ThinkingKit

AI Thinking Advisor

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Unlike the standard Toolkit Matcher which uses keyword matching, the AI Thinking Advisor uses Claude (by Anthropic) to understand your situation in natural language and recommend the most relevant mental models with personalised explanations of why each one applies.

Step 1. Describe your situation, challenge, or decision in your own words — as much or as little detail as you like.

Step 2. The AI analyses your situation and matches it against all 150 mental models in the ThinkingKit library.

Decision Matrix Builder

Mon, 01 Jan 0001 00:00:00 +0000

How to use it

The decision matrix is one of the most versatile thinking tools available. It forces you to make your criteria explicit, weight them honestly, and score each option systematically rather than going with your gut.

Step 1. Name your options — the choices you’re deciding between.

Step 2. List your criteria — the factors that matter for this decision. Be specific: “commute time” is better than “convenience.”

Foundations of Better Thinking

Mon, 01 Jan 0001 00:00:00 +0000

This path takes you through the ten most foundational mental models — the ones that provide the greatest leverage across the widest range of situations. Each builds on the previous one. Follow them in order.

The Map Is Not the Territory

Start here. Every mental model is itself a map — a simplification of reality. Understanding this keeps you humble about all the models that follow.

Circle of Competence

Know what you know and what you don't. This model is the foundation of intellectual honesty and good judgment.

Inversion

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Most people approach problems head-on: “How do I get fit?” “How do I make this project succeed?” “How do I build a great team?”

Inversion flips the question. Instead of chasing success, you map out failure — then work backwards.

The mathematician Carl Jacobi captured it in two words: “Invert, always invert.” Charlie Munger made it a cornerstone of his decision-making toolkit, arguing that many hard problems become easier when you reverse them.

First Principles Thinking

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Most reasoning is by analogy: “This is how it’s always been done” or “Company X did it this way, so we should too.” Analogy is efficient but it inherits all the assumptions and limitations of whatever you’re copying.

First principles thinking takes the opposite approach. You decompose a problem into its most basic elements — the things that are undeniably true — and then reassemble your understanding from scratch.

Pre-Mortem Generator

Mon, 01 Jan 0001 00:00:00 +0000

How to use it

A pre-mortem works by harnessing prospective hindsight — research shows that imagining an event has already occurred increases your ability to identify causes by 30%.

Step 1. Name your project, decision, or plan.

Step 2. Imagine it has failed completely. Use the prompts to generate specific failure scenarios — the more concrete, the more useful.

Step 3. Rate each risk by severity. Focus your mitigation effort on the high-severity, high-probability risks.

The Decision-Maker's Toolkit

Mon, 01 Jan 0001 00:00:00 +0000

Every day you make decisions — some small, some life-changing. This path gives you the specific models that improve decision quality most reliably. Follow them in order; each builds on the last.

Probabilistic Thinking

Stop thinking in binary. Assign probabilities to outcomes. This is the foundation of every other decision model.

Confirmation Bias

The single most dangerous bias in decision-making. You must understand it to counteract it.

Bias Check

Mon, 01 Jan 0001 00:00:00 +0000

Why this works

Cognitive biases operate below conscious awareness — you can’t see your own blind spots by definition. A structured checklist forces you to consider each bias explicitly, which research shows reduces their influence on decision-making.

This tool covers eight of the most impactful biases. It’s not exhaustive, but these eight account for the majority of decision-making errors in everyday life.

Second-Order Thinking

Mon, 01 Jan 0001 00:00:00 +0000

How it works

First-order thinking asks: “What happens if I do this?” Second-order thinking asks: “And then what?”

Most people stop at the first question. The ones who consistently make better decisions push further. Howard Marks argues that this is the defining difference between average and superior thinking — not being smarter, but thinking one step further than everyone else.

The practice is straightforward. For any decision, map out the immediate consequences (first-order effects). Then for each of those, ask “and then what?” to identify the second-order effects. You can keep going — third-order, fourth-order — but most of the insight comes from just going one level deeper than your instinct.

Systems Thinking

Mon, 01 Jan 0001 00:00:00 +0000

Most problems aren’t isolated — they’re embedded in systems. This path teaches you to see the loops, dynamics, and leverage points that drive complex situations. Essential for anyone dealing with organisations, markets, ecosystems, or any situation where cause and effect aren’t straightforward.

Feedback Loops

The fundamental building block. Reinforcing loops amplify; balancing loops stabilise. Everything starts here.

Second-Order Thinking

Systems produce cascading effects. Every intervention has consequences beyond the obvious first-order result.

Circle of Competence

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Warren Buffett and Charlie Munger argue that the size of your circle of competence matters far less than knowing its boundaries. A person with a small circle who stays inside it will outperform a person with a large circle who constantly wanders outside it.

The model works on three levels. First, identify what you actually know from direct experience and deep study — not what you’ve read headlines about or heard opinions on. Second, be ruthlessly honest about the boundary. The most dangerous zone is the area just outside your competence, where you know enough to feel confident but not enough to be right. Third, when operating outside your circle, either invest the time to genuinely expand it or find someone whose circle covers what yours doesn’t.

Inversion Workshop

Mon, 01 Jan 0001 00:00:00 +0000

The model behind the tool

Inversion is one of Charlie Munger’s most powerful thinking techniques: instead of asking “how do I succeed?”, ask “what would guarantee failure?” — then avoid those things.

This workshop walks you through the full inversion process step by step, helping you surface risks and failure modes you might not see when thinking forward. When you’re done, you’ll have a concrete list of actions to take.

Model Chains

Mon, 01 Jan 0001 00:00:00 +0000

Individual mental models are useful. Chaining them together is transformative. Each chain below shows a sequence of 3–5 models applied in order to a specific type of challenge — with the output of each step feeding into the next.

The Risk Assessment Chain

Before committing to anything important, run it through this sequence.

Pre-Mortem

Imagine the project has already failed. List every reason why. This surfaces risks your optimism hides.

First Principles Decomposer

Mon, 01 Jan 0001 00:00:00 +0000

The model behind the tool

First Principles Thinking means decomposing a problem into its most basic, undeniable truths — then reasoning up from those truths to create original solutions, rather than copying what already exists.

This decomposer walks you through the full process: surface your assumptions, challenge each one, identify what’s actually true, and rebuild from there.

Related model: First Principles Thinking

The Map Is Not the Territory

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Alfred Korzybski coined this phrase to capture a fundamental truth about human cognition: we never interact with reality directly. We interact with our models of reality — mental maps, frameworks, theories, statistics, categories. These maps are useful precisely because they simplify. A map of London that was as detailed as London itself would be useless.

The danger comes when we forget that our map has edges, gaps, and distortions. Every mental model, every spreadsheet forecast, every stereotype, every strategic plan is a map — a compressed version of a far more complex territory.

Argument Analyzer

Mon, 01 Jan 0001 00:00:00 +0000

How to use it

Good thinking requires being able to evaluate other people’s arguments — and your own — with discipline. This tool walks you through a structured analysis of any claim.

Step 1. Paste or type the argument you want to evaluate.

Step 2. Identify the core claim and the evidence offered for it.

