01
Read the filings
SEC documents, not headlines. Multi-year cash flows from EDGAR are normalized so one lucky year cannot become a permanent annuity. Incomplete facts stop the pipeline here.
methodology · no black box
The product in one pass.
The declared workflow reads filings, scores durability, values scenarios, lets risk veto and then waits. Coverage belongs to each recorded run, not to a permanent marketing number.
ticker → data → quant → council → risk → verdict
the pipeline
Five gates between a ticker and a verdict.
Unknown data remains unknown. Weak quality remains weak. A cheap-looking company can still be vetoed before a buy reaches the public output.
02
Score the business
Returns on capital, margins, leverage and consistency. Quality comes before cheapness because a deteriorating business can stay cheap all the way to zero.
03
Value it conservatively
Bear, base and bull scenarios. Growth is tied to what the business can actually reinvest, and sector routing keeps the wrong valuation model away from the wrong company.
04
Let risk veto
Accounting forensics, late filings and distress indicators can block a buy outright — regardless of popularity, narrative or apparent cheapness.
05
Wait
The verdict names a price and then, usually, nothing happens for months. Price decides. Doing nothing on purpose is part of the system.
what we never do
- 01No targets from vibes — every number traces to a filing and a formula.
- 02No AI-invented financials — language models debate; deterministic math owns the valuation.
- 03No reward for trading more — activity is not a product metric.
- 04Nobody pays MostlyNo to like a stock — verdicts are never sponsored.
- 05No rewritten history — calls are timestamped and the journal is append-only.
the limits, out loud
We miss winners on purpose.
A conservative model will call many excellent companies too expensive. That is the trade: fewer forced purchases, more explicit refusals and a visible too-hard pile when the math cannot defend the market price.
Challenger models run in public shadow until dated outcomes decide. Strictness is a choice; hiding it would be a lie.
the track-record contract
Timestamp first. Claim later.
The Council keeps recorded calls and later outcomes together instead of rewriting the method around a result.
record source
Project journal
recorded calls
Timestamped
outcome timing
After the stated horizon
The NO-ledger counts every refusal, the disagreement board scores the council against Wall Street, and the graveyard preserves killed signals.
Research & education only — not investment advice. Past results do not predict the future, and neither do we.