Fund Comparison Analysis

Problem

When choosing between funds, analysts often receive several fund codes but lack a consistent way to compare peer rankings, drawdown, holdings, manager quality, and fees. Asking a model directly can produce mixed data sources and filled-in numbers that look plausible but are not traceable. This skill grounds public fund comparison in the Hang Seng Polaris MCP financial data path, so the output reads more like a reviewable research note than a general product pitch.

Core Capabilities and Workflow

The skill organizes comparison into modules:
- Basic information: fund type, age, scale, benchmark, and positioning
- Performance and risk: interval returns, max drawdown, Sharpe ratio, Calmar ratio
- Allocation and holdings: asset mix, industry concentration, top holdings and bonds
- Fund managers: tenure, capability scores, managed scale, and period returns
- Conclusion: rating, star score, action suggestion, source list, and disclaimer

Operationally, it checks mcporter and JY_API_KEY, then fetches data fund by fund. To limit context load and hallucination risk, each call handles a single fund and usually no more than three indicators. If a field is missing, the report leaves it blank or marks it as unavailable instead of guessing. The default output is a Markdown report; on request, it can generate MD and HTML files.

Boundaries

It targets public equity, hybrid, bond, money market, ETF, ETF feeder, and QDII funds. Cross-category comparisons still require manual judgment about risk-return differences. Data availability depends on the Hang Seng Polaris MCP service, API key, and field freshness. The scoring formula is an aid for comparison, not investment advice.