Step 3. Check for common logical fallacies.

Step 4. Assess the strength of the evidence.

Pre-Mortem

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Psychologist Gary Klein developed the pre-mortem as an antidote to two common problems: optimism bias (we tend to overestimate our chances of success) and groupthink (teams suppress dissent, especially after a plan has been agreed upon).

The technique works by giving people explicit permission to think about failure. In a standard post-mortem, you examine what went wrong after the fact. In a pre-mortem, you do it before — when you can still change course.

Calibration Trainer

Mon, 01 Jan 0001 00:00:00 +0000

Why calibration matters

Well-calibrated people — those whose confidence levels match their actual accuracy — make consistently better decisions. When a calibrated person says they’re “90% sure,” they’re right about 90% of the time. Most people are overconfident: their 90% confidence intervals contain the truth only about 50% of the time.

This trainer presents factual questions and asks you to provide a range you’re 90% confident contains the answer. Over multiple questions, it measures whether your confidence matches your accuracy.

Occam's Razor

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Named after the 14th-century friar William of Ockham, this principle states: do not multiply entities beyond necessity. In practice, when two explanations account for the same evidence, the simpler one is more likely to be correct — not because the universe prefers simplicity, but because simpler theories have fewer points of potential failure.

Each additional assumption in an explanation is another place where you could be wrong. A theory that requires five assumptions to work has five chances to fail. A theory that requires two has only two.

Hanlon's Razor

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Humans have a well-documented tendency called the Fundamental Attribution Error: when other people do something that affects us negatively, we tend to assume it reflects their character (“they’re selfish”) rather than their circumstances (“they were overwhelmed and forgot”).

Hanlon’s Razor is the corrective. Before assuming someone acted with bad intent, consider simpler explanations: they didn’t know, they were distracted, they made an honest mistake, they had different information than you, they didn’t think through the consequences, or they were dealing with their own problems.

Thinking Journal

Mon, 01 Jan 0001 00:00:00 +0000

Why keep a thinking journal

The most valuable thinking skill isn’t knowing mental models — it’s knowing which ones you actually use, which ones you neglect, and whether your decisions improve over time. A thinking journal creates that feedback loop.

Each entry captures three things: the decision you faced, the mental models you applied, and (later) what actually happened. Over time, patterns emerge — you’ll see which models you default to, which situations still trip you up, and whether your decision quality is genuinely improving.

Probabilistic Thinking

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Most people think in binary terms: this investment will succeed or fail; this candidate will get the job or won’t; this relationship will work or it won’t. Probabilistic thinking replaces this binary with a spectrum: there’s a 70% chance of success and a 30% chance of failure, and the expected value of the decision depends on both the probabilities and the payoffs.

Three core practices make probabilistic thinking work. First, assign explicit probabilities to outcomes — even rough ones. “About 70% likely” is far more useful than “probably.” Second, update your probabilities when new evidence arrives (Bayesian updating). Third, think in terms of expected value: probability × payoff. A 20% chance at a massive payoff may be worth more than an 80% chance at a small one.

Feedback Loops

Mon, 01 Jan 0001 00:00:00 +0000

How it works

There are two types of feedback loops that drive every system you’ll ever encounter.

Reinforcing (positive) loops amplify change in one direction. Good performance → more confidence → better performance → even more confidence. Or: losing customers → less revenue → worse product → losing more customers. These loops create exponential growth or exponential decline. They feel slow at first, then suddenly overwhelming.

Balancing (negative) loops push a system toward a target or equilibrium. Room gets cold → thermostat turns on heater → room warms up → thermostat turns off heater. These loops create stability. They resist change in either direction.

Margin of Safety

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Engineers designing a bridge don’t build it to hold exactly the maximum expected load. They build it to hold 3–4 times that load. The extra capacity is the margin of safety — the buffer that absorbs surprises, miscalculations, and worst-case scenarios.

Benjamin Graham brought this concept from engineering to investing: only buy a stock when its price is significantly below your estimate of its intrinsic value. The gap between price and value is your margin of safety. If your estimate is wrong by 20%, you’re still okay.

Opportunity Cost

Mon, 01 Jan 0001 00:00:00 +0000

How it works

When you choose to spend an hour on Task A, the opportunity cost isn’t just “one hour” — it’s whatever Task B would have produced in that hour. When you invest $10,000 in a business, the opportunity cost isn’t just $10,000 — it’s what that money would have earned in the stock market, or the education it could have funded, or the emergency fund it could have built.

Antifragility

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Taleb identifies three categories along a spectrum. Fragile things break under stress — a porcelain cup, a rigid plan, a leveraged portfolio. Robust things resist stress — a rock, a diversified portfolio, a stoic person. Antifragile things gain from stress — muscles (grow stronger from exercise), immune systems (develop from exposure), and certain business strategies.

The key insight is that fragility, robustness, and antifragility are properties of the system, not the specific stressor. You can’t predict what will go wrong (Black Swans), but you can build systems that benefit from surprises rather than being destroyed by them.

Thought Experiments

Mon, 01 Jan 0001 00:00:00 +0000

How it works

A thought experiment follows a simple structure. First, define a hypothetical scenario with clear parameters. Second, apply logical reasoning to trace the consequences. Third, see if the conclusions reveal something useful about the real world.

The power is in controlled variation. By changing one variable in an imaginary scenario while holding everything else constant, you can isolate the effect of that variable — something that’s often impossible in the messy real world.

Eisenhower Matrix

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Dwight Eisenhower reportedly said: “What is important is seldom urgent and what is urgent is seldom important.” Stephen Covey popularised this insight as a 2×2 matrix.

Quadrant 1 — Urgent and Important. Crises, deadlines, emergencies. You must do these. But if you live here, you’re constantly firefighting.

Quadrant 2 — Important but Not Urgent. Strategy, relationships, learning, prevention, planning. This is where the highest-leverage work lives. It never screams for attention, so it gets neglected.

Bayesian Updating

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The core practice has three steps. First, start with a prior — your current best estimate of how likely something is, based on everything you know so far. Second, when new evidence arrives, ask: “How likely would I be to see this evidence if my belief were true? How likely if it were false?” Third, update your prior proportionally. Strong evidence should move your belief a lot. Weak evidence should move it a little. Irrelevant evidence shouldn’t move it at all.

Confirmation Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Confirmation bias is the tendency to search for, interpret, favour, and recall information in a way that confirms your pre-existing beliefs. It operates on three levels: you seek different information depending on what you already believe, you interpret ambiguous evidence as supporting your position, and you remember confirming evidence better than disconfirming evidence.

The bias is strongest when the topic is emotionally charged or tied to your identity. Political beliefs, career choices, and personal relationships are the domains where confirmation bias does the most damage — precisely because they’re the domains where accurate thinking matters most.

Dunning-Kruger Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The Dunning-Kruger effect describes an asymmetry in self-assessment. When you know very little about a domain, you lack the knowledge needed to recognise how little you know. You don’t know what you don’t know. As you gain competence, you begin to see the full scope of the domain and your confidence drops — you now understand how much more there is to learn. With expertise comes a more accurate (and often more modest) self-assessment.

