| name | plausibility-ranking |
| description | SOP: rank candidate explanations by plausibility using multi-dimensional weighted scoring |
| version | 1.0.0 |
| category | hypothesis-formation |
| type | sop |
| campaign | hypothesis-formulation |
| input | Candidate explanation list + relevant evidence (from explanation-generation output) |
| output | Plausibility-ranked list + per-dimension scores + ranking rationale |
| dependencies | {"skills":["subagent-spawning"]} |
Plausibility Ranking
Score candidate explanations on three dimensions — evidence consistency, parsimony, and explanatory scope — then weight and rank them to produce a priority list.
HARD-GATE
Preconditions (all must hold before starting):
1. ≥2 candidate explanations exist (from explanation-generation)
2. Each explanation has mechanism and predictions fields
Not satisfied → stop, return error: at least 2 candidate explanations are needed to rank.
Pipeline
- Precondition check: verify completeness of the candidate explanation list
- Evidence-consistency score (0-10): degree of agreement with known evidence
- Parsimony score (0-10): number of assumptions the explanation requires (fewer = higher score)
- Explanatory-scope score (0-10): how many related phenomena it can explain (not just the current anomaly)
- Weighted ranking: default weights 0.5/0.3/0.2 (evidence/parsimony/scope), adjustable
- Output ranked list + rationale
Output Format
{
"weights": {"evidence": 0.5, "parsimony": 0.3, "scope": 0.2},
"rankings": [
{
"rank": 1,
"explanation_id": "E1",
"statement": "...",
"scores": {
"evidence_consistency"
