How does this skill help with project planning?

It forces the use of probability ranges instead of single dates and uses 'pre-mortems' to imagine project failures in advance, surfacing risks that optimism usually hides.

What is the Overconfidence Effect in software development?

It is the systematic tendency to overestimate one's technical abilities and the precision of project timelines, which often leads to missed deadlines and under-budgeted resources.

What is Reference Class Forecasting?

It is a technique where you look at the 'base rate' of success for similar projects in the past rather than assuming your current project is a unique exception to the rule.

Why use this skill with an AI like Claude?

Claude can act as an objective third party to challenge your assumptions, suggest disconfirming evidence, and help you calculate more realistic confidence intervals.

Overconfidence Effect Mitigation

Name: Overconfidence Effect Mitigation
Author: lev-os

bylev-os

0•

生产力与工作流

Mitigates cognitive bias by implementing calibration tracking, probability ranges, and pre-mortem analysis for more accurate project forecasting.

The Overconfidence Effect skill provides a structured framework for Claude to help users counteract the universal tendency to overestimate knowledge and predictive accuracy. By implementing techniques such as reference class forecasting, three-point probability ranges, and systematic pre-mortems, this skill shifts decision-making from 'gut feelings' to evidence-based calibration. It is an essential tool for project managers and developers who need to set realistic expectations, identify hidden risks, and improve the reliability of technical roadmaps.

主要功能

01Disconfirming evidence discovery to challenge optimistic assumptions

02Structured pre-mortem analysis to identify specific failure modes before they occur

03Calibration tracking through prediction logging and accuracy reviews

04Reference class forecasting to align estimates with objective industry base rates

050 GitHub stars

06Probability range estimation using 50%, 75%, and 90% confidence intervals

使用场景

01Assessing technical risks and potential architectural failure points

02Estimating software development timelines and feature launch dates

03Evaluating strategic business decisions and market outcome predictions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lev-os/agents overconfidence-effect

For use in Claude.ai and ChatGPT

Download Skill