ABSTRACT

When it comes to asset allocation and portfolio management, Kelly criterion is a mathematical formula used to optimise expected log-returns over the long term. Nonetheless, not all stocks are well suited for analysis using Kelly criterion, due to their transient nature and noisy data. This paper presents an innovative index by which to assess the suitability of stocks for analysis using the Kelly criterion. When applied to real-world stock data, the correlation coefficient between the proposed KSI and log-returns based on the Kelly criterion was $-57.045\mathrm{\%}$ with a $p$-value of $1.215\times {10}^{-1}$. In a robustness test based on the Mid-Cap 100 dataset, the correlation coefficient was $-44.064\mathrm{\%}$ with a $p$-value of $2.438\times {10}^{-5}$. The results demonstrate the efficacy of the KSI for portfolio management.

摘要

在资产配置和投资组合管理方面，凯利准则是用于优化长期预期对数回报的数学公式。尽管如此，由于它们的瞬态性和嘈杂的数据，并非所有股票都非常适合使用凯利标准进行分析。本文提出了一个创新指数，通过该指数来评估股票是否适合使用凯利标准进行分析。当应用于真实世界的股票数据时，基于凯利标准的 KSI 和对数回报之间的相关系数为$-57.045\mathrm{\%}$与$p$-的价值$1.215\times {10}^{-1}$. 在基于 Mid-Cap 100 数据集的稳健性测试中，相关系数为$-44.064\mathrm{\%}$与$p$-的价值$2.438\times {10}^{-5}$. 结果证明了 KSI 对投资组合管理的有效性。