2025年，无疑是量化投资发展史上一个标志性的大年。年初，DeepSeek的横空出世为量化领域注入了颠覆性的AI动能；与此同时，A股市场震荡上行，整 体而言中小盘风格显著占优，中证2000与微盘股指数分别大涨超36%和80%，叠加市场流动性充裕、交投活跃，多重利好共振，令量化多头策略"如鱼得 水"，迎来全面爆发！ 为能更清晰了解量化私募基金的业绩，经笔者统计，在私募排排网上有业绩展示的量化产品共有1784只，2025年平均收益达30.28%，2025年平均超额 （几何）收益达10.83%；其中量化多头策略产品多达806只，2025年收益和超额（几何）收益分别为44.74%、16.46%，在私募二级策略中居前。 | 排 | 私募二级策略 | 有业绩展示的 | 2025年平均收益 | 2025 E 24 2 | | --- | --- | --- | --- | --- | | 序 | | 量化产品 | | 超额收益 | | 1 | 喜化多杀 | 806 | 44.74% | 16.46% | | 2 | 量化CTA | 375 | 20.21% | 14.04% | | 3 | 股票市场中性 | 186 | 9 ...

Core Viewpoint - The A-shares market has experienced a significant bull market over the past year, with a shift in investment focus from "chasing returns" to "selecting long-term core assets" as the market transitions from broad-based gains to differentiation [2] Group 1: Scale and Fund Behavior - Since the launch of the first CSI A500 ETF in September 2024, the total scale of A500-related ETFs and linked products has exceeded 160 billion yuan, marking it as one of the fastest-growing broad-based indices in the past decade [3] - In December 2025, the A500 ETF saw nearly 100 billion yuan in net subscriptions, primarily concentrated in the most liquid top products, indicating a strong institutional investment characteristic [3][5] - The rapid expansion of A500 ETF scale reflects institutional recognition of the index's investment value, as it is not driven by short-term sentiment or retail trading [5] Group 2: Quantitative and Private Equity Perspective - The increase in the number of A500 index-enhanced private equity products in 2025 directly reflects professional investment institutions' judgment on the index's "tradeability" [7] - The A500 index's structure is more favorable for quantitative models, with a broader coverage and a balanced industry allocation, making it a suitable base for beta strategies [10] - A500's characteristics allow for stable excess returns without significantly increasing systemic risk, positioning it as a new core index for quantitative institutions [11] Group 3: Importance of A500's Allocation Value - A500 has become a key "core equity vehicle" in the current market, offering threefold allocation value: as a central equity asset, as a quality base for alpha generation, and as a point of convergence for policy and capital preferences [12] - The rapid development of A500-related products indicates a clear market demand for balanced broad-based indices, shifting the focus from "whether to pay attention to A500" to "how much weight to allocate and how to enhance return efficiency" [12]

- The report highlights the performance of eight major stock selection factors across different stock pools (All A-shares, CSI 300, CSI 500, and CSI 1000). Among these, the value factor (17.88%) and size factor (11.88%) showed strong IC mean performance, while reversal and quality factors performed relatively weaker[54][55][66] - Small-cap value style dominated the market, with the small-cap size factor performing strongly across the entire market. Value factors also showed positive performance, indicating a market preference for low valuation stocks. Additionally, technical and low-volatility factors performed well, while consensus expectation factors weakened due to reduced focus on high-performance expectation sectors[54][55][66] - The report provides detailed definitions for various factors, such as size (logarithm of market capitalization), value (e.g., book-to-price ratio, earnings-to-price ratio), growth (e.g., net income growth), quality (e.g., ROE, gross margin), consensus expectation (e.g., changes in expected EPS), technical (e.g., volume skewness), volatility (e.g., 60-day return standard deviation), and reversal (e.g., 20-day return)[66][67] - The report also tracks the performance of convertible bond selection factors, which are constructed based on the relationship between convertible bonds and their underlying stocks. Factors include stock consensus expectation, stock value, and convertible bond valuation (e.g., parity premium rate). Among these, stock consensus expectation and stock value factors achieved higher IC mean values in the past week[59][60][62]

- The report highlights the performance of quantitative funds, particularly index-enhanced funds, with the China Securities 1000 Index Enhanced Fund achieving a weekly median excess return of 0.72%[3][4][7] - The China Securities 500 Index Enhanced Fund showed a weekly median excess return of -0.13%, while the CSI 300 Index Enhanced Fund achieved 0.42% during the same period[3][4][5] - Other index-enhanced funds recorded a weekly median return of 0.14%, with the best-performing fund achieving a return of 7.42% and the worst-performing fund at -3.19%[3][4][8] - Absolute return (hedging) funds had a weekly median return of 0.12%, while other active quantitative funds achieved a significantly higher weekly median return of 2.35%[3][8][10] - Multi-factor funds demonstrated strong performance with a weekly median return of 3.78%, and the best-performing fund in this category achieved a return of 5.32%[15][19][20] - Big data-driven active investment funds achieved a weekly median return of 1.50%, with the best-performing fund reaching 6.00% and the worst-performing fund at -1.59%[15][20]

