The S&P 500 (plus retained internationals) against each name's own valuation history: the price at which it reaches the cheapest 5% of that history, and the price at which it reaches the richest 3%. One rule a side, no hand-set discounts.
As of 2026-10-04 04:15 EDT · regime BOOM
buy 5% · sell 97% %ile
507 scanned · 506 valued
11 names at a buy — ADBE, TYL, LDOS, UHS, STE, DECK, INTU, AON +3 more.
362 more are richly valued vs their own history. Financials are valued on P/B, cyclical/high-growth names on normalized (median-margin) earnings, and ADRs on home-currency EV/EBIT handed back as a USD price — same percentile rule, different lens. 1 can't be given a rules-based target (earnings inflected too hard even after normalizing, class-split share data, or too little history); the most dislocated of those carry an expiring, thesis-backed hand target instead. A BUY is a signal, not an executed trade — Stage-2 diagnosis and staged entry still apply. 66 would be actionable on price but were vetoed — a value-trap disqualifier fired or the Stage-2 verdict came back uncertain/permanent — not buys (CTVA, CHTR, LULU, FIS, GPN +61 more). 7 cleared on price but missed the quality bar (ZTS, ROL, HBAN, VICI, IT +2 more).
The spread
each name vs its own history — reallocation read, no positions assumedCheapest vs own history destinations
Richest vs own history sources
Capital rotates right column → left column, whoever holds what: a name pinned at the top of its own range funds one pinned at the bottom — but only after Stage-2 says the cheap one is temporary, and only above the sell line + optimistic-DCF ceiling counts as "egregious" on the rich side. The left column shows only buyable ratings — rejected (AVOID) and vetoed (FLAGGED) names stay in the ledger, however cheap they look. Percentile = share of its own fiscal years on file at or below today's multiple.
How the buy price is set. Buy target = the price at the 5% percentile of a name's own EV/EBIT history — the level it has traded at or below only about 5% of its fiscal years on file — priced on today's earnings, so the target rises as the business grows. No hand-set discount: the single knob is that percentile. Financials use the same rule on P/B; erratic-EBIT growth/cyclical names use it on normalized (own median-margin) earnings. A strict owner-earnings DCF is shown as a deeper "floor" for scaling in — but only when it actually sits below the buy target (for strong cash generators it lands above and is hidden). FLAGGED = the price reached actionable territory but the buy is vetoed pending a fresh Stage-2 verdict: a deterministic value-trap disqualifier fired (F-Score ≤ 3, Altman Z < 1.8, Beneish M > −2.22, multi-year ROIC/revenue decline, heavy dilution) or the Stage-2 diagnosis could not call the dislocation temporary (uncertain/permanent). A BUY chip therefore always means: cheap on its own history and diagnosed temporary. * No value = no usable lens (extreme-growth, class-split shares, recent IPOs) — priced on dislocation + quality gates.
46 names would exit today — AKAM, GWW, MCHP, PM, NTAP, GRMN, FFIV, FAST +38 more.
Every name in the registry, run through the Stage-5A sell rule as if it were held at today's close — nothing here assumes a position, and no holdings are shown. EXIT means the frozen valuation exit would fire on that name today; it is not a trade and not an instruction to make one. A real sell also needs the position and its lots, a 56-day cooldown per name and reason, and — on the thesis lane only — a current Stage-5B verdict on that specific holding. The sell line is frozen at the 97% percentile and has deliberately no what-if slider.
Stage 2R — is the richness earned? The mirror of the buy side's "is this cheapness temporary?". 81 of 81 names at or above their sell line carry a verdict; the rest are awaiting one, and the count is shown rather than assumed. A verdict is informational: it cannot trigger, veto or delay the frozen exit, and no code that places an order reads it. It answers which of several dear names is the better source of capital — unearned means the multiple expanded without the business doing so, not that anything should be sold. The deterministic reasoning under each row stays primary.
