Intention Strength Factor** - **Construction Idea**: Measures the strength of buy intentions after market opening[5][6] - **Construction Process**: Calculated using high-frequency data to determine the strength of buy intentions after the market opens[5][6] - **Evaluation**: Showed moderate performance[5][6] 5. **Factor Name: Post-Opening Large Order Net Buy Proportion Factor** - **Construction Idea**: Measures the proportion of net buy orders of large size after market opening[5][6] - **Construction Process**: Calculated using high-frequency data to determine the proportion of net buy orders of large size after the market opens[5][6] - **Evaluation**: Demonstrated weak performance[5][6] 6. **Factor Name: Post-Opening Large Order Net Buy Strength Factor** - **Construction Idea**: Measures the strength of net buy orders of large size after market opening[5][6] - **Construction Process**: Calculated using high-frequency data to determine the strength of net buy orders of large size after the market opens[5][6] - **Evaluation**: Showed weak performance[5][6] 7. **Factor Name: Intraday Return Factor** - **Construction Idea**: Measures the return within a trading day[5][6] - **Construction Process**: Calculated using high-frequency data to determine the return within a trading day[5][6] - **Evaluation**: Demonstrated strong performance at the beginning of the year[5][6] 8. **Factor Name: End-of-Day Trading Proportion Factor** - **Construction Idea**: Measures the proportion of trading activity at the end of the day[5][6] - **Construction Process**: Calculated using high-frequency data to determine the proportion of trading activity at the end of the day[5][6] - **Evaluation**: Showed strong performance[5][6] 9. **Factor Name: Average Single Outflow Amount Proportion Factor** - **Construction Idea**: Measures the proportion of average single outflow amounts[5][6] - **Construction Process**: Calculated using high-frequency data to determine the proportion of average single outflow amounts[5][6] - **Evaluation**: Demonstrated moderate performance[5][6] 10. **Factor Name: Large Order Push-Up Factor** - **Construction Idea**: Measures the impact of large orders on price increases[5][6] - **Construction Process**: Calculated using high-frequency data to determine the impact of large orders on price increases[5][6] - **Evaluation**: Showed moderate performance[5][6] 11. **Factor Name: GRU(10,2)+NN(10) Factor** - **Construction Idea**: Combines GRU and neural network models to capture complex patterns in data[5][6] - **Construction Process**: Utilizes GRU with 10 units and 2 layers, followed by a neural network with 10 units[5][6] - **Evaluation**: Demonstrated average performance[5][6] 12. **Factor Name: GRU(50,2)+NN(10) Factor** - **Construction Idea**: Combines GRU and neural network models to capture complex patterns in data[5][6] - **Construction Process**: Utilizes GRU with 50 units and 2 layers, followed by a neural network with 10 units[5][6] - **Evaluation**: Showed weak performance[5][6] 13. **Factor Name: Multi-Granularity Model (5-Day Label) Factor** - **Construction Idea**: Uses multi-granularity approach to capture patterns over different time frames[5][6] - **Construction Process**: Trained using a 5-day label to capture short-term patterns[5][6] - **Evaluation**: Demonstrated average performance[5][6] 14. **Factor Name: Multi-Granularity Model (10-Day Label) Factor** - **Construction Idea**: Uses multi-granularity approach to capture patterns over different time frames[5][6] - **Construction Process**: Trained using a 10-day label to capture longer-term patterns[5][6] - **Evaluation**: Showed weak performance[5][6] Factor Backtest Results 1. **Intraday High-Frequency Skewness Factor**: IC -0.007, e^(-rank mae) 0.312, long-short return 0.29%, long-only excess return 0.99%, monthly win rate 1/1[9][10] 2. **Intraday Downside Volatility Proportion Factor**: IC -0.001, e^(-rank mae) 0.313, long-short return 0.22%, long-only excess return 0.95%, monthly win rate 1/1[9][10] 3. **Post-Opening Buy Intention Proportion Factor**: IC 0.032, e^(-rank mae) 0.324, long-short return 1.04%, long-only excess return -0.41%, monthly win rate 0/1[9][10] 4. **Post-Opening Buy Intention Strength Factor**: IC 0.027, e^(-rank mae) 0.323, long-short return 0.65%, long-only excess return 0.62%, monthly win rate 1/1[9][10] 5. **Post-Opening Large Order Net Buy Proportion Factor**: IC -0.006, e^(-rank mae) 0.306, long-short return -0.52%, long-only excess return -0.53%, monthly win rate 0/1[9][10] 6. **Post-Opening Large Order Net Buy Strength Factor**: IC 0.004, e^(-rank mae) 0.308, long-short return -0.07%, long-only excess return -0.66%, monthly win rate 0/1[9][10] 7. **Intraday Return Factor**: IC 0.037, e^(-rank mae) 0.328, long-short return 1.77%, long-only excess return 1.89%, monthly win rate 1/1[9][10] 8. **End-of-Day Trading Proportion Factor**: IC 0.084, e^(-rank mae) 0.334, long-short return 2.67%, long-only excess return 1.35%, monthly win rate 1/1[9][10] 9. **Average Single Outflow Amount Proportion Factor**: IC 0.013, e^(-rank mae) 0.319, long-short return 0.45%, long-only excess return 0.14%, monthly win rate 1/1[9][10] 10. **Large Order Push-Up Factor**: IC -0.007, e^(-rank mae) 0.327, long-short return 0.22%, long-only excess return 0.43%, monthly win rate 1/1[9][10] 11. **GRU(10,2

