Research designed to survive scrutiny.

This section is reserved for white papers and substantial original studies. Every release must state its question, method, evidence, limitations, version, and correction history.

Publication gate

Explicit research question and method
Traceable evidence and reproducible scope
Limitations, conflicts, and correction lineage
Stable owned record suitable for citation

Research library

Consensus & Disagreement

When AI Advisors Agree: Distinguishing Forecast Risk from Forecast Reliability

Seven AI frameworks can agree because an asset is easier to forecast. This outcome study tests agreement against visible risk and a no-change baseline, then distinguishes direction from magnitude.

Methodology & Explainability

When the Horizon Changes the Verdict: A Paired Study of 14,000 Multi-Agent Investment Reports

Across 1,247 matched one-year and five-year consensus pairs covering 377 assets, the exact five-level verdict changed in 56.5% of pairs and the broader direction changed in 50.0%.

Supporting record

Lessons Learned

One searchable ledger for research findings and internal workflow lessons, including evidence, limits, lineage, and consequences.

Open

Methodology

The practical production architecture, governance, model roles, and change history.

Open

Concepts

Evergreen educational articles and theoretical foundations organized by topic.

Open

Insights

Chronological commentary, analysis, opinion, and company updates.

Open