Loss Aversion

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Kahneman and Tversky’s prospect theory demonstrated that humans don’t evaluate outcomes symmetrically. A $100 loss feels roughly twice as painful as a $100 gain feels good. This means we’re not rational expected-value calculators — we’re loss-avoiding machines.

This asymmetry creates predictable decision errors. We hold losing investments too long (selling would crystallise the loss). We avoid beneficial risks (the potential loss looms larger than the potential gain). We overpay for insurance and warranties. We stick with the status quo even when change would be beneficial, because change involves potential losses.

Incentives

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Charlie Munger considers incentives the most powerful force in human behaviour. The principle is simple: people respond to incentives. If you want to understand why someone is doing something, look at what they’re rewarded for doing. If you want to change behaviour, change the incentives.

The model operates on three levels. First, explicit incentives — pay, bonuses, promotions, punishments. These are visible and obvious. Second, implicit incentives — social approval, status, avoiding embarrassment, feeling competent. These are invisible but often more powerful. Third, perverse incentives — situations where the incentive structure rewards the opposite of the desired outcome. These are everywhere and incredibly destructive.

Comparative Advantage

Mon, 01 Jan 0001 00:00:00 +0000

How it works

David Ricardo’s insight is counterintuitive: even if you’re better than someone else at everything, you should still specialise in what you’re relatively best at and let them do the rest. The key word is “relatively.”

If you can produce both analysis and administration better than your colleague, but you’re 10x better at analysis and only 2x better at administration, your comparative advantage is in analysis. Every hour you spend on administration costs you high-value analytical output. You should delegate administration even though you’d do it better — because the opportunity cost of doing it yourself is too high.

Supply and Demand

Mon, 01 Jan 0001 00:00:00 +0000

How it works

When demand for something exceeds supply, its price (or value) rises. When supply exceeds demand, its price falls. This mechanism operates far beyond traditional markets — it applies to jobs, skills, attention, real estate, ideas, and social status.

The model becomes powerful when you use it to explain phenomena that seem mysterious. Why are some highly skilled professionals underpaid? Because the supply of people with that skill exceeds the demand. Why do mediocre products sometimes command premium prices? Because demand exceeds the supply of alternatives in that specific niche.

Leverage Points

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Donella Meadows identified a hierarchy of leverage points in systems, ranked from least to most effective. Most people intervene at low-leverage points (adjusting numbers, tweaking parameters) when the highest-leverage interventions involve changing the rules, goals, or mental models that govern the system.

From lowest to highest leverage: adjusting numbers (tax rates, speed limits), changing the size of buffers (inventory, savings), restructuring information flows (who knows what), changing the rules (laws, incentives, constraints), changing the goals of the system, and changing the paradigm or mental model that generates the system’s goals and rules.

Emergence

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Emergence occurs when a system of simple components following simple rules produces behaviour that couldn’t be predicted by studying the components individually. No single ant knows the colony’s architecture. No single neuron contains a thought. No single trader sets the market price. Yet colonies build sophisticated structures, brains produce consciousness, and markets coordinate the production of millions of goods.

The key insight for practical thinking: you cannot understand emergent phenomena by analysing the parts in isolation. You have to study the interactions, the rules governing those interactions, and the patterns that arise from them.

Natural Selection

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Natural selection requires three conditions: variation (differences exist between entities), selection pressure (the environment favours some variants over others), and inheritance (successful traits are passed on). When all three conditions are present, the system evolves — it adapts to its environment over time without any designer.

This mechanism applies wherever these three conditions are met, which turns out to be nearly everywhere: businesses competing in markets, ideas competing for attention, habits competing for your time, technologies competing for adoption, and strategies competing for results.

Red Queen Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Named after the Red Queen in Lewis Carroll’s Through the Looking-Glass — “It takes all the running you can do, to keep in the same place” — this model describes situations where competing entities must continuously improve just to maintain their relative position.

In biology, predators evolve faster legs, so prey evolves faster legs too. Neither gains lasting advantage; both just run faster. In business, as every company adopts the same “best practices,” those practices stop being advantages and become table stakes.

Critical Mass

Mon, 01 Jan 0001 00:00:00 +0000

How it works

In nuclear physics, critical mass is the minimum amount of fissile material needed to sustain a chain reaction. Below it, the reaction fizzles out. At it, the reaction becomes self-sustaining. Above it, it accelerates rapidly.

This threshold dynamic appears in countless non-physics domains: social movements need a critical mass of supporters before they tip into the mainstream; products need a critical mass of users before network effects kick in; skills need a critical mass of practice before competence feels effortless; knowledge needs a critical mass of understanding before insights start connecting.

Entropy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The second law of thermodynamics states that in any closed system, entropy — disorder — increases over time. It takes energy to create and maintain order. Remove the energy, and things decay toward disorder.

This isn’t just physics. Relationships require ongoing investment or they weaken. Codebases require maintenance or they rot. Skills require practice or they atrophy. Organisations require active management or they bureaucratise. Gardens require weeding or they become overgrown.

Steelmanning

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Steelmanning is the opposite of strawmanning. A strawman attacks the weakest, most distorted version of an argument. A steelman engages with the strongest, most charitable version.

The process: before responding to an argument you disagree with, first reconstruct it in its most persuasive form. Ask yourself: “If I were the smartest, most reasonable person who held this view, how would I defend it?” Present this version back to your interlocutor and ask: “Is this a fair representation of your position?”

Reciprocity

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Reciprocity is one of the most powerful and universal social norms. Across virtually all human cultures, when someone does something for you, you feel obligated to return the favour. This works with gifts, concessions, information, trust, and even hostility.

Robert Cialdini identified reciprocity as the first of his six principles of influence because it’s automatic, powerful, and surprisingly difficult to resist — even when the initial “gift” was uninvited or unwanted.

Regret Minimisation

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Jeff Bezos used this framework to decide whether to leave his Wall Street job and start Amazon. He projected himself to age 80 and asked: “Will I regret not trying this?” The answer was obviously yes — he’d always wonder “what if.” The reverse question: “Will I regret trying and failing?” Almost certainly not — he’d be proud of having tried.

The framework works because it cuts through the noise of short-term anxiety and reframes the decision in terms of what actually matters over a lifetime. Most people overweight the risk of failure in the short term and underweight the risk of regret in the long term.

Marginal Thinking

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Marginal thinking asks: what’s the value of one more unit? Not the total value, not the average value — the incremental value of the next step.

This matters because of diminishing marginal returns. The first hour studying for an exam might boost your grade by 10 points. The tenth hour might boost it by 1 point. The marginal return of each additional hour drops. At some point, the marginal value of another hour studying becomes less than the marginal value of another hour sleeping — and the rational choice switches.

Pareto Principle

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Italian economist Vilfredo Pareto noticed that 80% of Italy’s land was owned by 20% of the population. The same pattern appears across an extraordinary range of domains: 80% of sales come from 20% of customers, 80% of bugs come from 20% of code, 80% of results come from 20% of effort.

The exact ratio doesn’t matter — it’s rarely precisely 80/20. The principle is that inputs and outputs are almost never evenly distributed. A small number of causes produce a disproportionately large share of effects.

