joellewis/decision-frameworks

Overview

Decision frameworks provide the structural rigor to make high-quality choices under uncertainty. This skill transforms decision-making from an emotional "gut feel" into a probabilistic process by decoupling outcomes from process, identifying cognitive biases, and utilizing a latticework of mental models to uncover second-order consequences.

Guiding Principles

Principle 1: Decisions are Bets (Source: Duke, Thinking in Bets)

Every decision is a bet on a specific future. Shift from binary "Right/Wrong" thinking to probabilistic "Confidence" thinking. Acknowledging uncertainty (e.g., "I am 70% confident") improves decision quality and reduces the sting of bad outcomes.

Principle 2: Anti-Resulting (Source: Duke, Thinking in Bets)

Do not judge a decision solely by its outcome. A "good" result can come from a "bad" process due to luck. Evaluate the process used at the time of the decision, not the final result.

Principle 3: Latticework of Models (Source: Munger, Poor Charlie's Almanack)

You cannot make wise decisions with a single mental model. You must have a "latticework" of models from multiple disciplines (math, physics, biology, psychology) to avoid the "Man with a Hammer" syndrome.

Principle 4: Circle of Competence (Source: Parrish, Great Mental Models v1)

Know the boundaries of your knowledge. Decisions made outside your circle of competence have a much higher risk of failure. If you must decide outside your circle, seek a partner with the missing expertise.

Principle 5: Pragmatic Truth (Source: Sivers, Useful Not True)

When faced with multiple interpretations of a situation, choose the one that is most useful for making progress, even if it is not "100% true." Belief is a tool for action.

When to Use This Skill

• When a decision involves significant "downside" risk or high uncertainty.
• When choosing between mutually exclusive strategic paths.
• When building a "Decision Journal" for organizational learning.
• When an individual or team is stuck in "Analysis Paralysis."

When NOT to Use This Skill

• For low-stakes, reversible decisions (e.g., "Type 2" decisions). (Source: Bezos, Amazon Letter)
• When speed is the primary requirement and the cost of being wrong is negligible.

Core Process

Step 1: Frame the Decision

Clearly state the choice being made and the desired outcome.

1. The Happiness Test: Will this decision matter in 10 minutes? 10 months? 10 years? Use the result to determine how much time to spend on the decision. (Source: Duke, How to Decide)
2. First Principles: Strip away the analogies and conventional wisdom. What are the fundamental truths of this situation? (Source: Parrish, v1)

Step 2: Build the Decision Matrix

Map the possibilities and probabilities. (Source: Duke, How to Decide)

1. Options: List all possible paths (including "Do Nothing").
2. Outcomes: For each path, list the possible results (Best, Worst, Average).
3. Probabilities: Assign a percentage chance to each outcome.

Step 3: Apply Mental Model Stress-Tests

Use the latticework to uncover hidden flaws. (Source: Munger, Parrish)

1. Inversion: What would cause this decision to be a disaster? How do we avoid that?
2. Second-Order Thinking: Ask "And then what?" for the preferred outcome. What are the long-term consequences?
3. Occam's Razor: If there are two competing solutions, is the simpler one more likely to be correct? (Source: FS Blog)

Step 4: Perform a Pre-Mortem

Imagine the decision has failed 12 months from now. (Source: Duke, How to Decide)

1. Identify the Cause: Why did it fail? (e.g., market shift, execution gap, technical debt).
2. Mitigate: Adjust the current plan to prevent those identified failure modes.

Step 5: Document in the Decision Journal

Record the reasoning at the time to prevent hindsight bias. (Source: Duke, Thinking in Bets)

1. The Context: What did we know at the time?
2. The Reasoning: Why did we choose this path?
3. The Confidence: What was our percentage confidence in the result?

Frameworks & Models

The Decision Matrix (Source: Duke, How to Decide)

A structured table mapping:

• Option A vs. Option B.
• Probabilities of Success/Failure.
• Expected Value (Value * Probability).

Hanlon's Razor (Source: FS Blog)

"Never attribute to malice that which is adequately explained by stupidity (or incompetence)." Use this to prevent emotional bias in interpersonal or competitive decisions.

Map vs. Territory (Source: Parrish, Great Mental Models v1)

The model (the map) is not the reality (the territory). Always check if the "map" you are using for a decision matches the actual "territory" on the ground.

Cross-Skill Invocations

• REQUIRED SUB-SKILL: devils-advocate — To stress-test the preferred option in the matrix.
• RECOMMENDED SUB-SKILL: assumption-audit — To validate the probabilities assigned in Step 2.
• RECOMMENDED SUB-SKILL: mental-model-library — For a broader set of models beyond the core three.

Common Mistakes

1. Resulting: Judging a decision based on its outcome rather than the process. (Source: Duke, Ch. 1)
2. Analysis Paralysis: Spending more time on a decision than the "Happiness Test" justifies. (Source: Duke, Ch. 5)
3. Overconfidence: Stating 100% certainty in an uncertain future. (Source: Duke, Ch. 2)
4. Hindsight Bias: Believing you "knew it all along" after the result is known. (Source: Duke, Ch. 6)

Diagnostic Checklist

• [ ] Have we decoupled the outcome from the decision process?
• [ ] Have we assigned probabilistic confidence scores (e.g., 60%, 80%) to our claims?
• [ ] Have we performed an "Inversion" or "Pre-Mortem" to identify failure paths?
• [ ] Is this decision within our "Circle of Competence"?
• [ ] Have we considered second-order consequences ("And then what?")?

Sources

• Duke, Thinking in Bets, Ch. 1, 2, 6 — Resulting, Confidence, Decision Journal.
• Duke, How to Decide, Ch. 3, 4, 5 — Decision Matrix, Pre-Mortem, Happiness Test.
• Parrish, Great Mental Models v1, Ch. 1, 2, 4 — First Principles, Circle of Competence, Second-Order Thinking.
• Munger, Poor Charlie's Almanack, Ch. 2 — Latticework, Inversion.
• Sivers, Useful Not True — Pragmatic belief as a decision tool.