Core Viewpoint - The primary focus of the article is on the concept of risk in investing, particularly the risk of permanent capital loss, and how it can be minimized through various measures and indicators [3][11]. Group 1: Understanding Risk - "Risk" is often used in investing but lacks a clear definition, with the most significant aspect being the risk of permanently losing capital [3]. - Two dominant measures of risk in professional investing are equity beta, which indicates a stock's sensitivity to market movements, and Value at Risk (VaR), which estimates the maximum expected loss over a specific time frame at a defined confidence level [4][5]. Group 2: Limitations of Risk Measures - Both equity beta and VaR are criticized for being historical measures that may not predict future outcomes accurately, as they rely on past patterns [6]. - A Monte Carlo model can be used to enhance VaR calculations, but it does not guarantee protection against all potential outcomes [6]. Group 3: Risk Management Strategies - Private investors are encouraged to adopt a pragmatic approach to risk management by identifying key indicators that historically signal the end of a bull market [8]. - The author lists ten indicators that suggest a potential end to the current secular bull market, noting that all indicators are currently met, indicating a possible market downturn [11]. Group 4: Market Behavior and Speculation - The article highlights a prevailing speculative mentality among investors, where the desire to follow trends can lead to dangerous market behavior [14]. - Evidence of speculative fever is illustrated through the performance of unprofitable Nasdaq stocks, which have shown high returns, signaling excessive risk-taking [19]. Group 5: Current Investment Strategy - The company maintains a nearly fully invested position while adopting a defensive strategy, focusing on low beta equities and commodities, particularly gold, to mitigate risks [23]. - Despite a defensive approach, the company achieved a remarkable return of +29.24% for USD investors in 2025, raising concerns about the level of risk taken [24].

以上的传统方法所计量的VaR在剧烈波动的金融市场下很容易失效，核心原因在于其静态分布的假设在一定程度上忽略了 波动的时变性，导致无法量化波动率的动态变化。一般认为，金融市场中高波动率时期往往伴随着持续的高波动，低波动率时 期则延续低波动，即金融市场中资产价格的变化具有一定的趋同性，这种现象被称为波动率聚集，最早由法国数学家曼德布洛 特（Benoit Mandelbrot）提出，在数学上可以表示为在一段时间内，资产收益率绝对值的自相关函数为正，这种性质使得资产 收益率的波动率存在一定程度上的可预测性，而不是完全随机变化。GARCH类模型作为现代的金融时间序列模型，便是基于 波动聚集特性进行建模，能够更好地处理金融数据存在的尖峰肥尾特征和异方差问题。 GARCH模型（广义自回归条件异方差模型）是ARCH模型（自回归条件异方差模型）的延伸，即广义上的ARCH模型，相 比于ARCH模型，GARCH模型通过加入 项，能够捕捉历史波动对当前波动的累计效应，将ARCH模型 所依赖的高阶残差项通过历史残差平方的递归函数进行替代，更为符合金融数据波动的持续性特征。 一、引言 随着资本市场的不断深化发展，证券公司所面临市场风险的 ...