StockSemáforo uses a transparent, deterministic rule-based system: no black boxes. Each stock is scored from 0 to 100 across four areas. The overall score gives more weight to valuation —the "cheap or expensive?" question— than to the average of quality (growth, profitability and solidity). The fundamentals come from the official reports companies file with the U.S. SEC (EDGAR), which are in the public domain: they aren't editable, we trust the source. The only thing you add is the current price.
This area needs the current price, which you enter (we link you to Google Finance). With it we compute the P/E (price ÷ earnings) and compare it to a "fair P/E": the multiple the company deserves. That fair P/E isn't a fixed number: it starts from its sector reference (a P/E of 30 is normal in software, expensive in banking), but it adjusts to the company's growth. A fast-growing company deserves a higher multiple than its sector average —that's the "growth at a reasonable price" (GARP) logic—. Specifically, the fair P/E is the higher of the sector reference and ~1.2 times its annual growth. Trading below the fair P/E scores high; above, low.
A key nuance to keep credibility: an outrageous price sinks the score, however good the company. When the P/E far exceeds what growth justifies, we apply a penalty thatcaps the overall grade: quality can't "rescue" an absurd price (paying ten times fair value is a bad investment even for an excellent company). That penalty grows smoothly and continuously with the price, with no jumps.
To judge growth within valuation we use a prudent mix of 1- and 5-year growth (if the company has accelerated, the 5-year; if it has decelerated, the average of both), so that neither an exceptional year nor a weak one distorts the result. Without a price, this area doesn't score and the grade rests on the other three.
Average of the annualized growth (about a 5-year CAGR) of earnings per share and of revenue. We use several years, not the last one, so we're not fooled by a single exceptionally good or bad year —typical in cyclical sectors—. Growing 30% a year or more scores 95; flat, around 30; declining, less.
Combines net margin (25%+ is excellent), gross margin (60%+ signals pricing power) and ROE. We cap ROE at 40%: above that it usually reflects leverage or aggressive buybacks (which shrink equity), not more real quality.
Net debt/EBITDA (below 0.5x is strength; above 4x, risk) and free cash flow margin (20%+ is elite; negative is a warning sign).
Raising the dividend for many consecutive years demands growing cash generation and balance-sheet discipline: a quality signal that single-year ratios don't capture. So the score adds +1 point for every 4 consecutive years of increases, up to +4(per the SEC filing series, which covers about 15 years). It's a bonus, never a penalty: not paying a dividend doesn't subtract —Amazon or Berkshire aren't worse for reinvesting everything—. And it has a lock: no bonus if the payout exceeds 90% of earnings (a streak bought by paying out nearly everything has an expiry date). Nor can the bonus rescue an absurd price: the overvaluation cap rules.
The sector P/E reference is an approximation based on the company's classification, not a live market average. The growth, margin and debt thresholds are still general. The system doesn't assess qualitative factors (competitive advantages, management, regulatory risks). For companies without profits, the P/E doesn't apply. Fundamentals are computed over the last twelve months (TTM): we combine the latest closed fiscal year with the most recent quarters, so they reflect the current situation, not that of a year ago. That TTM may include one-off items (charges or gains) that distort a specific period. And in companies with several share classes, EPS may not be available, so those can't be valued precisely. Use the traffic light as a first filter, never as a final decision.
This tool is educational and does not constitute financial advice.