experiment-rigor

What problem does it solve?

It prevents Limina’s research from producing ambiguous, non-decisive, or invalid conclusions by enforcing method validity, fair baselines, decisive thresholds, and careful interpretation of negative results.

Core Features & Use Cases

• Decision-grade H/E/F design: Structures hypotheses, experiments, and findings around a specific adoption/rejection decision, with thresholds and guardrails.
• Method-validity gating: Forces explicit checks that the experiment exposes the method’s real intended capability and that a negative result is not due to an invalid setup.
• Comparative, controlled experiments: Requires named baselines, controlled variables, metrics (primary + guardrails), and a trial plan with stopping/budget rules.
• Canonical evidence management: Keeps narrative in H/E/F while storing raw per-run metrics under kb/research/data/ for traceability across sessions.
• Negative-result classification: Separates invalid tests, implementation failures, insufficient signal, trade-off failures, and true negatives before concluding REJECTED.

Quick Start

Use experiment-rigor to design or review a hypothesis or experiment by writing a one-sentence decision, specifying the method-valid setup, and setting confirm/reject/inconclusive rules with raw metrics stored under kb/research/data/.