Research / experimental
Statistical Research & Audit
Descriptive statistics and audited model output only — no outcome claims.
Out-of-sample audit
Pre-draw records only, scored after the draw. No post-draw fitting.
Baselines
Read from the research API — never hard-coded.
V9 Intelligence
GET /api/research/model-performance → models.v9
V9 Transition Walk-Forward Research
GET /api/v9/wf-results — transition signals evaluated walk-forward. Research only; no signal is promoted on this evidence.
V9 research queue
GET /api/v9/draw-sequence — conditioning source and signal per draw type. Ordering is a research queue, not a performance ranking.
Candidate Engine v3
GET /api/research/model-performance → models.candidate_v3
Descriptive v1
GET /api/research/model-performance → models.descriptive_v1 · historical comparison only
V8 Weighted Ensemble
GET /api/v8/results — archived holdout evaluation. Not promoted.
Experiment counts
Each research family uses its own experiment unit. Counts are reported separately and never summed.