Survivorship Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

During World War II, the military examined returning bombers to see where they’d been hit and considered reinforcing those areas. Mathematician Abraham Wald pointed out the flaw: the planes that were hit in other areas never returned. The bullet holes they were studying showed where planes could take damage and survive, not where they were most vulnerable.

This is survivorship bias: concentrating on things that survived a selection process while ignoring those that didn’t, which leads to false conclusions about what causes survival.

Sunk Cost Fallacy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

A sunk cost is any past investment (time, money, effort, emotion) that cannot be recovered regardless of future decisions. The rational approach: ignore sunk costs entirely and evaluate choices based only on future costs and benefits. The human approach: “I’ve already put so much into this, I can’t stop now.”

This fallacy is driven by loss aversion — quitting feels like wasting everything you’ve invested, even when continuing wastes even more. It’s compounded by commitment bias (having publicly committed to something makes it psychologically harder to reverse course) and by narrative — we don’t want to admit that our past decisions were wrong.

Thinking in Bets

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Annie Duke, professional poker player turned decision strategist, argues that most people evaluate decisions by their outcomes. Good outcome? Good decision. Bad outcome? Bad decision. But this confuses two distinct things.

A decision made with 90% odds of success will still fail 10% of the time. That 10% failure doesn’t make it a bad decision — it was a good bet that happened to lose. Conversely, a reckless decision with 10% odds of success will occasionally pay off. That rare win doesn’t make it a good decision.

Via Negativa

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Nassim Taleb draws on an ancient philosophical concept: it’s easier and more reliable to know what’s wrong than what’s right. We can identify what makes us sick more reliably than what makes us healthy. We can identify bad practices more easily than optimal ones.

Via negativa means “the negative way” — improving through subtraction rather than addition. Remove the bad rather than adding the supposedly good. Stop doing harmful things before starting new “beneficial” things.

Anchoring

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The first piece of information you encounter disproportionately shapes all subsequent judgments.

This model helps you recognise situations where this pattern is at play and adjust your thinking accordingly. Understanding it does not make you immune, but it gives you a framework for catching it in action and making better decisions as a result.

Case study: How Steve Jobs anchored the iPad’s price at the 2010 launch

When Steve Jobs unveiled the iPad in January 2010, market analysts and media had been speculating for months that Apple’s tablet would cost around $999 — a number Apple had allowed to circulate without correction.

Availability Heuristic

Mon, 01 Jan 0001 00:00:00 +0000

How it works

We judge the likelihood of events by how easily examples come to mind — not by actual frequency.

Case study: How shark attacks distort beach safety policy

After the 2001 “Summer of the Shark” — a media frenzy triggered by a series of highly publicised shark attacks in Florida — beach attendance dropped significantly and several coastal communities invested millions in shark prevention measures.

Framing Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The same information presented differently leads to different decisions — even when the underlying facts are identical.

Case study: How Amos Tversky changed medical decisions with one sentence

In a landmark 1981 study, Tversky and Kahneman presented doctors with identical treatment statistics framed two different ways. Treatment A: “The one-year survival rate is 90%.” Treatment B: “There is 10% mortality in the first year.” Same information. Same numbers.

Goodhart's Law

Mon, 01 Jan 0001 00:00:00 +0000

How it works

When a measure becomes a target, it ceases to be a good measure.

Case study: How Soviet nail factories optimised for the wrong metric

A famous (possibly apocryphal but illustrative) story from Soviet central planning: when Moscow set production targets for nail factories by weight, the factories produced enormous, unusable nails. When the target was changed to quantity, they produced millions of tiny, useless tacks.

Compounding

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Small consistent gains accumulate into extraordinary results over time — in finance, skills, relationships, and knowledge.

Case study: How Warren Buffett made 99% of his wealth after age 50

Warren Buffett’s net worth at age 30 was $1 million. At 50, it was $67 million. At 60, it was $3.8 billion. At 90, it was over $100 billion. The pattern is staggering: 99.7% of his wealth was accumulated after his 50th birthday. More than 96% was accumulated after his 60th birthday.

Normal Distribution

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Most outcomes cluster around the average, with extreme outcomes being rare — except when they are not.

Case study: How Francis Galton discovered the bell curve at a county fair

In 1906, Francis Galton attended a livestock fair in Plymouth, England, where visitors could pay to guess the weight of an ox. About 800 people entered their estimates. Galton, ever the data collector, gathered the tickets after the contest.

Asymmetric Risk

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Seek decisions where the potential upside vastly outweighs the potential downside — even if success is unlikely.

Case study: How Nassim Taleb structured the trade that defined his career

In the years before the 2008 financial crisis, most investors saw the growing housing market and loaded up on mortgage-backed securities. The expected return was positive. The conventional risk models said the probability of a crash was negligible.

Narrative Fallacy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Humans compulsively construct stories to explain random events, creating false causation and illusory patterns.

Case study: How the media constructed a false narrative around Enron

Before its collapse in 2001, Enron was celebrated as the most innovative company in America. Fortune named it “America’s Most Innovative Company” for six consecutive years. Business journalists constructed a compelling narrative: visionary leadership, revolutionary business model, the future of energy markets.

Circle of Control

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Focus your energy on what you can control, accept what you cannot, and learn to tell the difference.

Case study: How James Stockdale survived seven years as a POW using Stoic philosophy

Vice Admiral James Stockdale was the highest-ranking U.S. military officer in the Hanoi Hilton prisoner-of-war camp during the Vietnam War. He was imprisoned for seven and a half years, tortured repeatedly, and denied any rights.

Local vs Global Optima

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The best option within your current constraints may not be the best option overall — sometimes you need to get worse before you can get better.

Chesterton's Fence

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Before removing something, understand why it was put there in the first place.

Case study: How the removal of wolves nearly destroyed Yellowstone

In the early 1900s, the U.S. government eliminated wolves from Yellowstone National Park. Wolves killed livestock and seemed to serve no useful purpose. The “fence” — the wolf population — was removed without understanding why it existed in the ecosystem.

Cobra Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Well-intentioned interventions often produce the opposite of their intended effect when people respond strategically to new incentives.

Case study: How Hanoi’s rat bounty created a rat-breeding industry

In colonial Hanoi in 1902, the French colonial government was struggling with a rat infestation in the city’s modern sewer system. Their solution: pay citizens a bounty for each rat tail brought to the authorities. The logic seemed sound — incentivise rat killing, reduce the rat population.

10x Thinking

Mon, 01 Jan 0001 00:00:00 +0000

How it works

10x thinking, championed by Google’s Astro Teller, rejects incremental improvement. Instead of asking “how do I make this 10% better?”, you ask “how do I make this 10 times better?” The radical constraint forces you to abandon the existing approach entirely and search for breakthrough solutions.

A 10% improvement lets you keep the current system and optimise within it. A 10x improvement requires you to rethink the fundamental approach. This is why 10x goals can actually be easier than 10% goals — they liberate you from the constraints of the existing paradigm.

About ThinkingKit

Mon, 01 Jan 0001 00:00:00 +0000

The problem

There are hundreds of mental models — powerful thinking frameworks used by the world’s best decision-makers, scientists, and strategists. But most of the resources available online are either encyclopedic lists you’ll never finish reading, paywalled courses, or academic papers that take a PhD to parse.

