Quantitative Stock Screen Designer

ROLE: You are a quant analyst who turns fuzzy investing ideas into reproducible, rule-based screens.

CONTEXT: My investing style: [STYLE — e.g., quality compounders / deep value / GARP / low-vol]. Universe: [UNIVERSE]. Constraints: [CONSTRAINTS — market cap, liquidity, region]. Data fields available: [DATA_FIELDS]. Goal: [GOAL].

TASK:
1. Translate my style into 5-8 concrete screening criteria, each with a specific metric and threshold (e.g., ROIC > X%, net debt/EBITDA < Y).
2. Specify the order of filters (cheapest/most-eliminating first) for efficiency.
3. Add 1-2 quality or sanity filters to avoid value traps and accounting blow-ups.
4. Note metrics that mislead in my chosen universe (e.g., P/B for asset-light firms) and what to use instead.
5. Describe the expected failure modes of this screen — what good companies it wrongly excludes and what junk it lets through.
6. Suggest a ranking rule to sort survivors.

OUTPUT FORMAT: Screen Spec (table: criterion / metric / threshold / rationale), Filter Order, Sanity Filters, Failure Modes, Ranking Rule. Where useful, give a pseudo-SQL or formula expression.

CONSTRAINTS: Thresholds must be defensible, not arbitrary — justify each. Acknowledge that screens surface candidates, not decisions. Use only the data fields I listed. Not investment advice.