Calibration
7 calibrated stages
Each gate has explicit pass / caution / fail criteria — the same questions, asked the same way, every time. Mood doesn't get a vote.
LI — AI Systems Studio / 02 · Investment Pipeline /Educational · not financial advice Fairfax, VA · EN · 中文
Judgment as a system.
A 7-stage decision-support pipeline that encodes investing frameworks — macro → technicals → valuation → risk → a structured report. This is the rigor, not the recommendation.
Educational framework illustration — not financial advice. The report below is a redacted sample run through a public-redaction tool. All real holdings, position sizing, and P&L have been removed. It demonstrates the structure and discipline of the process, never a recommendation to buy or sell any security.
Proof exhibit · the signature
A sample report, redacted.
Scroll the report — each of the seven stages reveals in sequence, the way the pipeline runs them. Sensitive figures are blacked out by design.
Tessera Terminal NFLX_PUBLIC_20260529.json · v4.121 Redacted
Pipeline v4.121·2026-05-29·public report
NFLXNetflix, Inc. — illustrative
Watch Conviction · ▮▮ redacted
A high-quality compounder priced for perfection. The framework's verdict is process, not position — valuation stage gates further sizing until the envelope resets.
Pipeline gatesWATCH
01Sentiment & MacroPass
02Carry & SystemicPass
03Stock ClassificationPass
04Technical StructureCaution
05Valuation EnvelopeCaution
06Risk & SizingRedacted
07Action & ExitRedacted
01
Sentiment & Macro
PassFramework question
Is the macro regime a tailwind or a headwind for this kind of asset right now?
Method
Classify the regime (growth / inflation / liquidity), then check whether the thesis depends on the regime holding. A thesis that needs a specific macro to be true is a weaker thesis.
02
Carry & Systemic
PassFramework question
What breaks if I'm wrong — and does it break quietly or all at once?
Method
Screen for systemic fragility and crowded positioning. Prefer exposures whose failure mode is gradual and observable over ones that gap.
03
Stock Classification
PassFramework question
What kind of business is this — and am I judging it on the right axis?
Method
Bucket as compounder / cyclical / turnaround / optionality. The bucket sets which later gates matter most and which metrics are noise.
04
Technical Structure
CautionFramework question
Is price telling me anything the fundamentals aren't?
Method
Read trend and drawdown context — an external de-rating is different from a broken thesis. Used to time, never to decide.Levels redacted
05
Valuation Envelope
CautionFramework question
What has to be true at this price — and is that a forecast or a hope?
Method
Reverse-DCF the implied expectations and compare to a calibrated band. Above the band, the gate caps new sizing regardless of conviction.Targets redacted
06
Risk & Sizing
RedactedFramework question
How much, given everything above — and what would make me trim?
Method
Size from conviction × envelope × portfolio context, against backtested thresholds. Actual sizing is private and removed from this sample.Sizing redacted
07
Action & Exit
RedactedFramework question
What is the decision, and what pre-commitment governs the next one?
Method
Emit a structured verdict with pre-set triggers for add / trim / exit — so the next decision is made now, calmly, not later in the moment.Triggers redacted
Open the full public-safe report ↗ The real stage-4 artifact, exported by the pipeline's
--public redaction — verified zero-holdings leak. Why trust the process
Discipline, made repeatable.
Evidence
Backtested thresholds
Sizing and valuation bands are set from historical behavior, not gut. Thresholds are reviewed, not improvised mid-decision.
Versioning
Versioned releases
The framework ships like software — v4.121 here. Changes are deliberate and traceable, so a past decision can be read in its own context.