Group 1 - The core focus of the article is on Yang Dong, a seasoned fund manager at GF Fund, who currently manages a total of 360.45 billion yuan in assets and is recognized for his "steady" investment style that combines subjective long positions with quantitative strategies [1][2][3] - Yang Dong's investment strategy has led to significant performance, with his flagship fund, GF Multi-Factor, achieving a return of 89.03% in 2021, and his newly launched GF Value Navigator fund yielding a cumulative return of 126.75% since inception [2][3] - The article highlights the challenges Yang faces in adapting his investment style to changing market conditions, particularly as he has shifted from traditional value stocks to high-dividend assets in response to market trends [4][5] Group 2 - Yang Dong's product lineup includes both "core" and "style-enhanced" funds, with the former benchmarked against major indices like CSI 800 and the latter focusing on enhancing returns based on beta [3][4] - Recent changes in Yang's portfolio management include a significant reduction in traditional value stocks and an increase in high-dividend stocks, which has resulted in a net asset value growth rate of 25.52% for his GF Steady Strategy fund in 2024 [5][6] - The introduction of new funds, such as GF Quality Selection and GF Research Selection, aims to address the challenges of maintaining a steady investment style while adapting to market volatility, with a combined scale of 109.8 billion yuan [6]

证券研究报告 金工 量化指增基金超额呈现边际修复 2026 年 1 月 24 日│中国内地 量化投资周报 本月以来估值因子偏弱，成长因子相对走强 本月以来估值因子整体偏弱，月初的回撤影响较大；波动率、换手率等防御 性量价因子同样表现不佳，仅在沪深 300 成分股票池取得正收益。成长、 盈利、小市值及反转因子相对走强，在中证 500 以外的成分股票池均呈现 正收益。预期类因子中，超预期因子在沪深 300 以外的股票池呈现正收益； 预期估值因子在沪深 300 和全 A 股股票池取得正收益，在其余股票池回撤； 预期增速因子则呈现普遍回撤。 本月以来小市值因子多空表现靠前，量价因子承压 从平均多空收益来看，本月以来小市值因子表现靠前，平均取得较明显的正 收益，但主要源于在沪深 300 成分股票池中的收益优势；预期增速和超预 期因子紧随其后；成长因子同样呈现正向的平均多空收益。反转、波动率和 换手率等量价因子整体承压，平均呈现较大回撤。 本月以来指增超额呈现边际修复，300 指增超额领先 我们重点跟踪以沪深 300、中证 500、中证 1000 和中证 A500 指数为基准 的量化指数增强基金。基于公募指增基金的复权净 ...

1. Report Industry Investment Rating No relevant information provided in the content. 2. Core View The report summarizes the performance, scale changes, net subscriptions/redemptions, valuation, margin trading, and fund managers of ETFs in the past week. It also mentions recent market events and new fund issuances [1][2][5]. 3. Summary by Related Catalogs ETF Performance - Last week (January 19 - 23, 2026), the median weekly return of equity ETFs was 0.95%. Among broad - based ETFs, the median return of CSI 500ETF was 4.34%, the highest. By sector, the median return of cyclical ETFs was 3.52%, the highest. By theme, the median return of photovoltaic ETFs was 6.87%, the highest [1][12][15]. ETF Scale Changes and Net Subscriptions/Redemptions - Last week, equity ETFs had a net redemption of 338.098 billion yuan, with the overall scale decreasing by 309.829 billion yuan. Among broad - based ETFs, CSI 500ETF had the least net redemption of 8.427 billion yuan; by sector, cyclical ETFs had the most net subscriptions of 29.338 billion yuan; by theme, chip ETFs had the most net subscriptions of 12.774 billion yuan [2][27][30]. ETF Benchmark Index Valuation - In broad - based ETFs, the valuation quantiles of ChiNext and SSE 50ETF were relatively low. By sector, the valuation quantiles of large - finance and consumer ETFs were relatively moderate. Among sub - themes, the valuation quantiles of liquor and new energy vehicle ETFs were relatively low. Compared with the previous week, the valuation quantiles of consumer and bank ETFs decreased significantly [3][33][45]. ETF Margin Trading - From Monday to Thursday last week, the margin trading balance of equity ETFs decreased from 53.910 billion yuan to 53.456 billion yuan, and the short - selling volume increased from 2.276 billion shares to 2.303 billion shares. Among the top 10 ETFs in terms of average daily margin purchases and short - selling volume, Science and Technology Innovation Board ETFs and CSI 500ETF had relatively high average daily margin purchases, while CSI 1000ETF and CSI 500ETF had relatively high average daily short - selling volumes [4][46][51]. ETF Managers - As of last Friday, Huaxia, E Fund, and Huatai - PineBridge ranked in the top three in terms of the total scale of listed non - monetary ETFs. This week, 6 ETFs will be issued, including Huaan CSI Non - ferrous Metal Mining Theme ETF, ICBC Credit Suisse CSI All - Index Electric Power and Utilities ETF, etc. [5][52][55].