How the sell price is set. Sell line = the price at the 97% percentile of a name's own valuation history — richer than it has traded for all but about 3% of its own history. Reaching it is only the first leg. The exit needs a second, independent one: price must also clear an owner-earnings DCF ceiling built on assumptions chosen to flatter the company, so a name is sold only when it is dear on its own history and dear against a generous estimate of what it can earn. Three fallbacks cover the names with no DCF anchor. Thin history, ADRs, financials and erratic EBIT run on the history leg alone. A company earning nothing must instead clear a record price and a record market cap — a multiple divided by near-zero earnings prints "richest ever" on a business at the start of a recovery, and without the ratchets the simulation sold names there (energy in 2017, GFC survivors in 2010). Names with no EV/EBIT at all use an EV/S line that already has both ratchets baked in. ONE LEG = above the sell line with the corroborating leg missing: reported, never sold on. Thesis breaks and value-trap disqualifiers do not sell mechanically — every tested policy containing mechanical trap-flag exits lost in every configuration — so they are recorded here and routed to Stage 5B for judgment.
Glossary — what the acronyms and terms mean
- P/B Price-to-Book
- Share price vs. the accounting value of the company's assets minus its debts. Used for financials (banks, insurers), whose assets are mostly financial and marked near real value.
- EV/EBIT Enterprise Value / Earnings Before Interest & Taxes
- EV = the cost to buy the whole business (market cap + debt − cash). EBIT = operating profit before interest and tax. Together: how expensive the business is per dollar of operating profit — like a P/E, but debt-aware. This is the default valuation lens here.
- ADR American Depositary Receipt
- A US-traded, US-dollar stand-in for a foreign company's shares (e.g. TSMC, ASML), so it can be bought on a US exchange.
- FX Foreign Exchange
- Currency conversion. For ADRs, the valuation is built in the company's home currency and then FX-converted into US dollars.
- USD US Dollars
- The reporting currency for every target price shown on the board.
- IPO Initial Public Offering
- When a private company first sells shares to the public and starts trading. Recent IPOs lack the years of history this method needs.
- Normalized (median-margin) earnings
- Profit smoothed to the company's typical (median) profit margin, so one freak good or bad year doesn't distort the valuation. Used for erratic-EBIT growth/cyclical names.
- Percentile rule
- Ranking today's valuation against the name's own past — e.g. the buy line is the price at which it's cheaper than ~80% of its own history.
- Cyclical
- A company whose profits rise and fall with the economy (automakers, chipmakers, industrials).
- Spin-off
- A new standalone company split out of an existing one (e.g. GE Vernova from GE) — so it has little independent trading history.
- Sell line
- The mirror of the buy target: the price at which a name reaches the 97th percentile of its own valuation history — richer than it has traded for all but about 3% of its own history. Clearing it is one leg of the sell rule, never the whole of it.
- Owner-earnings DCF ceiling
- What the business is worth if you discount the cash an owner could actually take out of it, under assumptions deliberately chosen to flatter the company. It is the second, independent leg of the sell rule: price above a generous estimate of the company's own worth, not just above its own past.
- Ratchet record price / record market cap
- A parameter-free corroboration used when a company earns nothing, so no DCF ceiling exists. A multiple divided by near-zero earnings prints "richest ever" on a business at the start of a recovery; requiring an all-time-high price and market cap as well is a test that a recovering company cannot pass partway up.
- Stage 5A / Stage 5B
- 5A is the deterministic sell monitor — the two-leg valuation exit plus any thesis or disqualifier warnings. 5B is the judgment layer that diagnoses those warnings on a real holding. Warnings never sell mechanically; only the valuation exit does.
- Stage-2 diagnosis / staged entry
- The follow-up process after a BUY signal: a deeper check (is the cheapness temporary or a real deterioration?) before buying in gradually rather than all at once. A BUY is a signal, not a trade.