Market View
A quick view of market direction and environment derived deterministically from BriefRooms data. The LLM neither sets nor overwrites the result.
Loading data…
History will be fetched only after expansion.
Research · shadow only
BriefRooms LAB
BriefRooms' central research laboratory — one place for monitoring forecasts, Evidence Patterns, Strategy Research, BRACE-SPX, the Experiment Registry and Experience Store. Research engines remain isolated. Verified changes may reach production only through the BriefRooms Evolution Controller and the component owner's promotion gate.
DATA-DRIVEN
Frozen forecasts
Real prospective forecasts from the internal Belief Core, including target, outcome and individual Brier Score.
Loading data…
Hypothesis history — last 30 daysPublic rolling 30D archive. Older records remain only in the internal analytical history.
Research status
The model collects outcomes without affecting production.
BELIEF CORE · SHADOW EVALUATION
Results & calibration
Metrics calculated only from forecasts frozen before the outcome.
Calibration Curve
Forecast probability versus observed outcome frequency. The diagonal represents perfect calibration.
Breakdown
Belief Core quality by instrument, horizon and individual belief.
Rolling Brier + ECE
Quality stability across successive groups of 50 resolved forecasts.
Belief Core v3 Candidate Library
11 Belief Core v3 candidates: 4 are wired into prospective SHADOW collection and 7 are waiting for required real data sources. Quality review begins at n≥50 resolved forecasts; an incremental-information test is then required. Candidates do not affect Belief Core v2 or production before passing prospective OOS and the central Evolution Gate.
Loading candidates…
Evolution Controller · Closed Loop
Shared BriefRooms lifecycle: history → hypothesis → challenger → prospective OOS → promotion gate → production version → monitoring → rollback / retirement.
Loading closed-loop status…
Evidence Patterns
Automatic discovery of Evidence combinations that prospectively relate to later outcomes and pass a separate holdout/OOS test. Selection uses discovery data only; later holdout/OOS data do not participate in pattern selection.
Loading patterns…
Loading…
Loading Strategy Research…
Loading BRACE-SPX…
Loading FSE…
Loading Shadow Engines…
Loading Registry & Experience Store…
BriefRooms research environment. Results are not investment recommendations. Decision LAB does not execute trades. Evidence Patterns describe statistical relationships and do not establish causality.