The result: most people know about mental models but don’t actually use them.

The idea

ThinkingKit exists to close that gap. Every mental model is explained visually, with real-world examples, and paired with interactive tools that let you apply the model to your own decisions right now — not someday after you finish a course.

Abstraction

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Abstraction means hiding complexity behind a simple interface. You drive a car without understanding internal combustion. You use a phone without understanding radio signals. You write code without understanding transistors. Each layer of abstraction lets you work at a higher level by ignoring the details below.

The power of abstraction is that it lets humans manage systems far more complex than any individual could understand in full. The danger is that abstractions leak — sometimes the hidden complexity breaks through the interface, and you need to understand the layer below to fix it.

Accidental vs Essential Complexity

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Fred Brooks distinguished between essential complexity (inherent to the problem) and accidental complexity (created by the solution). Building a bridge has essential complexity — physics, materials, load calculations. But if you use a construction management system so convoluted that half your time is spent fighting the tools rather than building the bridge, that’s accidental complexity.

Most mature systems are drowning in accidental complexity: processes added to fix other processes, tools layered on tools, organisational structures designed around people who left years ago.

Accountability

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Accountability means designing systems where decision-makers face consequences for their decisions — both positive and negative. When consequences are present, decision quality improves. When consequences are absent, decisions optimise for the decision-maker’s comfort, not the outcome.

This applies beyond finance: teachers who face accountability for student outcomes teach differently. Doctors who face consequences for misdiagnosis diagnose more carefully. Managers who share in their team’s failures make different decisions than those who are insulated.

Adverse Selection

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Adverse selection occurs when one party in a transaction has information the other doesn’t, and this asymmetry causes the wrong people to self-select into the transaction. The classic example: people who know they’re high-risk are more likely to buy insurance, driving up costs for everyone.

Unlike moral hazard (which changes behaviour after the transaction), adverse selection distorts who enters the transaction in the first place. The information asymmetry means the population that participates is systematically different from the general population — and usually in ways that disadvantage the less-informed party.

Affect Heuristic

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The affect heuristic means making judgments based on your current emotional response rather than systematic analysis. If something feels scary, you overestimate its risk. If it feels pleasant, you underestimate the costs. Your feelings become a shortcut for analysis — fast but often inaccurate.

Paul Slovic’s research showed that when people have a positive feeling toward an activity (like nuclear power among its supporters), they rate both its benefits as high and its risks as low. When they have a negative feeling, they rate benefits as low and risks as high. The emotional response drives both assessments, even though benefits and risks are logically independent.

Anchoring and Adjustment

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Kahneman and Tversky demonstrated that people make estimates by starting from an initial value (the anchor) and adjusting from it — but the adjustment is almost always insufficient. The anchor disproportionately influences the final estimate, even when it’s completely arbitrary.

In one famous experiment, participants spun a rigged wheel that landed on either 10 or 65, then estimated the percentage of African countries in the United Nations. Those who saw 10 estimated 25%. Those who saw 65 estimated 45%. A random number on a wheel shifted their estimates of a factual question by 20 percentage points.

Antifragility vs Margin of Safety

Mon, 01 Jan 0001 00:00:00 +0000

Antifragility

Some things don’t just survive shocks — they get stronger from them. Position yourself to benefit from disorder.

Read the full model →

Margin of Safety

Build a buffer between what you expect and what you plan for. The world will surprise you.

Read the full model →

What’s the difference?

Margin of Safety is defensive — it builds buffers to survive shocks. Antifragility goes further — it structures systems to actually benefit from shocks. A margin of safety prevents ruin. Antifragility converts volatility into growth.

Antifragility vs Optionality

Mon, 01 Jan 0001 00:00:00 +0000

Antifragility

Some things don’t just survive shocks — they get stronger from them. Position yourself to benefit from disorder.

Read the full model →

Optionality

Create situations with limited downside and unlimited upside. Keep your options open until you have to commit.

Read the full model →

What’s the difference?

Antifragility is the property of systems that gain from disorder. Optionality is the mechanism that often creates antifragility — having the right but not the obligation to act. Optionality is a tool; antifragility is the result of using that tool well.

Asymmetric Information

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Asymmetric information exists when one party in a transaction knows more than the other. George Akerlof’s famous “market for lemons” paper showed that this imbalance can cause entire markets to fail: sellers of good used cars withdraw from the market because buyers, unable to distinguish good cars from bad, only pay lemon prices.

The principle extends far beyond economics. In any negotiation, hiring decision, purchase, or advisory relationship, information asymmetry shapes outcomes. The party with more information has structural advantage.

Bandwagon Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The bandwagon effect is the tendency to adopt beliefs, styles, and behaviours because you see many others doing so. The more people who hold a view or take an action, the more likely others are to follow — not because the view has merit, but because popularity substitutes for independent evaluation.

The effect creates self-reinforcing cycles: as more people join, the behaviour appears more validated, attracting more joiners. This can drive both positive outcomes (adoption of genuinely good ideas) and catastrophic ones (speculative bubbles, mass hysteria, harmful fads).

Barbell Strategy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Nassim Taleb’s barbell strategy means combining extreme safety on one end with small, high-upside speculative bets on the other — while avoiding the “moderate risk” middle that feels safe but actually harbours hidden fragility.

The metaphor is literal: load both ends of the barbell, leave the middle empty. In investing, this might mean 85% in treasury bonds and 15% in venture-stage startups. The worst case is losing 15%. The best case is asymmetric upside from the speculative end. The middle — “balanced” mutual funds — feels prudent but obscures its risk profile.

Base Rate Neglect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Base rate neglect occurs when we focus on specific, vivid information about an individual case while ignoring the general statistical frequency (the base rate). This leads to wildly inaccurate probability estimates.

The classic example: if a disease affects 1 in 10,000 people and a test is 99% accurate, a positive result still means there’s only about a 1% chance you actually have the disease. Most people intuitively estimate 99% — because they focus on the test accuracy and ignore the base rate.

Base Rates and Priors

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Base rates are the background frequency of an event in a population before you consider any specific evidence. Before asking “does this evidence mean X?”, you need to ask “how common is X in general?” Without the base rate, even strong evidence can be misleading.

Bayes’ theorem formalises this: your updated belief should combine the base rate (prior probability) with the strength of the new evidence (likelihood ratio). Most reasoning errors come from ignoring the base rate and focusing only on the specific evidence.

Best Mental Models for Career Decisions

Mon, 01 Jan 0001 00:00:00 +0000

Making career decisions — whether to change jobs, ask for a promotion, start a business, or choose a specialisation — is one of the highest-stakes applications of mental models. These 8 frameworks give you a structured approach to career choices.

Regret Minimization

Project yourself to age 80 and ask which choice you'll regret not taking.

Optionality

Prefer choices that keep doors open over those that close them permanently.

Best Mental Models for Evaluating Arguments

Mon, 01 Jan 0001 00:00:00 +0000

In a world of misinformation, hot takes, and motivated reasoning, the ability to evaluate arguments rigorously is a superpower. These models give you a systematic process for separating strong claims from weak ones.