Core Insights - The article highlights the vision and strategic approach of Mingxi Capital, led by Chen Haowei, to build a world-class quantitative investment methodology in China, emphasizing a collaborative and innovative framework [1][2]. Group 1: Company Overview - Mingxi Capital was founded in 2014 by Zhang Xiangfang, who established a strong foundation in the futures quantitative field, leading to deep market insights and initial capital accumulation [2]. - The company transitioned from a single-engine strategy to a dual-core driving model, integrating the expertise of partners with backgrounds from top hedge funds to enhance its operational capabilities [2]. Group 2: Strategic Development - The partnership between Zhang Xiangfang and Chen Haowei, along with other key members, is characterized by a complementary strategy that aims to elevate the firm from futures to a larger stock market capacity [2]. - The introduction of the NOVA system represents a significant technological advancement, designed to serve as an "AI partner" that enhances the investment research process through advanced data handling and strategy development [4][5]. Group 3: Technological Innovation - The NOVA system consists of three core intelligent modules: NOVA Matrix for data structuring, NOVA Go for strategy generation, and NOVA Pilot for risk control and portfolio optimization, creating a closed-loop system for investment operations [4][5]. - The system has proven effective during market volatility, allowing the firm to manage risks proactively and maintain operational stability [6]. Group 4: Organizational Culture - Mingxi Capital fosters a collaborative environment akin to a "Bell Labs" ecosystem, promoting knowledge sharing and innovation among team members [6][7]. - The firm employs a unique contribution attribution system that rewards not only final outcomes but also the entire process of research and development, encouraging a culture of collaboration and respect for foundational work [6][7]. Group 5: Future Vision - Looking ahead, Mingxi Capital aims to continue upgrading the NOVA system and invest heavily in computational infrastructure to enhance AI capabilities and attract top talent globally [7]. - The overarching goal is to maintain a competitive edge through continuous evolution, driven by a cohesive partnership, an autonomous intelligent system, and a supportive organizational ecosystem [7].

- The report mentions the use of a **BARRA factor model** to construct ten style factors for the A-share market, including Size (SIZE), Beta (BETA), Momentum (MOM), Residual Volatility (RESVOL), Non-linear Size (NLSIZE), Valuation (BTOP), Liquidity (LIQUIDITY), Earnings Yield (EARNINGS_YIELD), Growth (GROWTH), and Leverage (LVRG) [60] - The **construction of the A-share sentiment index** is based on market volatility and trading volume changes, dividing the market into four quadrants. Only the quadrant with increasing volatility and decreasing trading volume shows significant negative returns, while the other three quadrants show significant positive returns. The sentiment index includes bottoming and peaking warning signals [36][39] - The **A-share prosperity index** is constructed using the YoY growth of the net profit attributable to the parent company of the Shanghai Composite Index as the Nowcasting target. The index reflects the high-frequency prosperity trend of the A-share market [31][34] - The **theme mining algorithm** identifies investment opportunities by processing news and research report texts, extracting theme keywords, exploring relationships between themes and individual stocks, constructing active theme cycles, and building theme influence factors. The report highlights the "Commercial Aerospace" theme as a recent opportunity [49] - The **enhanced index portfolios** for CSI 500 and CSI 300 are based on strategy models. The CSI 500 enhanced portfolio achieved a cumulative excess return of 48.49% since 2020, with a maximum drawdown of -9.51%, while the CSI 300 enhanced portfolio achieved a cumulative excess return of 45.73% since 2020, with a maximum drawdown of -5.86% [49][55] - The **performance of style factors** over the past week shows that Beta factor had the highest excess return, while Size factor exhibited a significant negative excess return. High Beta stocks performed well, while Size and other factors underperformed [61][67] - The **factor exposure correlation matrix** indicates that liquidity is positively correlated with Beta, Momentum, and Residual Volatility, while valuation is negatively correlated with Beta, Residual Volatility, and Liquidity [62][63] - The **factor performance attribution for major indices** reveals that indices like CSI 500, ChiNext, and Wind All A benefited from high Beta exposure, while indices like Shanghai Composite and SSE 50 underperformed due to lower Beta exposure [69][70][73]

