Learn it → Apply it → Save as Skill
You understand it — until you have to explain it. Most people feel they grasp how everyday things and familiar plans work. That feeling collapses, fast, the moment someone asks for the actual step-by-step mechanism.
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.
You can recognize something, use it daily and talk about it smoothly, and still have no real model of how it works underneath. The gap only shows up when you're forced to produce the mechanism step by step, out loud, for someone who knows nothing.
Before explaining self-rated understanding is high. After explaining — once you try to write the actual mechanism — it drops. The drop size varies by person and topic; what matters is that it drops at all.
The model is most useful whenever someone (including you) claims to understand a mechanism, system or plan, but that claim has never actually been tested by producing the explanation.
The method is a small, repeatable test you can run on yourself or, carefully, invite someone else to run on their own claim.
Rate how well you understand something, on a simple scale, before you try to explain it.
Write or say the actual step-by-step mechanism, out loud, with no hand-waving over any step.
Rate your understanding again, honestly, right after attempting the explanation.
The drop between the two ratings points to exactly which step was never actually understood.
Open a use case to see what the model changes.
Fluent description is not proof of depth. Ask the candidate to walk through what happens, step by step, when a specific input hits the system — including the parts that usually get summarized rather than explained.
Before approving a plan, ask one person to produce the actual causal chain on a whiteboard. The exercise reliably surfaces the step everyone assumed someone else owned.
Write the mechanism you believe drives a metric, step by step, before proposing a fix. If the explanation stalls, the fix is premature — test the mechanism first.
Replace "do you understand?" with "explain it back to me, step by step." The request itself, done kindly, is the diagnostic.
Research on this effect has been studied most in politics: asking people to explain exactly how a policy would work, mechanism by mechanism, tends to moderate overconfident positions more than asking them to justify why they hold them.
Before presenting, draft the explanation cold, without notes. Gaps found in private cost nothing. Gaps found live cost credibility.
This is an illustrative example designed to teach the method.
Expand for the distinction between the original study, a famous related illustration, and two commonly confused neighbors.
In 2002, Yale psychologists Leonid Rozenblit and Frank Keil coined "the illusion of explanatory depth" after asking a small group of undergraduates to rate how well they understood everyday devices and mechanisms, then write a detailed, step-by-step explanation of how each one worked, then rate their understanding again. Ratings were consistently lower after the explanation attempt than before it. The researchers concluded that the act of generating a mechanistic explanation — not just thinking about the topic — is what exposes how little is actually known.
No, and this is the most common mix-up. The bicycle illustration comes from a separate, later study: Rebecca Lawson's 2006 paper "The Science of Cycology," which asked people — including regular cyclists — to complete a simple diagram of a bicycle's frame and chain. Many got the mechanism wrong despite years of daily familiarity. It is a related, independently useful demonstration of the same underlying gap between felt and demonstrated understanding, not a replication of the original Rozenblit–Keil study.
Dunning-Kruger is about self-assessed ability, and is usually framed around people of low-to-moderate competence overestimating their skill relative to others. The illusion of explanatory depth is about self-assessed explanatory knowledge of a mechanism, and research suggests it affects nearly everyone — including genuine experts, though typically less severely in their own specialty.
The curse of knowledge is about forgetting what it was like not to know something once you do know it, which makes your explanations to novices too compressed. The illusion of explanatory depth is about not actually possessing the mechanism in the first place, no matter who you're explaining it to. The two often travel together — a true expert can suffer from the curse of knowledge about parts they genuinely understand, while still having illusory depth about parts they don't.
Smooth, confident talk about a topic is not evidence of a working causal model. Only the step-by-step explanation tests that.
The effect is reduced, not eliminated, for genuine subject-matter experts — and can return the moment they're asked about an adjacent part of their own field.
The point is recalibration, not "gotcha." Run it on yourself first, and frame it gently when used with others.
Reciting correct facts about a topic is not the same as explaining how its parts cause its outcomes. Ask "what happens next, and why," not "what is it called."
Research on political topics found that asking people to justify why they hold a position can backfire and entrench it further. Asking them to explain how the mechanism works is what tends to moderate overconfidence.
Being unable to explain how bicycle gears work is not the same failure as being unable to ride one. Don't use this model to doubt a demonstrated, practiced skill.
the stakes require demonstrated performance, not self-rated understanding — watch the person actually do the task instead of asking them to explain it.
the claim is about procedural or motor competence rather than explanatory knowledge — riding, typing, or driving don't require a verbal mechanism.
time is short and action is urgent — demanding a full mechanistic explanation mid-crisis can stall necessary decisions rather than improve them.
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 apply the illusion of explanatory depth, 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.
Writing a private, step-by-step explanation is usually enough to trigger the recalibration — the audience doesn't have to be real. What matters is forcing the mechanism into explicit, ordered steps rather than letting it stay as a vague felt sense.
Silent reflection lets vague, compressed mental shortcuts stand in for a mechanism without ever being checked. Producing an explicit, ordered explanation forces each step to be stated, which is where the missing links become visible.
No. Confidence and accuracy are correlated in many domains, especially with real practice and feedback. The illusion specifically targets a narrower claim: confidence about possessing a step-by-step causal mechanism, which is a different and more demanding kind of understanding than general competence or familiarity.