What is 'resulting' in the context of this skill?

Resulting is the tendency to judge the quality of a decision based on its outcome rather than the quality of the process used to make it. This skill helps you separate the two so you can learn from every 'bet'.

How does this skill calculate Expected Value (EV)?

It uses the formula EV = (Probability of Win × Win Amount) - (Probability of Loss × Loss Amount) to help you choose the path with the highest long-term value, even if it feels riskier.

Can I use this for software architecture choices?

Yes, it is highly effective for technical decisions where outcomes are uncertain, helping you quantify the risks of different tech stacks or patterns before committing resources.

What is a Bayesian update in this framework?

A Bayesian update is the process of adjusting your initial probability 'bet' as soon as you receive new evidence, ensuring your strategy evolves with the project's reality.

Thinking in Bets Framework

Name: Thinking in Bets Framework
Author: lev-os

bylev-os

0•

Productivity & Workflow

Implements a probabilistic decision-making framework to separate process quality from outcome bias using Annie Duke's principles.

The Thinking in Bets skill equips Claude with a rigorous framework for navigating decision-making under uncertainty. Inspired by Annie Duke’s poker-derived strategies, it transforms binary 'right or wrong' judgments into probabilistic assessments. By helping users calculate Expected Value (EV), perform Bayesian updates as new information arrives, and differentiate between decision quality and outcome quality, this skill enables teams to avoid 'resulting'—the trap of judging a past choice solely by its result. It is an essential tool for high-stakes strategic planning, technical architecture pivots, and building a culture that values sound reasoning over lucky outcomes.

Key Features

010 GitHub stars

02Identifies and corrects cognitive biases like hindsight bias and overconfidence

03Analyzes decisions using a 2x2 quality/outcome matrix to combat 'resulting'

04Quantifies beliefs as probabilistic odds rather than binary certainties

05Facilitates Bayesian updating to refine confidence as new data emerges

06Calculates Expected Value (EV) to compare strategic options mathematically

Use Cases

01Conducting post-mortems to distinguish between poor process and simple bad luck

02Evaluating high-stakes technical architecture or hiring decisions under uncertainty

03Calibrating team confidence levels for product roadmap priorities and feature bets

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lev-os/agents thinking-in-bets

For use in Claude.ai and ChatGPT

Download Skill