Learn it → Apply it → Save as Skill
Before you commit, assign someone to genuinely attack the plan. A decision that has never been seriously argued against has only been tested once. Structured dissent — a real, resourced effort to find the strongest case against your own plan — surfaces weak assumptions before reality does.
The Dr. Wallace’s Favorite Mental Models is designed to turn useful ways of thinking into reusable tools. Each model gives you a fast visual overview, deeper interactive learning, practical application exercises and an AI-ready Markdown skill you can save for future use.
Instead of merely reading about mental models, you can gradually build your own AI skills library: a collection of portable reasoning skills that can help ChatGPT, Claude, Grok and other AI assistants apply the mental models you value.
Before a consequential decision ships, give someone real standing and real time to build the strongest possible case against it. Treat what survives that attack as the plan. Treat what does not as a warning you got for free.
Build the strongest case FOR the plan
Build the strongest case AGAINST it
Important: an assigned critic who does not really try is weaker than a sincere disagreement. The role only works if it is resourced, respected and allowed to change the outcome.
Most useful when a plan looks settled, confidence is high, and the cost of being wrong in public is far higher than the cost of being challenged in private.
The method works by separating the plan's defenders from a resourced, respected challenger, so the hardest objection gets said out loud before it gets expensive.
Give the challenger role to someone with real standing, enough time, and no incentive to go easy on the plan.
The challenger builds the strongest case against the plan: weak assumptions, missing data, unexamined risks.
The original team answers the strongest version of each objection with evidence, not with dismissal or status.
Decision-makers weigh the plan against what survived the attack, and revise what did not.
Open a use case to see what the model changes.
Fast, early agreement is often a sign the plan was never seriously tested, not that it is correct. Assign a genuine challenger before the plan hardens into a commitment.
A technical red team attempts an authorized intrusion — digital, physical, or social — and reports what actually worked, rather than what the design assumed would hold.
Assign one interviewer to argue the strongest case against the hire using the actual evidence gathered, so early enthusiasm does not quietly become the final answer.
A standing or ad hoc review group builds the strongest counter-argument on the merits, independent of who proposed it or how senior they are.
Require someone to write the detailed case against the investment before the vote, using the same data the bulls are using.
Have a colleague argue the other side's position as persuasively as possible, so your opening position is not built only on your own optimism.
This is an illustrative example designed to teach the method.
Expand for where the practice comes from and where a single assigned critic tends to fall short.
The practice traces to Cold War wargaming in the United States. RAND Corporation simulations for the military used "red" to represent a Soviet adversary and "blue" for US forces; U.S. Secretary of Defense Robert McNamara separately used red and blue teams to evaluate a competing aircraft design. The label later spread into military alternative analysis, cybersecurity, and corporate strategy.
Social psychologist Charlan Nemeth's research is often summarized as skeptical of the token version of this role: a person merely play-acting disagreement does not reliably trigger the same reconsideration that a sincere, genuinely-held dissenting view does, because others can tell the critic does not actually believe the objection. Nemeth points to one practical substitute: some investment firms require analysts to spend real time writing the detailed case against a position they personally favor, which she describes as about as close as you can get to authentic dissent without a real dissenter in the room.
A red team is typically a standing or well-resourced group operating under its own rules of engagement, actively trying to break the plan or system rather than simply voicing an objection in a meeting. Israel's intelligence services created a standing contrarian unit, nicknamed Ipcha Mistabra ("the opposite is reasonable"), after the 1973 Yom Kippur War intelligence failure. U.S. military red-team programs and intelligence "alternative analysis" teams work the same way: structured, resourced, and ongoing rather than a single assigned role in one meeting.
A critic who is only going through the motions rarely produces the same rethinking as someone who actually believes the objection.
Assigning the role to the most junior or least powerful person in the room limits how hard they can afford to push.
The goal is a better plan, not a winner. A challenge that becomes personal stops being useful to anyone.
Responding to the easiest form of an objection, rather than its strongest form, defeats the purpose of raising it.
A single pass right before sign-off catches far less than challenge built into the planning cycle itself.
This model is adversarial testing of one plan, not generative idea collection. Mixing the two produces neither a sharper plan nor more ideas.
the decision is low-stakes, cheap to reverse, and speed matters more than certainty.
you want the team itself to generate failure modes before committing, rather than assign an external challenger.
the task is procedural with a known correct answer, not a judgment call worth arguing about.
This browser tool structures your thinking. It does not send your information anywhere or perform AI analysis.
Save as Skill converts the model into portable Markdown instructions that teach an AI when and how to run a devil's-advocacy or red-team pass on a plan, what questions to ask, what evidence to seek, what mistakes to avoid and how to structure its analysis.
Save the skills you find useful and organize them into your own reusable library. The Markdown is intentionally model-neutral, so it can be used wherever an AI system accepts uploaded files, project knowledge, custom instructions or reusable text context. Platform capabilities differ, so this is a portable skill file rather than a claim of one-click installation.
In cybersecurity, a red team tries to compromise a system the way an attacker would, while a blue team defends it. A "purple team" is a temporary pairing of the two so lessons transfer directly into better defenses, rather than staying siloed in a report.
Standing structures outlast any one strong-willed critic leaving the room. Three separate review teams examined the intelligence behind the raid that killed Osama bin Laden before the operation went ahead — alternative analysis as an ongoing institutional habit, not a one-time favor from a brave colleague.
Partly. Writing out the strongest case against your own favored option, in real detail, captures some of the value even alone. What is harder to replace is the social pressure of someone else's genuine disagreement, which research suggests does more to change minds than content delivered without real disagreement behind it.