The Guide

1. $1B+ market cap and 10+ years public — enough history to judge.
2. 8+ years of financial statements with positive revenue throughout.
3. Revenue CAGR ≥ 0% over available history — no long-term shrinkers.
4. Diluted EPS CAGR ≥ 3%, and the latest year's EPS must be positive.
5. No 4+ consecutive years of revenue decline anywhere in the record.
6. ROIC above the cost of capital (WACC) in at least 6 of the last 10 years — proof the business creates value, with slack for investment years.
7. Positive free cash flow in at least 7 of the last 10 years.
8. Debt/EBITDA ≤ 4× and interest coverage ≥ 5× (debt-free companies skip coverage).
9. Stock-based compensation ≤ 12% of revenue — heavier SBC is also penalized again in scoring.
10. Composite quality score ≥ 55 after archetype-specific scoring (below).

Banks and insurers are excluded at the industry level (their statements don't fit this framework) — but payment networks, exchanges, and ratings agencies stay in. Occasionally the admin adds a stock that fails a gate on explicit judgment; those rows carry an EXCEPTION badge naming exactly which gate they failed, and show no fabricated score. Wondering why a specific stock is out? Search it on the universe page and hit "Understand why" for a per-ticker diagnostic.

A toll road shouldn't be judged on growth rate; a cyclical shouldn't be judged on current-quarter margins. Each business is classified first, then scored on the metrics that matter for that kind of business.

Companies that don't fit any archetype cleanly are labeled uncategorized rather than force-classified. That's a feature, not a bug.

The number in the Universe table is a composite of 5 archetype-specific quality checks — see it broken down on any ticker page.

A score in the 80s for a Toll Road means something different than 80 for a Cyclical — same number, different criteria. Compare within an archetype, not across.

Every surface of this product serves one discipline, shown on each stock's page as The Setup — three conditions, each a checkable fact with dates, never a score:

1. Unusually cheap — in the deep Buy Zone: the cheapest decile of its own 5-year valuation history. Cheapness is the cushion: at a 5-year extreme, a lot of bad news is already in the price.
2. Fears named — the specific, dated reasons it's cheap are written on the fear ledger. If you can't name the fear, you can't recognize the moment it dies.
3. A fear easing — a dated, sourced event recently made one of those named fears smaller. Not a price bounce, not a feeling: a fact.

When all three lights are on, the product says exactly that — and points you at a fresh dive deep, because that's the moment research pays. It never says buy: cheapness and catalysts are observable; the decision is yours. Most stocks, most days, have one or zero lights on — the checklist's power is how rarely it completes.

The Buy Zone percentile, the 52-week slider, and the trend bar each answer one question — unusually cheap vs its own history? where in the year's price range? are the reported numbers improving? The Story chip runs the synthesis for you and names the pattern (computed from those same signals — no AI, no score):

• Fresh dislocation — near its 52-week low while grading cheap: the classic window. Check the fears.
• Outgrew its price — earnings rose faster than the stock, so it's cheap without being near its lows. A discount created by growth.
• Chronic discount — cheap has been this stock's normal for months; the market wants proof. Watch the fear ledger, not the price.
• Cheap for a reason? — the discount comes with weakening numbers: the value-trap pattern. Dive deep before acting.
• Drifting in — approaching its zone; watching territory, not acting territory.
• Repriced — the discount you were watching closed.

One subtlety worth knowing: the percentile compares today against a rolling 5-year window, so a stock that stays cheap long enough makes cheap its own new normal and drifts out of the deep zone without the price recovering — that is exactly what the “Chronic discount” chip flags.

For every stock we store 60 monthly snapshots of P/E (or dividend yield for Cash Cows, or P/Normalized-E for cyclicals). When the daily gate runs, today's fresh price is divided by cached TTM earnings, and we compute where that value ranks within the 60 historical ones: 0 = bottom of the 5yr range, 100 = top. Everywhere a percentile is DISPLAYED, we normalize it to one scale: p0 = the cheapest the stock has been against its own history, p100 = the priciest — for every stock, whatever metric graded it. (Under the hood, yield-graded stocks are inverted: yield rises as price falls. The grading thresholds below are stated in raw metric terms.)

Cyclical override

Cyclicals use P/Normalized Earnings, then add a guard: if operating margin is at or above its 10-year average, the grade is downgraded one step and flagged om_at_peak. This is the anti-trap — a cheap P/E on peak earnings is exactly when cyclicals look most tempting and most dangerous.

Entry, days in zone, events

When a stock first enters a grade we snapshot its price, P/E (or yield), and the date — that's the entry, and "days in zone" counts from it. Zone transitions (entered / more attractive / less attractive / exited) appear in the activity feed, in each stock's permanent Buy Zone activity log, and in its Notable Movements journal.