home/roleplay/evidence-weighting

Evidence Weighting

GPTClaudeGemini··316 copies·updated 2026-07-14
evidence-weighting.prompt
# 📊 Phase 4 — Evidence Weighting & Confidence Grading

> **Purpose:** Apply a rigorous 4-factor confidence matrix to every claim that survived the adversarial debate. Make uncertainty *visible and traceable*.

---

## 📋 Prerequisites

- ✅ Master Directive loaded
- ✅ Phases 1-3 complete (Grounding → Archetypes → Debate)

---

## 🎯 The 4-Factor Confidence Matrix

Every claim is graded on four orthogonal dimensions:

| Factor | What It Measures |
|---|---|
| **Evidence Strength** | Quality, replicability, and quantity of supporting evidence |
| **Source Quality** | Credibility of the sources (peer-review, primary data, expert consensus) |
| **Recency** | How current the evidence is — for fast-moving domains this matters enormously |
| **Domain Expertise Alignment** | Whether the evidence comes from people who actually work in this domain |

Final confidence label is the **minimum** across the four factors — a single weak dimension drops overall confidence.

---

## 📜 The Prompt (Copy Below)

when to use it

Community prompt sourced from the open-source GitHub repo gouthamlabs/Nexus-Method (MIT). A "Evidence Weighting" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.

tags

roleplaycommunitygeneral

source

gouthamlabs/Nexus-Method · MIT