Core Viewpoint - The article emphasizes that relying solely on historical performance data to select consistently excellent active equity funds is challenging, and understanding the reasons behind a fund's strong performance is more crucial than analyzing past results [1]. Group 1: Fund Performance Analysis - The article reviews the performance and distribution characteristics of active equity funds over the past decade, indicating that understanding the investment philosophy and framework is essential for investors [1]. - A sample of 466 funds was analyzed, focusing on those established before June 30, 2015, with performance benchmarks including over 60% of the CSI 300 or CSI 800 indices [2]. - The cumulative return distribution of active equity funds was examined over two periods: from October 2015 to September 2020 (a bullish market) and from October 2020 to September 2025 (a challenging market) [4]. Group 2: Return Distribution Characteristics - The distribution of fund returns is right-skewed, indicating that most funds cluster around lower return values, with a long tail on the right side representing a few extreme high performers [7]. - In the difficult market environment from October 2020 to September 2025, the right-skewed distribution is particularly pronounced, suggesting that most funds will tend to perform around the market average [7]. Group 3: Mean Reversion in Fund Performance - The analysis categorized funds into five performance tiers based on cumulative returns, revealing that only 16% of the top-performing funds from the first period remained in the top tier in the second period, while 27% of the lowest-performing funds improved to the top tier [10]. - The distribution of funds across performance tiers in the second period shows no clear persistence, indicating that historical performance is not a reliable indicator for future fund selection [11]. Group 4: Investment Recommendations - For most investors, attempting to pick the next star fund is discouraged; instead, understanding the underlying philosophy and maintaining conviction during tough times is recommended [13]. - Utilizing broad-based index funds with controlled tracking error is suggested as a more rational and lower-risk long-term investment strategy for those unwilling to invest time in active management [13].

人类最伟大的能力，不是知道答案，而是永远保持提问与重构答案的可能。 L. G. Valiant 在规则明确、数据丰富的任务中AI已展现出越来越强大的超人类能力。比如OpenAI宣称，GPT-4参加了多 种基准考试，包括美国律师资格考试、法学院入学考试LSAT、SAT数学部分和证据性阅读与写作部分考 试等。在这些测试中，GPT-4得分高于88%的人类应试者。 AI技术的快速进展和成熟，无疑从很多方面影响和改变着我们的生活和工作方式，尤其是在 整合共识和 主流叙事、优化"已知"领域的运行效率 方面，现有的AI模型已体现出强大的能力。 相比与当前的AI，人类的价值和竞争优势体现在哪里？个人赞同以下观点， 保持对异常现象的好奇心， 勇于挑战共识，提出有价值的假说，仍是人类最值得珍视的能力。 AI的优势：高效整合历史经验和共识， 优化"已知"领域的运行效率 当前以大语言模型（LLMs）为代表的AI模型，仍然是一个 数据和计算驱动，基于历史频率、相关性及均 值的统计归纳系统 。它的"智能"，是基于"同一事物可以通过无限种方式被陈述、表达和表征"这一事 实。其能力源于语言表征的泛化能力，即能够将一种表达方式转化为另一种表达 ...

对于不少投资者而言，量化投资像是个带着神秘色彩的"黑匣子"。近年来，随着量化投资的蓬勃发展及人 工智能的加持，越来越多的投资者开始主动关注、理解这类投资策略。 但对于不同的基金经理，量化工具发挥的作用并不相同。本期，我们邀请中泰证券资管量化公募部总经理 李玉刚做客直播间，就以下问题进行解答。 李玉刚 我们可以把投资当作认知跟决策的一个分支，大家经常说投资是认知的变现。说到对投资的评 价，首先要知道投资是一个属于不确定性的领域的游戏， 在这样的游戏中，我们更应该关注过程，而不 是注重结果 。 主播 为什么？作为投资者，肯定是关注收益的。 李玉刚 对，但短期的收益是受很多因素影响。短期好的结果，不代表整个过程是好的。从长期来说，我 们知道长期投资收益来源于较高质量的决策过程，这是一个前提。 主播 也就是说长期的收益不是短期收益积累来的。 李玉刚 它不是简单加权的过程。比如一只股票可能长期表现不好，或处于下跌状态，但投资这只股票的 收益，在这种不确定性领域的收益，不是时间的加权（通过时间简单累积收益），这点是比较明显的。 讨论这个问题，大家应该接受一个前提，就是在投资上，尤其是注重长期收益的投资者，我们更应该关注 ...