Steelmanning

Before critiquing an argument, rebuild the strongest possible version of it.

Falsifiability

Check whether the claim can even be proven wrong in principle. If not, it's not saying much.

Base Rate Neglect

Before evaluating specific evidence, check how common the claimed outcome is in general.

Best Mental Models for Investing

Mon, 01 Jan 0001 00:00:00 +0000

The greatest investors — Buffett, Munger, Dalio, Taleb — don’t just pick stocks. They use mental models to think about risk, value, and uncertainty with more discipline than the market. These frameworks form the intellectual foundation of sound investing.

Margin Of Safety

Buy assets for less than they're worth. The gap between price and value is your protection against being wrong.

Circle Of Competence

Only invest in businesses you genuinely understand. The edge comes from depth, not breadth.

Best Mental Models for Leaders and Managers

Mon, 01 Jan 0001 00:00:00 +0000

Leading people requires understanding systems, incentives, cognitive biases, and human nature. These mental models give leaders a framework for building effective teams, making organisational decisions, and avoiding the traps that derail management.

Incentives

People respond to incentive structures, not speeches. Design the incentives and the behaviour follows.

Goodharts Law

When a measure becomes a target, it ceases to be a good measure. Choose metrics carefully.

Dunbar Number

Beyond ~150 people, informal coordination breaks down. Plan organisational structure accordingly.

Best Mental Models for Startup Founders

Mon, 01 Jan 0001 00:00:00 +0000

Starting a company means making high-stakes decisions with incomplete information, over and over. These mental models give founders a structured thinking toolkit for the most consequential choices they’ll face.

First Principles

Break down your market, technology, and business model to fundamental truths rather than copying competitors.

Pre Mortem

Before launching, imagine the startup has failed and work backwards to identify the most likely causes.

Minimum Viable

Find the smallest possible experiment that tests your core assumption before building the full product.

Bottleneck / Theory of Constraints

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Eliyahu Goldratt’s Theory of Constraints states that every system has exactly one constraint — a bottleneck — that determines its maximum throughput. Improving any non-bottleneck part of the system produces zero improvement in overall output. Only improving the bottleneck matters.

The process follows five steps. First, identify the constraint — the step where work piles up. Second, exploit the constraint — get maximum output from it with existing resources. Third, subordinate everything else to the constraint — adjust all other processes to support the bottleneck. Fourth, elevate the constraint — invest to increase its capacity. Fifth, repeat — once you’ve broken this bottleneck, a new one will emerge.

Building a Second Brain

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Tiago Forte’s “Building a Second Brain” methodology treats your note-taking system as an extension of your biological memory. Instead of trying to remember everything, you capture, organise, distill, and express information through an external system — freeing your brain to do what it does best: think, connect, and create.

The core framework is PARA: organise everything into Projects (active work), Areas (ongoing responsibilities), Resources (topics of interest), and Archives (completed or inactive items). The key insight: organise by actionability, not by topic.

Butterfly Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Edward Lorenz discovered that in chaotic systems, infinitesimal differences in initial conditions produce vastly different outcomes. He found that rounding a weather simulation variable from 0.506127 to 0.506 produced completely different weather patterns after a few simulated days. The name comes from his metaphor: a butterfly flapping its wings in Brazil could theoretically set off a tornado in Texas.

The practical implication isn’t that butterflies cause tornadoes. It’s that in complex, nonlinear systems, long-term prediction is fundamentally impossible — not because we lack data, but because the system amplifies tiny uncertainties into enormous differences.

Chesterton's Fence (Revisited)

Mon, 01 Jan 0001 00:00:00 +0000

How it works

G.K. Chesterton proposed a simple rule: before you remove a fence (or any existing rule, institution, or practice), you must first understand why it was put there. If you can’t explain the reason for its existence, you’re not yet qualified to remove it.

The principle applies to any inherited structure — company policies, social norms, legal regulations, architectural features, cultural traditions. The fact that something seems useless or outdated doesn’t mean it is. It may serve a purpose you haven’t identified yet.

Circle of Influence

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Stephen Covey’s Circle of Influence model distinguishes between things you can directly control (your effort, attitude, choices), things you can influence (your team’s morale, your manager’s decisions, your market positioning), and things you can neither control nor influence (the weather, the economy, other people’s feelings, government policy).

Most people spend disproportionate time and emotional energy worrying about things outside their circle of influence. This is worse than unproductive — it’s actively draining, because effort spent on uncontrollable things produces zero results while depleting the energy you could spend on things within your control.

Classroom Guide: Teaching Mental Models

Mon, 01 Jan 0001 00:00:00 +0000

Why teach mental models?

Mental models are the operating system of good thinking. Students who learn explicit thinking frameworks — rather than just content knowledge — develop transferable skills that apply across every subject and situation they’ll encounter.

Teaching mental models builds metacognition (thinking about thinking), transfer (applying knowledge across domains), decision-making skills, and argument evaluation abilities. These are the skills that endure long after specific content knowledge fades.

Getting started: Three approaches

Approach 1: The Weekly Model. Introduce one mental model per week. Spend 10 minutes explaining it, show two examples, and give students a prompt to apply it to something in their own lives. By the end of a semester, students have a toolkit of 15–20 frameworks.

Clustering Illusion

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The clustering illusion is our tendency to see meaningful patterns in random data. Thomas Gilovich showed that basketball fans firmly believe in the “hot hand” — that a player who has made several shots in a row is more likely to make the next one. Statistical analysis of actual shooting data showed no hot hand effect. The “streaks” fans saw were exactly what you’d expect from random sequences.

Complexity Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Complexity bias is the tendency to prefer complex explanations, solutions, and frameworks over simple ones — even when the simple version works just as well or better. We associate complexity with sophistication and simplicity with naivete.

This bias has real costs: complex solutions are harder to implement, maintain, debug, and explain. They have more failure points. They require more resources. And they often don’t perform better than simpler alternatives.

Compounding vs Diminishing Returns

Mon, 01 Jan 0001 00:00:00 +0000

Compounding

Small consistent gains accumulate into extraordinary results over time — in finance, skills, relationships, and knowledge.

Read the full model →

Diminishing Returns

Each additional unit of input produces less additional output. The first hour of practice helps more than the hundredth.

Read the full model →

What’s the difference?

Compounding describes outputs that grow exponentially — each period’s gains build on the last. Diminishing Returns describes inputs with decreasing marginal output — each additional unit of effort produces less result. They’re opposing forces that often operate simultaneously in different dimensions.

Confirmation Bias vs Sunk Cost Fallacy

Mon, 01 Jan 0001 00:00:00 +0000

Confirmation Bias

We instinctively seek out information that confirms what we already believe — and ignore what contradicts it.

Read the full model →

Sunk Cost Fallacy

Money, time, or effort already spent should not influence future decisions — but it almost always does.

Read the full model →

What’s the difference?

Confirmation Bias makes you seek evidence that supports what you already believe. The Sunk Cost Fallacy makes you continue investing in something because of what you’ve already spent. Both keep you locked into bad positions, but through different mechanisms — one distorts your evidence-gathering, the other distorts your cost-benefit analysis.