Quantitative Models and Construction Methods - **Model Name**: CSI 300 Enhanced Portfolio **Model Construction Idea**: The model aims to enhance the performance of the CSI 300 Index by leveraging quantitative strategies to generate excess returns over the benchmark index[5][9][15] **Model Construction Process**: The portfolio is constructed by applying quantitative stock selection and weighting methodologies to the CSI 300 Index constituents. The process involves identifying stocks with favorable factor exposures and optimizing the portfolio to maximize risk-adjusted returns while maintaining a low tracking error relative to the benchmark[9][15] **Model Evaluation**: The model demonstrated strong performance with positive excess returns over the benchmark index, indicating effective factor utilization and portfolio construction[15] - **Model Name**: CSI 500 Enhanced Portfolio **Model Construction Idea**: Similar to the CSI 300 Enhanced Portfolio, this model focuses on enhancing the performance of the CSI 500 Index by employing quantitative strategies[5][9][15] **Model Construction Process**: The portfolio is built by selecting stocks from the CSI 500 Index based on quantitative factors and optimizing the portfolio to achieve excess returns while controlling tracking error[9][15] **Model Evaluation**: The model underperformed the benchmark index during the observed period, suggesting potential challenges in factor effectiveness or market conditions[15] - **Model Name**: CSI 1000 Enhanced Portfolio **Model Construction Idea**: This model targets the CSI 1000 Index, aiming to generate excess returns through quantitative enhancements[5][9][15] **Model Construction Process**: The portfolio construction involves selecting stocks from the CSI 1000 Index using quantitative factors and optimizing the portfolio for risk-adjusted returns and low tracking error[9][15] **Model Evaluation**: The model showed a slight underperformance relative to the benchmark index, indicating room for improvement in factor application or portfolio optimization[15] - **Model Name**: GARP Portfolio **Model Construction Idea**: The GARP (Growth at a Reasonable Price) portfolio combines growth and valuation factors to identify stocks with strong growth potential at reasonable valuations[32] **Model Construction Process**: Stocks are selected based on a combination of growth metrics (e.g., earnings growth) and valuation metrics (e.g., price-to-earnings ratio). The portfolio is then optimized to balance growth and valuation exposures[32] **Model Evaluation**: The portfolio achieved positive excess returns over the CSI 300 Index, demonstrating the effectiveness of the GARP strategy in the observed period[32] - **Model Name**: Small-Cap Growth Portfolio **Model Construction Idea**: This portfolio focuses on small-cap stocks with strong growth characteristics, aiming to capture the growth premium in the small-cap segment[37] **Model Construction Process**: Stocks are selected from the small-cap universe based on growth factors such as earnings growth and revenue growth. The portfolio is optimized to maximize growth exposure while managing risk[37] **Model Evaluation**: The portfolio delivered positive excess returns over the micro-cap index, indicating the effectiveness of the growth factor in the small-cap segment[37] Model Backtesting Results - **CSI 300 Enhanced Portfolio**: Weekly return 4.69%, excess return 1.90%, tracking error 4.71%, maximum drawdown 1.68%[9][15][22] - **CSI 500 Enhanced Portfolio**: Weekly return 6.34%, excess return -1.58%, tracking error 4.07%, maximum drawdown 3.11%[9][15][16] - **CSI 1000 Enhanced Portfolio**: Weekly return 6.17%, excess return -0.86%, tracking error 5.31%, maximum drawdown 4.45%[9][15][18] - **GARP Portfolio**: Weekly return 3.62%, excess return 0.84%, tracking error 13.93%, maximum drawdown 4.04%[32][33] - **Small-Cap Growth Portfolio**: Weekly return 4.95%, excess return 0.49%, tracking error 11.60%, maximum drawdown 9.76%[37][40] Quantitative Factors and Construction Methods - **Factor Name**: Market Capitalization (Size) Factor **Factor Construction Idea**: This factor captures the size effect, where smaller companies tend to outperform larger companies over time[42] **Factor Construction Process**: Stocks are ranked by their market capitalization, and the top 10% (large-cap) and bottom 10% (small-cap) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the size factor's performance[41][42] **Factor Evaluation**: The factor showed mixed performance, with large-cap stocks outperforming small-cap stocks in the observed period[42] - **Factor Name**: Price-to-Book Ratio (PB) Factor **Factor Construction Idea**: This factor identifies undervalued stocks based on their price-to-book ratios[42] **Factor Construction Process**: Stocks are ranked by their PB ratios, and the top 10% (high PB) and bottom 10% (low PB) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the PB factor's performance[41][42] **Factor Evaluation**: The factor underperformed during the observed period, with high PB stocks outperforming low PB stocks[42] - **Factor Name**: Expected Net Profit Adjustment Factor **Factor Construction Idea**: This factor captures the impact of expected net profit adjustments on stock performance[53] **Factor Construction Process**: Stocks are ranked by their expected net profit adjustments, and the top 10% (high adjustment) and bottom 10% (low adjustment) are selected to form long and short portfolios, respectively. The difference in returns between these portfolios represents the factor's performance[41][53] **Factor Evaluation**: The factor delivered positive returns, indicating its effectiveness in identifying stocks with favorable profit adjustments[53] Factor Backtesting Results - **Market Capitalization (Size) Factor**: Multi-market excess returns: -0.79% (All Market), 4.83% (CSI 300), -5.59% (CSI 500), -2.47% (CSI 1000)[42][43][48] - **Price-to-Book Ratio (PB) Factor**: Multi-market excess returns: -4.01% (All Market), -5.52% (CSI 300), -6.06% (CSI 500), -5.68% (CSI 1000)[42][43][48] - **Expected Net Profit Adjustment Factor**: Multi-market excess returns: 0.57% (All Market), 0.86% (CSI 300), 1.89% (CSI 500), -0.58% (CSI 1000)[53][54][55]