Why Dynamic Portfolio Optimisation Can Fail for Small Accounts
Summary
This post examines dynamic portfolio optimisation for a relatively small trading account. It describes a historical backtest setup that estimates instrument correlations and covariance, derives expected returns from existing portfolio weights and risk aversion, and converts contract values and trading costs into portfolio terms. The optimiser searches a grid of feasible positions while accounting for risk, costs, maximum instrument exposure, sign restrictions, and prior-period weights.
The document also compares original, rounded, and optimised positions through turnover and portfolio-risk calculations. Its title signals a negative result, and the author frames a static subset of markets as the better approach to examine in a later post. The supplied extract is incomplete: the performance and conclusion sections contain no results, and the optimisation depends on historical estimates and chosen constraints. It therefore explains a test design and trade-offs, but does not provide enough outcome data here to judge the method's performance.
Key ideas
- The proposed optimiser searches discrete position choices using expected return, risk, and transaction cost terms.
- Covariance estimates and implied expected returns are inputs to the portfolio decision.
- Position constraints include risk ceilings, instrument limits, sign restrictions, and trade restrictions.
- Passing prior optimised weights into later periods makes the procedure dynamic and allows costs to depend on turnover.
- The extract describes a backtest design but omits its performance results and detailed conclusion.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.