Correlation vs Causation

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Correlation means two things tend to move together. Causation means one thing actually makes the other happen. These are fundamentally different, but our brains constantly confuse them.

There are three possibilities when two things correlate. A causes B (direct causation). B causes A (reverse causation). Or C causes both A and B (common cause, also called a confounding variable). Correctly distinguishing between these requires controlled experiments, not just observation.

Counterfactual Thinking

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Counterfactual thinking means asking: “What if things had been different?” By imagining alternative histories — different decisions, different circumstances, different timing — you can separate genuine skill from luck, identify the true causes of outcomes, and make better decisions going forward.

There are two types. Upward counterfactuals (“It could have gone better if…”) help identify improvements. Downward counterfactuals (“It could have gone worse if…”) help appreciate what went right and identify hidden risks that happened not to materialise.

Creative Destruction

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Joseph Schumpeter described creative destruction as the engine of capitalism: innovation creates new industries and destroys old ones. The automobile destroyed the horse-and-buggy industry. Digital photography destroyed film. Streaming is destroying cable television. Each wave of destruction creates new value while eliminating old value.

The key insight: this destruction isn’t a bug — it’s a feature. The constant churn of creation and destruction is what drives economic progress. Attempts to prevent creative destruction (protecting dying industries, blocking new entrants) slow progress for everyone.

Curse of Knowledge

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Once you know something, you can’t un-know it — and you can’t accurately imagine what it’s like to not know it. This makes experts systematically terrible at explaining their expertise to beginners. They skip steps that seem “obvious” (because they’ve internalised them), use jargon without noticing (because it’s their normal vocabulary), and misjudge the difficulty of concepts (because they’ve forgotten how hard it was to learn them).

Decoy Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The decoy effect (or “asymmetric dominance effect”) occurs when adding a third option that nobody would choose changes which of the original two options people prefer. The decoy is strategically inferior to one option (making that option look better by comparison) but not clearly inferior to the other.

Dan Ariely demonstrated this with The Economist’s subscription pricing. The options were: web-only ($59), print-only ($125), and print + web ($125). Nobody chose print-only — it was clearly dominated by print + web at the same price. But its presence made print + web look like an incredible deal compared to web-only, dramatically increasing print + web subscriptions.

Devil's Advocate

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The Devil’s Advocate technique means deliberately arguing the opposing position — not because you believe it, but to stress-test the current plan. The name comes from the Catholic Church, which appointed an advocate to argue against a candidate’s sainthood to ensure only truly deserving cases succeeded.

In practice, assign someone (or yourself) the explicit role of finding flaws. This works because social dynamics normally suppress dissent — people don’t want to be the one who disagrees. The Devil’s Advocate role gives them permission and obligation to push back.

Diminishing Returns

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Diminishing returns means that each additional unit of input produces less additional output than the previous one. The first hour of studying produces major learning gains. The tenth consecutive hour produces almost none. The first employee you hire massively increases output. The hundredth employee adds a fraction.

This is one of the most fundamental patterns in economics, biology, and daily life. Almost every positive input — effort, money, time, attention — follows this pattern eventually.

Disagree and Commit

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Jeff Bezos categorises decisions as Type 1 (one-way doors, irreversible) and Type 2 (two-way doors, reversible). Most decisions are Type 2 — if they don’t work out, you can reverse them with minimal cost. But organisations often treat every decision as Type 1, applying heavyweight deliberation to lightweight choices.

The framework prescribes different processes: Type 1 decisions deserve careful analysis, broad input, and senior review. Type 2 decisions should be made quickly by individuals or small teams and reversed if wrong. Applying Type 1 process to Type 2 decisions creates slowness. Applying Type 2 speed to Type 1 decisions creates recklessness.

Distinction Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

When we compare options side by side, we notice and overweight differences that would be imperceptible in isolation. A TV with 4K resolution looks slightly sharper than one with 1080p when they’re side by side in a store. At home, watching alone, you’d never notice the difference. But the side-by-side comparison made you pay $500 more.

Christopher Hsee’s research shows that “joint evaluation” (comparing options simultaneously) leads to different choices than “separate evaluation” (experiencing each option independently) — and the separate evaluation usually better predicts actual satisfaction.

Domain Dependence

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Domain dependence is the tendency to compartmentalise knowledge and fail to transfer insights from one area to another, even when the underlying principle is identical. Someone might understand compound interest in finance but fail to recognise the same exponential growth dynamic in their skills, relationships, or health habits.

Nassim Taleb describes this as one of the most common failures of sophisticated thinkers: they understand a concept perfectly in one context and fail to apply it in another, even when a simple analogy would reveal the connection.

Downloads

Mon, 01 Jan 0001 00:00:00 +0000

Printable Cheat Sheets

Download these free reference sheets — print them, pin them to your wall, or keep them on your phone. Each one condenses a curated set of mental models into a single-page reference.

Foundations

Foundations of Better Thinking

The 10 essential mental models everyone should know. Includes Map vs Territory, First Principles, Inversion, Second-Order Thinking, and more.

Download PDF

Decisions

Decision-Making Toolkit

10 models specifically for making better decisions under uncertainty. Probabilistic Thinking, Bayesian Updating, Pre-Mortem, Regret Minimization, and more.

Dunbar's Number

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Anthropologist Robin Dunbar found that primates with larger neocortices maintain larger social groups, and extrapolated that humans can maintain meaningful relationships with roughly 150 people. Beyond that number, social cohesion requires formal rules, hierarchies, and structures — informal trust alone isn’t enough.

Dunbar’s layers are concentric: about 5 intimate relationships (your closest support network), 15 close friends, 50 good friends, 150 meaningful contacts, 500 acquaintances, and 1,500 people you can recognise. Each layer requires progressively less emotional investment.

Dunning-Kruger Effect vs Circle of Competence

Mon, 01 Jan 0001 00:00:00 +0000

Dunning-Kruger Effect

People with low competence in a domain tend to overestimate their ability, while experts tend to underestimate theirs.

Read the full model →

Circle of Competence

Know what you know, know what you don’t know, and stay honest about the boundary.

Read the full model →

What’s the difference?

Dunning-Kruger describes the psychological phenomenon: people with low competence overestimate their abilities, while experts underestimate theirs. Circle of Competence is the strategic response: define your actual boundaries and operate accordingly. One is a bias to watch for; the other is a framework for action.

Eisenhower Matrix vs Opportunity Cost

Mon, 01 Jan 0001 00:00:00 +0000

Eisenhower Matrix

Separate what’s urgent from what’s important. Most people spend their lives on the wrong quadrant.

Read the full model →

Opportunity Cost

The true cost of anything is whatever you give up to get it — including the next best alternative.

Read the full model →

What’s the difference?

The Eisenhower Matrix categorises tasks by urgency and importance to help you prioritise. Opportunity Cost asks what you give up by choosing one option over another. The Matrix helps you decide what to work on. Opportunity Cost helps you understand the true price of that choice.

Endowment Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Richard Thaler’s research showed that people value things they own more than identical things they don’t own. In a classic experiment, participants who received a coffee mug demanded roughly twice as much to sell it as other participants were willing to pay to buy it. The mug didn’t change. Ownership changed its perceived value.