Quantitative Models and Construction Methods Model: AI CSI 1000 Enhanced Portfolio - **Construction Idea**: The model is based on the full-spectrum fusion factor, which integrates both high-frequency and low-frequency price-volume data using deep learning and multi-task learning techniques[6][7] - **Construction Process**: 1. Train 27 high-frequency factors using a deep learning model to obtain high-frequency deep learning factors 2. Use multi-task learning to extract end-to-end features from low-frequency price-volume data, resulting in low-frequency multi-task factors 3. Combine the high-frequency and low-frequency factors to form the full-spectrum fusion factor[6] - **Evaluation**: The model shows significant excess returns and a high information ratio, indicating strong performance and effective risk management[1][7] - **Backtest Results**: - Annualized excess return: 21.60% - Annualized tracking error: 6.06% - Information ratio (IR): 3.57 - Maximum drawdown of excess return: 7.55% - Calmar ratio of excess return: 2.86[1][7] Model: LLM-FADT Text Stock Selection Strategy - **Construction Idea**: The model enhances the BERT-FADT strategy by incorporating additional interpretations from a large language model (LLM), including new title interpretations, market catalysts, implied meanings, potential risks, and return guidance[2][14][17] - **Construction Process**: 1. Input six types of text into a fine-tuned FinBERT model: original text, new title interpretations, market catalysts, implied meanings, potential risks, and return guidance 2. Convert these texts into text feature vectors 3. Train an XGBoost model using these enriched text features[17] - **Evaluation**: The LLM-FADT strategy is more stable and has smaller excess drawdowns compared to the BERT-FADT strategy, showing better performance in extreme market conditions[2][14][20] - **Backtest Results**: - Annualized return: 30.10% - Annualized excess return: 25.52% - Sharpe ratio: 1.18 - Information ratio (IR): 2.00[2][20][24] Model: AI Industry Rotation Model - **Construction Idea**: The model uses the full-spectrum fusion factor to score 32 primary industries and constructs a weekly rebalancing strategy by equally weighting the top 5 industries[3][38] - **Construction Process**: 1. Score each industry using the full-spectrum fusion factor based on the industry component stocks 2. Select the top 5 industries with the highest scores 3. Equally weight these industries and rebalance weekly[38][43] - **Evaluation**: The model effectively utilizes AI's feature extraction capabilities to capture patterns in multi-frequency price-volume data, complementing top-down strategies[3][38] - **Backtest Results**: - Annualized return: 26.87% - Annualized excess return: 19.02% - Maximum drawdown of excess return: 12.43% - Sharpe ratio of excess return: 1.85[3][41] Model: AI Thematic Index Rotation Model - **Construction Idea**: The model scores 133 thematic indices using the full-spectrum fusion factor and constructs a weekly rebalancing strategy by equally weighting the top 10 thematic indices[4][28] - **Construction Process**: 1. Score each thematic index using the full-spectrum fusion factor based on the index component stocks 2. Select the top 10 thematic indices with the highest scores 3. Equally weight these indices and rebalance weekly[28][31] - **Evaluation**: The model leverages AI to identify and capitalize on trends in thematic indices, providing a diversified and dynamic investment approach[4][28] - **Backtest Results**: - Annualized return: 16.92% - Annualized excess return: 9.37% - Maximum drawdown of excess return: 20.79% - Sharpe ratio of excess return: 0.73[4][30] Model Backtest Performance AI CSI 1000 Enhanced Portfolio - Annualized excess return: 21.60% - Annualized tracking error: 6.06% - Information ratio (IR): 3.57 - Maximum drawdown of excess return: 7.55% - Calmar ratio of excess return: 2.86[1][7] LLM-FADT Text Stock Selection Strategy - Annualized return: 30.10% - Annualized excess return: 25.52% - Sharpe ratio: 1.18 - Information ratio (IR): 2.00[2][20][24] AI Industry Rotation Model - Annualized return: 26.87% - Annualized excess return: 19.02% - Maximum drawdown of excess return: 12.43% - Sharpe ratio of excess return: 1.85[3][41] AI Thematic Index Rotation Model - Annualized return: 16.92% - Annualized excess return: 9.37% - Maximum drawdown of excess return: 20.79% - Sharpe ratio of excess return: 0.73[4][30]