Strativest Capital
An independent research desk. We study companies and markets in depth, publish everything we find in full and for free, write our predictions down before events happen, and publish the score afterwards — hits and misses together.
What we do
Five research teams each watch one question: where the next supply-chain bottleneck forms, whether crash warning signs are flashing, what the market is thinking, which stocks are coiling before a move, and what is really moving the names under watch. Their work appears here as reports and as formal numbered studies (SIP-2026-001, SIP-2026-002, and so on).
How we work, in plain words
- Best sources first. Company filings and official data before commentary. Every figure is dated.
- Predictions are written down in advance. No calling it after the fact.
- No quiet edits. A published study never changes in place; if we got something wrong, the correction is published, dated, and labelled.
- The score is published whole. Cherry-picking would destroy the point.
- Every report says what would change its mind.
- When we cannot confirm a figure, we say UNVERIFIED — and no conclusion rests on it.
The standing notice
Everything here is research and education — not personalised buy/sell advice. Reports present evidence and track records; they are not recommendations for any individual. See the Disclosures page for the full wording.
The Standard — the rules every report follows
Four study types, and the rules that do not bend. This is what makes a score meaningful: the method is fixed before the results arrive.
One per module — the deep-research study the module is built on: method, historical evidence, playbook, baseline.
Monthly or quarterly per module: what changed, and what the module got right and wrong since the last study.
When something resolves — a 3x call hits or misses, a warning level fires and plays out, or doesn't.
Numbers only. Every open call in the module, marked against outcomes. No narrative rescue.
No silent edits
A published study never changes in place. Corrections go in a dated errata entry on the study and in the index; substantive revisions are a new version, with the change stated.
Figures re-verified at publication
Every number is checked against its source before a study ships. Anything that cannot be re-verified is marked UNVERIFIED in the study — never silently used, never silently dropped.
Predictions are pre-registered
Calls are entered in the module's ledger before outcomes can be known. Studies may only cite ledger entries — there is no rewriting history after the fact.
Primary sources first
Evidence is dated and sourced, verbatim where a quote is the evidence. Conflicting figures are held as UNVERIFIED and kept out of rankings until they resolve.
The track record is published whole
Hits and misses in the same scorecard. Selectivity is the strategy; cherry-picking would destroy the point.
Every claim carries its falsifier
Each study states what would change its mind, its limitations, and the items it held out as unverified — explicitly, in the study.