aiqbench analysis
This paper applies reinforcement learning, specifically an iterated-regression method, to a dynamic portfolio problem: choosing among equity factors each month under a switching cost. The switching cost regularises the learned policy against estimation noise, and the authors give asymptotic consistency results. The out-of-sample gains over a myopic policy and the best fixed factor are backtested on about 60 years of U.S. data, so they warrant scrutiny before anyone treats them as an edge. For ML readers, the transferable idea is using an explicit cost on action changes to stabilise a learned policy in a noisy, low-signal environment.
Dynamic Strategy Selection
arXiv:2307.04754v3 Announce Type: replace Abstract: We consider the problem of dynamically selecting among competing strategies using observable predictors. To mitigate estimation noise and excessive switching, we regularise the estimated policy with a switching cost, which may exceed actual transaction costs when these are too small to provide sufficient regularisation. Since switching costs make current choices affect future decisions, we formulate the problem using dynamic programming. The optimal policy is estimated using a reinforcement learning algorithm based on iterated regressions, making the procedure simple to implement. We apply the method to monthly rotation among momentum, value, profitability, and defensive U.S. equity factors over a sample of about 60 years. Using a parsimonious set of state variables, the estimated policy outperforms both the myopic dynamic policy and the best fixed factor in hindsight in out-of-sample evaluation. Under weak conditions, we show theoretically that the decision rule estimated from empirical data is asymptotically consistent for the unknown optimal decision rule. Simulation results illustrate the finite-sample relevance of these asymptotic results.
Abstract:We consider the problem of dynamically selecting among competing strategies using observable predictors. To mitigate estimation noise and excessive switching, we regularise the estimated policy with a switching cost, which may exceed actual transaction costs when these are too small to provide sufficient regularisation. Since switching costs make current choices affect future decisions, we formulate the problem using dynamic programming. The optimal policy is estimated using a reinforcement learning algorithm based on iterated regressions, making the procedure simple to implement. We apply the method to monthly rotation among momentum, value, profitability, and defensive U.S. equity factors over a sample of about 60 years. Using a parsimonious set of state variables, the estimated policy outperforms both the myopic dynamic policy and the best fixed factor in hindsight in out-of-sample evaluation. Under weak conditions, we show theoretically that the decision rule estimated from empirical data is asymptotically consistent for the unknown optimal decision rule. Simulation results illustrate the finite-sample relevance of these asymptotic results.
Substantially revised version. Proofs and bounds have been corrected, the focus and application have changed, and a new coauthor has been added
Machine Learning (cs.LG); Portfolio Management (q-fin.PM); Methodology (stat.ME); Machine Learning (stat.ML)
arXiv:2307.04754 [cs.LG]
(or arXiv:2307.04754v3 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2307.04754 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 284 words
Related coverage
Parallelism Strategy Chaining for Fast Training Convergence
arXiv · Sep 9, 2026
Soft Strategy Selection for Batch-Mode Active Learning
arXiv · Oct 7, 2026
LLM-Guided Reinforcement Learning for Adaptive NPC Behavior in Multi-Agent Combat Games
arXiv · Sep 4, 2026
Select and Improve: Understanding the Mechanics of Post-Training for Reasoning
arXiv · Jun 12, 2026