菜单
Percentile-based equity scoring methodology — 18 indicators across 6 categories, with sector-adaptive weights and outlier cap rules
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Xiaohongshu (小红书) browser automation skill. Use when browsing/searching content, extracting post data, or interacting with the platform via playwright. Requires pre-authenticated storage state.
Chinese traditional calendar and BaZi (Four Pillars of Destiny) skill. Provides CLI scripts for birth chart analysis, marriage compatibility, and auspicious date selection. Wraps the china-testing/bazi library with lunar-python.
Ctrip (携程) hotel price query tool. Use when checking hotel prices, comparing rates, or monitoring price changes on Ctrip via Playwright browser automation. Requires pre-authenticated storage state.
Amap (Gaode Maps) Web Service API skill. Provides CLI scripts for POI search, route planning, geocoding, and weather queries. Zero dependencies — uses Node.js native fetch (Node 18+).
Single-file HTML slide-deck template for Tao Te Ching chapter decks — ink-style, 10-page narrative, hand-rolled engine, reusable CSS variables and animation classes
Send emails via QQ Mail SMTP using nodemailer. Use when sending notifications, reports, or alerts via email.
| name | sector-equity-scoring |
| scope | langsensei |
| description | Percentile-based equity scoring methodology — 18 indicators across 6 categories, with sector-adaptive weights and outlier cap rules |
| version | 1.0.0 |
| dependencies | {"skills":["https://github.com/LangSensei/emploke-marketplace/tree/main/skills/eastmoney-data"]} |
A reusable methodology for building quantitative value-scoring formulas for any A-share sector. Each formula produces a 0-100 composite score per stock, combining valuation, profitability, growth, cash flow, risk, and industry quality indicators via percentile ranking.
Every sector formula targets 18 indicators organized into 6 categories:
| Category | Standard Indicators | Typical Weight |
|---|---|---|
| Valuation (3) | PE-TTM, PB, Dividend Yield | ~15-25% |
| Profitability (5) | ROE, ROA, ROIC, Gross Margin, Net Margin | ~25-35% |
| Growth (2) | Revenue Growth YoY, Net Profit Growth YoY | ~10-25% |
| Cash Flow (3) | Operating CF / Net Profit, FCF Yield, Operating CF / Revenue | ~8-15% |
| Risk (3) | Debt-to-Asset Ratio, Current Ratio, Interest-bearing Debt Ratio | ~8-12% |
| Industry Quality (2) | Gross Margin YoY Change, Quick Ratio | ~5-10% |
Sector-specific substitutions are allowed. Financial sectors (banking, insurance, securities) replace inapplicable indicators with industry-specific ones (e.g., NPL ratio, provision coverage, net interest margin for banks; combined ratio, investment yield for insurance). The total must still target 18 indicators.
Each indicator is scored by ranking all stocks in the sector and converting to a 0-100 percentile:
score_i = percentile within sector (0-100)
For "lower is better" indicators (PE, D/A, etc.):
score = 100 - percentile
For "higher is better" indicators (ROE, dividend yield, etc.):
score = percentile
percentile = (rank - 1) / (N_valid - 1) * 100
Tie-breaking: When multiple stocks share the same indicator value, they receive the same rank (standard competition ranking). The next rank is offset by the number of tied stocks (e.g., if two stocks tie at rank 3, the next stock receives rank 5).
Edge case — N_valid = 1: When only one stock has a valid value for an indicator, the percentile formula produces 0/0. In this case, assign a score of 50 (neutral) for that indicator.
The composite score is a weighted sum:
Value_Score = Sum(weight_i * score_i) / Sum(weight_i_participated)
Missing indicators are excluded — the denominator adjusts to only include weights of indicators with valid data.
