Devil's Advocate

Overview

The devil's advocate is a structured dissent role: argue against a proposal as strongly as possible to expose weaknesses before reality does.

It's not negativity — it's disciplined critical thinking with rules: steelman first, attack the idea not the person, and end with a verdict.

Best used before big commitments: strategies, hires, purchases, launches, arguments you're about to make publicly.

When to use

• Stress-testing a business strategy or plan
• Evaluating a major decision (hire, purchase, pivot)
• Preparing for objections to a proposal or argument
• Team decisions suffering from groupthink
• Personal choices where you might be rationalizing

Core concepts

• Steelman first. State the proposal in its strongest form before attacking. Attacking a strawman proves nothing and insults everyone.
• Separate idea from person. The role attacks the proposal, never the proposer. Establish this explicitly or the exercise becomes political.
• Assigned dissent. Rotate an explicit devil's advocate role. Voluntary dissent gets punished socially; assigned dissent is safe and expected.
• Premortem. 'It's a year later and this failed spectacularly. Why?' Working backward from failure surfaces risks optimism hides.
• Inversion. 'What would guarantee this fails?' List failure modes, then check which are present. Charlie Munger's razor for plans.
• Base rates. 'How often do things like this succeed?' Replace inside-view optimism with outside-view statistics before deciding.
• Disconfirming evidence. Actively seek the strongest counter-evidence and counter-arguments. Confirmation bias is the default; counter-search is the discipline.
• Verdict, not just critique. End with: proceed, proceed with mitigations, or kill. Critique without a verdict is just anxiety with extra steps.

Practical workflow

1. State the proposal crisply. One paragraph: what, why, expected outcome. The advocate and everyone else work from the same text.
2. Assign the role explicitly. 'You are the devil's advocate for this decision.' Explicit assignment makes dissent safe and thorough.
3. Steelman it. Advocate restates the strongest case first. Ensures the critique targets the real proposal.
4. Run the attacks. Premortem + inversion + base rates + disconfirming evidence. Write down every weakness found, ranked by severity.
5. Proposer responds. The proposal's champion answers each critique: mitigation, acceptance, or rebuttal with evidence.
6. Render the verdict. Proceed / proceed with conditions / pivot / kill. The advocate gives a clear recommendation with reasoning.
7. Record the reasoning. Document the critique and verdict. Future you will want to know why — especially if it fails.
8. Rotate the role. Next time, someone else advocates. Permanent critics get tuned out; rotation keeps it fresh and fair.

Common pitfalls

• Strawmanning. Attacking a weak version of the proposal. Steelman first or the exercise is theater.
• Personal attacks. Critique sliding into 'this is a dumb idea, you're dumb.' Idea-person separation must be explicit and enforced.
• Permanent devil. One person always dissenting gets dismissed as 'just negative.' Rotate the role.
• Critique without verdict. Listing risks with no recommendation. End with proceed/pivot/kill — decisions need conclusions.
• Only pre-decision. Never reviewing after outcomes. Postmortems ('were the advocate's warnings right?') train better advocates.
• Groupthink politeness. Nobody wanting to be the dissenter. Assigned roles + leader explicitly inviting critique fix this.
• Advocate as formality. Going through motions to rubber-stamp. If the advocate can't find real weaknesses, the proposal may be fine — or the advocate is phoning it in.
• Ignoring base rates. All inside-view argument, no outside statistics. 'Our plan is different' is usually optimism, not insight.