Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 

README.md

📊 Rubric

"Measurable criteria translate qualitative debate into quantitative decisions."

MOOLLM K-Lines

K-Line Why Related
adversarial-committee/ What gets scored
evaluator/ Who applies the rubric
scoring/ Generic scoring principles
advertisement/ Objects advertise against criteria
designs/mike-gallaher-ideas.md Mike Gallaher's methodology

Quick Links:

Overview

A RUBRIC defines measurable quality criteria that translate qualitative discussion into quantifiable scores. The committee knows they'll be scored but doesn't see the evaluator's reasoning.

Why Rubrics?

Without explicit criteria:

  • "Good" means different things to different personas
  • Debate has no grounding
  • Decisions can't be compared across time
  • Success is undefined

With rubrics:

  • Criteria are explicit and weighted
  • Each option scores against the same scale
  • Decisions become defensible
  • Learning accumulates

Example Rubric

rubric:
  name: "Client Engagement Evaluation"
  
  criteria:
    resource_efficiency:
      weight: 0.20
      scale: 1-5
      anchors:
        5: "Fits perfectly within current capacity"
        3: "Manageable with some adjustment"
        1: "Would require significant new resources"
        
    risk_level:
      weight: 0.30
      scale: 1-5
      anchors:
        5: "Minimal risk, strong track record"
        3: "Moderate risk, manageable concerns"
        1: "High risk, major red flags"

Credits

Mike Gallaher — RUBRIC-based scoring as bridge between qualitative and quantitative.

See: designs/mike-gallaher-ideas.md