What is base rate neglect?

Base rate neglect is a cognitive bias where individuals ignore general statistical information (the base rate) in favor of specific, vivid information about a particular case, leading to inaccurate probability estimates.

Do I need to be a mathematician to use this skill?

No. While it includes Bayesian formulas for precision, the primary benefit is the structured habit of asking 'what usually happens in these situations?' before diving into case-specific details.

What is a reference class in Base Rate Analysis?

A reference class is the group of similar past events or cases used to determine a baseline frequency. For example, 'B2B SaaS startups' would be a reference class for a new software venture.

How does this skill improve software project planning?

It forces teams to look at the 'outside view'—how often similar projects have succeeded or failed in the past—preventing the planning fallacy where unique details make a team over-optimistic.

Base Rate Analysis

Name: Base Rate Analysis
Author: lev-os

bylev-os

0•

Data Science & ML

Combats decision-making bias by anchoring probability assessments on statistical baseline frequencies before incorporating specific case details.

Base Rate Analysis is a cognitive framework designed to improve decision-making and forecasting accuracy by prioritizing statistical 'base rates' over vivid, individual anecdotes. By identifying a relevant reference class and establishing a prior probability based on historical data, this skill helps users avoid 'base rate neglect'—the common human tendency to overweight unique information while ignoring the broader context. It provides a structured process for identifying reference classes, finding frequency data, and applying Bayesian logic to adjust estimates only when diagnostic evidence is present, making it invaluable for high-stakes decisions in software project planning, hiring, and risk assessment.

Key Features

01Bayesian Probability Anchoring to combat representativeness heuristics

02Reference Class Identification for defining relevant statistical populations

030 GitHub stars

04Formal Bayes Rule Application for precise mathematical forecasting

05Diagnostic Evidence Calibration to weight new information appropriately

06Structured Communication Framework for transparent decision reasoning

Use Cases

01Estimating project success rates by benchmarking against similar historical software rewrites

02Improving hiring accuracy by anchoring candidate potential on team-wide retention baselines

03Evaluating technical risk and fraud detection by calculating the probability of rare events

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lev-os/agents base-rate-analysis

For use in Claude.ai and ChatGPT

Download Skill