Relative Returns

Title: Relative Returns

Main Research Question: How can we develop a framework to universalize investment strategies, allowing them to approach the optimal performance of strategies with known optimal parameters?

Methodology: The authors propose a general framework for universalizing investment strategies. This involves developing conditions under which investment strategies are universalizable and creating examples of common investment strategies that fit into this framework. These examples include both trading strategies that decide positions in individual stocks and portfolio strategies that allocate wealth among multiple stocks. The authors discuss the runtime efficiency of universalization algorithms, addressing the issue of their potential exponential complexity.

Results: The authors present a set of examples of investment strategies that can be universalized. These include the constantly rebalanced portfolio (CRP) strategy and various trading strategies. They show that their framework can be applied to a wide range of investment strategies, making it a powerful tool for universalizing investment strategies.

Implications: The development of this framework has significant implications for the field of investment strategies. It allows for the creation of online algorithms that can approach the optimal performance of strategies with known optimal parameters, even when these parameters are unknown at the time of execution. This can lead to more effective investment strategies and better management of investment portfolios. Additionally, the general nature of the framework means that it can be applied to a wide range of investment strategies, making it a versatile tool in the field.

Link to Article: https://arxiv.org/abs/0204019v1 Authors: arXiv ID: 0204019v1