Purpose
The question this method should answer
Use market behavior as dated evidence and keep it separate from the company thesis.
Step-by-step method
Four passes through the evidence
Record the signal date
Write the source, period, calculation, and result before moving to the next step.
Choose a benchmark
Write the source, period, calculation, and result before moving to the next step.
Define the holding window
Write the source, period, calculation, and result before moving to the next step.
Write the failure condition
Write the source, period, calculation, and result before moving to the next step.
Evidence hierarchy
Prefer the source closest to the claim
Reader worksheet
Record the chain of reasoning
| Field | Research note | Provenance |
|---|---|---|
| Research question | ____________________________ | Source URL, filing page, and as-of date |
| Current evidence | ____________________________ | Source URL, filing page, and as-of date |
| Peer or benchmark | ____________________________ | Source URL, filing page, and as-of date |
| Counter-evidence | ____________________________ | Source URL, filing page, and as-of date |
| Next checkpoint | ____________________________ | Source URL, filing page, and as-of date |
| Invalidation rule | ____________________________ | Source URL, filing page, and as-of date |
Failure modes
Ways this method can mislead
- Backfilling an entry after seeing the result
- Ignoring the benchmark
- Turning a confirmation tool into a forecast
- Changing the question after the outcome becomes visible.
- Reporting a raw return without the market or sector benchmark.
Primary sources first
Evidence trail
Documents and market data change. Verify the latest filing and state the observation date before relying on a conclusion.
Practical application
How to complete Watchlist Maintenance Routine
Start with a decision-sized question: What role does the exposure serve, and which benchmark and risk limit make that role testable? A broad theme is not yet a research result. Name the company or peer set, reporting period, unit, accounting basis, and the date on which the answer will be reviewed.
| Research stage | What a complete result includes | Quality check |
|---|---|---|
| 1. Scope | One claim, one decision it informs, the measurement period, and the evidence that could contradict it. | Could two readers interpret the question in the same way? |
| 2. Primary evidence | Use position weights, factor and sector exposures, benchmark returns, drawdowns, correlations across regimes, liquidity, and a written decision log. Record the document, filed or published date, covered period, page or section, and URL. | Is the source close enough to the claim, or is it merely repeating another source? |
| 3. Comparable calculation | Use total returns over identical dates, avoid treating recent correlation as permanent, and separate security research from portfolio sizing. Preserve the raw inputs beside the result. | Would a change in definition, fiscal calendar, currency, or business mix reverse the comparison? |
| 4. Interpretation | Separate reported fact, calculation, management explanation, analyst inference, uncertainty, and evidence that would change the view. | Can another reader reproduce the number without guessing an input? |
Illustrative reasoning
What the finished analysis should sound like
Two different tickers can still express the same economic bet; map shared revenue drivers, rates, currencies, commodities, and valuation sensitivity before calling the portfolio diversified. This is an illustrative reasoning pattern, not a factual statement about a current company. Replace it with dated, sourced company evidence before publication.
Strong answer
States the value, period, definition, calculation, benchmark, uncertainty, and next checkpoint. It identifies both confirming and conflicting evidence and explains why the chosen source is appropriate.
Weak answer
Repeats a management adjective, mixes periods, uses an adjusted value on one side and a reported value on the other, or presents a precise conclusion without exposing the inputs.
Publication test
Five checks before relying on the result
- Definition: write exactly what is included and excluded.
- Period: align dates and distinguish quarterly, year-to-date, trailing, and forward figures.
- Source: link the primary document and retain the page or section reference.
- Counter-evidence: document the strongest reasonable alternative explanation.
- Update rule: name the next filing, event, or threshold that requires a fresh conclusion.
Watchlist Maintenance Routine is useful only when the trail remains understandable after the market outcome is known. Keep the original question and inputs so later review measures the method rather than rewriting the premise.