Scoring Multi-Strategy Portfolios by Correlation, Coverage, and Diversity
Summary
The document describes a tool for assessing portfolios of expert advisors (EAs) across three dimensions: correlation between daily profit-and-loss series, the hours and weekdays when strategies trade, and diversity across asset classes. It combines dimension scores into a letter grade and reports correlation flags, coverage summaries, and recommendations. The correlation threshold and the relative weights assigned to each dimension are configurable.
The analyzer reads daily results from CSV files, with fields for date, daily P&L, trade hour, and weekday. It requires a configurable minimum number of days per EA and can generate sample data when input files are absent. The document gives no empirical validation, performance results, or detailed scoring formula, so the grade should be treated as a screening aid rather than evidence that a portfolio is profitable or robust. Coverage by hour and asset category also does not, by itself, establish that strategies will diversify losses under changing market conditions.
Key ideas
- The analyzer evaluates portfolios using strategy correlation, temporal coverage, and asset class diversity.
- Daily P&L series are compared pairwise using Pearson correlation, with configurable threshold flags.
- Hourly and weekday summaries show when the included strategies are active.
- A weighted composite score converts the dimensions into a letter grade and recommendations.
- The document does not provide performance validation or enough detail to reproduce the scoring formula.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.