[Senate] Backfill falsifiable predictions for hypotheses flagged by quality gates

Goal

Reduce the no_predictions Senate quality gate backlog by adding falsifiable, testable predictions to hypotheses that lack them. Each prediction should be tied to a measurable expected observation so the hypothesis can be evaluated by downstream governance and market mechanisms.

Acceptance Criteria

☐ At least 25 hypotheses from the no_predictions backlog are inspected

☐ Each selected hypothesis receives one or more falsifiable predictions or an explicit exception note

☐ New predictions are inserted through the existing hypothesis prediction pathway and linked to hypothesis IDs

☐ Before/after no_predictions counts are recorded

Approach

Query hypotheses with predictions_count = 0 or no linked hypothesis_predictions rows.

Prioritize active, evidence-backed hypotheses with target genes or pathways.

Add concise predictions with measurable endpoints, expected direction, and evaluation context.

Verify prediction links and refresh/record the quality gate count.

Dependencies

6da303b5-62fc-4c88-9aff-9345784b5e8a - Failed quality gate triage
de85e156-b7e3-4513-a586-4f3762a824f7 - Hypothesis predictions table

Dependents

Senate rigor scoring
Exchange market resolvability

Work Log

2026-04-21 - Created from quality gate triage

Triage source: no_predictions had 44 fail rows in quality_gate_results; latest system-check payload reported 122 current affected hypotheses.

File: qg_no_predictions_backfill_spec.md

Modified: 2026-04-28 03:24

Size: 1.6 KB