Fragility checks score 0 (robust) to 4 (fragile); upside checks 0 (none) to 4 (high). Numeric inputs map smoothly onto 0–4 and are averaged per check; in rate sensitivity and the zombie test the worst input counts half, so a harmless input cannot cancel a serious one; subsidy dependence takes the strongest channel. A clicked rating overrides inputs. Fragility and upside are weighted averages × 25. Overall = upside × weight + (100 − fragility) × (1 − weight). Verdicts: ≥70 attractive, ≥55 interesting, ≥40 neutral, else unattractive.
- Stress. Higher real rates lift rate, zombie and credit scores (more for weak ones) and cut valuation upside. Subsidy cuts scale subsidy dependence; extra inflation lifts debasement and lowers the Crunch level when margins fell in 2021–23.
- Valuation impact. P/E read as a perpetuity: price change = 1 / (1 + P/E × Δ) − 1.
- Rate ladder. All scores recomputed at each rate shock with the current subsidy and inflation settings.
- Subsidies. When a company loads, Claude first classifies its sector. In sectors that rarely get subsidies one quick search checks for company-specific rescues or state stakes; if none turn up, Claude gives a sector-based estimate (command subsidies forces the full search). In subsidy-associated sectors (renewables, utilities, autos and EVs, semiconductors, airlines, defence, heavy industry, Chinese state-linked firms and others) a wide search runs: Parallel Search looks for grants, state guarantees and state ownership in aid registers (Subsidy Tracker, EU State Aid database), annual reports and filings. Claude extracts amounts, and the page only accepts an amount if Claude's quote appears word for word in the source; each value links to it. Command subsidies reruns it.
- Price chart. Prices for context, never shorter than one year: FMP where its free plan allows (a few US symbols), otherwise Scalable Capital by ISIN, in the currency Scalable quotes (EUR for most US shares). The regime bar at the right edge has no timing. Sound: latest-year free cash flow (operating cash flow minus capex) at the revenue growth actually delivered over five years, at your discount rate. Crunch: 3-year average free cash flow at the weakest 3-year growth of the last seven years, minus up to 5 pp for fragility, at the discount rate +2 pp (or the rate-stress slider if higher), cut further by the subsidy and inflation sliders. Boom: the highest price the market paid per unit of 3-year average free cash flow in the last five years, applied to today's 3-year average (modelled when there is no price history). Free cash flow is after interest, so the values are per share without a separate debt adjustment. For sectors the chart is a member index, weighted by market cap (equal-weighted across currencies) and set to 100 at the start of the range.
- Valuation. Scored only from reported figures: owner earnings (operating cash flow minus maintenance capex) against enterprise value, a reverse DCF on enterprise value (what growth the price assumes, compared with the revenue growth actually delivered) and Spitznagel's Faustmann ratio against what ROIC justifies. The discount rate is 9 % unless you set it with discount 8; discount default resets it. Analyst consensus is shown but not scored. Banks and insurers get no cash-flow valuation.
- Sectors. Claude only chooses the member companies (at most the max you set, fewer if fewer clearly belong; conglomerates and part-time players are left out). You can untick or add members. Each member is loaded from EODHD with the same formulas as a single company, and each sector input is the median across the members. Share of firms with interest coverage below 1 is counted directly.
- Source order. Search, statements, consensus and macro: EODHD first, FMP as fallback for US. Share price: FMP for US (current), otherwise EODHD's last close. P/E is calculated here as price ÷ trailing EPS, except where statements and share price are in different currencies. Your own entries override everything.
- Data. Tickers like AAPL, ENR.XETRA, ADX.MC, 1211.HK. Latest annual statements, price and analyst consensus; EODHD's P/E is its own trailing figure. Ratios are calculated here; each input shows its formula and period. Inflation and GDP come from EODHD macro data; demand growth and expected return use analyst consensus (forward earnings yield + inflation as a rough proxy). Subsidies and state support are not in any database: enter them or let Claude rate the check as a judgment. Moat, catalysts and debasement are scored from data (margins, ROIC persistence, estimate revisions, margin recovery, capex vs depreciation, margins through the 2021–23 inflation, net cash, real assets). Claude only rates subsidy dependence, as a judgment, never figures. Your plans decide coverage.
- Commands. load <ticker or name> · verified on|off · company <name> · sector <name> · fill · save · new · open <saved> · rate <check> <0-4> · stress rates|subsidy|inflation <n> · stress clear · upside <0-70> · screen · compare · notes. ↑ recalls.
- Limits. Thresholds are judgment calls. Scores show exposure, not timing. Not investment advice.