These rules prevent outliers from distorting rankings:
| Rule | Description |
|---|---|
| Growth ±200% cap | Revenue and net profit growth rates are capped at ±200% before ranking. Display the capped value (e.g., 200.0, not the raw 1500%) |
| ROE/ROIC ≤60% cap | ROE and ROIC values exceeding 60% are capped at 60% before ranking |
| PE<0 exclusion | Loss-making stocks (negative PE) are excluded from PE ranking entirely (they receive no PE score) |
| Coverage penalty | If a stock has fewer than 10 valid indicators (out of 18), apply penalty: score *= valid_count / 18 |
| ST/*ST exclusion | All ST and *ST stocks are excluded from the ranking universe |
| B-share exclusion | B-share stocks (codes starting with 9xxxxx) are excluded |
FCF_Yield = FCFF_BACK / f20
FCFF_BACK = Free cash flow to firm (from datacenter-web API: RPT_F10_FINANCE_MAINFINADATA)f20 = Total market capitalization (from push2 batch API)f116 (circulating market cap) — use total market capWeights must sum to 100% and reflect sector economics. Use these guidelines:
| Category | Default | Range |
|---|---|---|
| Valuation | 20% | 15-35% |
| Profitability | 30% | 25-35% |
| Growth | 15% | 10-25% |
| Cash Flow | 10% | 8-15% |
| Risk | 10% | 8-12% |
| Industry Quality | 8% | 5-10% |
There is no single correct weight allocation — it should be driven by sector characteristics.
| Sector Type | Adjustment | Rationale |
|---|---|---|
| Consumer staples (food, beverage, household) | Raise Profitability to ~35%, emphasize gross margin | Brand moat drives pricing power |
| Heavy-asset / PPP sectors (utilities, construction, transportation) | Raise Cash Flow to ~13-15%, raise Risk to ~12% | Capital-intensive, cash flow quality critical |
| Cyclical sectors (steel, coal, nonferrous, chemicals) | Lower Growth to ~10-12%, raise Valuation to ~25% | Growth is volatile and mean-reverting |
| Growth sectors (electronics, computer, pharma) | Raise Growth to ~20-25%, lower Valuation to ~15% | Revenue trajectory is the primary value driver |
| Financial sectors (banking, insurance, securities) | Replace Cash Flow/Risk with industry-specific indicators; Valuation ~30-35% | Standard cash flow metrics don't apply |
Within each category, distribute weight based on indicator importance:
Every sector formula document follows this structure:
Perform all 7 standard checks:
| # | Check | What to Verify |
|---|---|---|
| 1 | Loss-making stocks | All PE<0 stocks should cluster near the bottom of rankings |
| 2 | High-score logic | Top-ranked stocks should have strong fundamentals across multiple categories, not just one extreme metric |
| 3 | Low-score logic | Bottom-ranked stocks should show clear weaknesses (losses, high debt, declining revenue) |
| 4 | Industry leader ranking | Well-known sector leaders should appear in a reasonable range (top quartile, though not necessarily #1) |
| 5 | Score distribution | Mean ~49-51, median ~49-52, >70 scores <15% of total, healthy spread |
| 6 | Cap validation | Verify that capped values appear correctly in ranking table (200.0 for growth, 60.0 for ROE/ROIC) |
| 7 | Sub-industry coverage | Check that the formula doesn't systematically favor one sub-industry over others within the sector |
Present checks as a verification table: dimension, expected, actual, conclusion (pass/fail).
Include a Limitations section noting known weaknesses (e.g., YoY vs CAGR, missing indicators, cross-sectional only).
| API | Endpoint | Purpose |
|---|---|---|
| Batch quote API | push2.eastmoney.com/api/qt/clist/get | Sector stock list + basic indicators (PE, PB, market cap, dividend yield) |
| Financial data API | datacenter-web.eastmoney.com/api/data/v1/get | Detailed financials per stock (RPT_F10_FINANCE_MAINFINADATA) |
Refer to the eastmoney-data skill for API usage details, field mappings, and request patterns.
See reference/a-share/ for 32 completed A-share sector formulas (01-banking through 32-conglomerate). These serve as concrete examples of the methodology applied to different sector types.