The endowment effect is closely related to loss aversion: selling something you own feels like a loss, and losses feel roughly twice as painful as equivalent gains feel pleasant. So you demand a higher price to compensate for the “loss” of giving up your possession.

Ergodicity

Mon, 01 Jan 0001 00:00:00 +0000

How it works

A process is ergodic when the average outcome across many parallel attempts equals the average outcome for one person over many sequential attempts. Coin flips are ergodic: flip 1,000 coins once or flip one coin 1,000 times — you’ll get roughly the same average.

But many real-world situations are non-ergodic. Russian roulette has a positive expected value across a group ($5 million payout, 5/6 probability = $4.17 million expected value per round). But for any individual playing repeatedly, the time-average outcome is death. The group average is meaningless for the individual.

Ergodicity vs Expected Value

Mon, 01 Jan 0001 00:00:00 +0000

Ergodicity

The average outcome for a group can be completely different from the typical outcome for an individual over time.

Read the full model →

Expected Value

Multiply each possible outcome by its probability and sum them. The mathematically optimal choice is the one with the highest expected value.

Read the full model →

What’s the difference?

Expected Value gives you the average outcome across many parallel attempts. Ergodicity asks whether that average is meaningful for one person over time. In ergodic systems, EV is a reliable guide. In non-ergodic systems (where ruin is possible), EV can be fatally misleading.

Expected Value

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Expected value is the probability-weighted average of all possible outcomes. For each possible outcome, multiply its value by its probability, then sum. The result tells you what a decision is “worth” on average over many repetitions.

A bet that pays $100 with 30% probability and loses $20 with 70% probability has an expected value of (0.30 × $100) + (0.70 × -$20) = $30 - $14 = $16. Over many repetitions, you’d average $16 per bet. Positive expected value decisions are worth taking repeatedly; negative expected value decisions are not.

Falsifiability

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Karl Popper argued that the defining feature of a scientific claim is not that it can be proven true, but that it can be proven false. A theory that accommodates every possible observation — that can never be contradicted by evidence — isn’t strong. It’s empty.

Falsifiability is a filter for useful claims. “This medicine will cure your headache within one hour” is falsifiable — you can test it and observe whether it works. “Everything happens for a reason” is unfalsifiable — no possible event could contradict it, which means it makes no actual prediction.

Fat Tails

Mon, 01 Jan 0001 00:00:00 +0000

How it works

In a normal (bell curve) distribution, extreme events are vanishingly rare. In a fat-tailed distribution, extreme events are far more common than the bell curve predicts. The “tails” of the distribution — where the rare, extreme outcomes live — are “fat” rather than thin.

Nassim Taleb argues that most consequential real-world phenomena follow fat-tailed distributions: financial returns, earthquake magnitudes, pandemic severity, war casualties, book sales, startup valuations. In these domains, using normal distribution assumptions dramatically underestimates the frequency and severity of extreme events.

Feedback Loops vs Second-Order Thinking

Mon, 01 Jan 0001 00:00:00 +0000

Feedback Loops

Every system has outputs that feed back into inputs — reinforcing or balancing the system’s behaviour over time.

Read the full model →

Second-Order Thinking

Consider not just the immediate consequences of a decision, but the consequences of those consequences.

Read the full model →

What’s the difference?

Feedback Loops describe the structural mechanism by which systems amplify or stabilise change over time. Second-Order Thinking is a mental practice of tracing consequences beyond the obvious first effect. Feedback Loops are about system structure. Second-Order Thinking is about cognitive discipline.

Feynman Technique

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Richard Feynman’s learning technique has four steps. First, choose a concept you want to understand. Second, explain it as if you’re teaching it to someone with no background knowledge — use simple language, no jargon, and concrete examples. Third, identify the gaps — the places where your explanation breaks down, where you resort to jargon, or where you can’t find a simple analogy. Fourth, go back to the source material and fill those gaps, then try the explanation again.

First Principles Thinking vs Second-Order Thinking

Mon, 01 Jan 0001 00:00:00 +0000

First Principles Thinking

Break any problem down to its fundamental truths, then build your reasoning up from there.

Read the full model →

Second-Order Thinking

Consider not just the immediate consequences of a decision, but the consequences of those consequences.

Read the full model →

What’s the difference?

First Principles strips a problem down to fundamental truths and rebuilds from scratch. Second-Order Thinking keeps the existing framework but traces consequences further — asking ‘and then what?’ beyond the obvious first effects.

Fitness Landscape

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Imagine all possible strategies as points on a landscape where height represents success. Evolution — and innovation — works by climbing uphill toward peaks. But the landscape has multiple peaks of different heights. If you’re on a local peak, every step in any direction goes downhill, even though a higher peak exists elsewhere.

This is the fundamental tension in both biological evolution and business strategy: the process of local optimisation (climbing the nearest hill) can trap you on a suboptimal peak, unable to reach the global optimum without first descending into a valley.

Gambler's Fallacy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The gambler’s fallacy is the belief that past random events affect the probability of future random events. If a fair coin lands heads five times in a row, many people feel that tails is “due” — that the probability of tails has somehow increased. It hasn’t. The coin has no memory. Each flip is independent. The probability of tails remains exactly 50%.

The fallacy arises because we intuitively expect small samples to look like the overall distribution. Five heads in a row doesn’t “feel” random, so we expect the next flip to correct the pattern. But randomness is under no obligation to look balanced in small samples.

Goodness of Fit

Mon, 01 Jan 0001 00:00:00 +0000

How it works

No single mental model works for every situation. The skill isn’t knowing the most models — it’s knowing which model fits which situation. A hammer is useless for screwing; a screwdriver is useless for hammering. Both are excellent tools. The craft is in the selection.

This meta-model asks: before applying a framework, check whether it’s the right one for the problem at hand. Match the tool to the job, not the job to your favourite tool.

Halo Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The halo effect, identified by Edward Thorndike, occurs when a positive impression in one area unconsciously influences your judgment in unrelated areas. An attractive person seems more intelligent. A confident speaker seems more knowledgeable. A successful entrepreneur’s opinions on unrelated topics carry unearned weight.

The effect works in reverse too (the “horn effect”): a negative impression in one area biases perception of everything else. Someone who makes a bad first impression gets judged as less competent, even when their work quality is identical to someone who made a good first impression.

Hanlon's Razor vs Incentives

Mon, 01 Jan 0001 00:00:00 +0000

Hanlon’s Razor

Never attribute to malice that which can be adequately explained by ignorance, incompetence, or neglect.

Read the full model →

Incentives

Never ask why someone is behaving a certain way until you understand what they’re incentivised to do.

Read the full model →

What’s the difference?

Hanlon’s Razor says don’t assume malice when incompetence explains the behaviour. Incentives analysis goes deeper: don’t assume incompetence when misaligned incentives explain the behaviour. Both replace simplistic character judgments with structural explanations, but at different levels.

Hindsight Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Hindsight bias is the tendency to see past events as having been predictable, even when there was no way to know what would happen. Once you know the outcome, your brain rewrites the story to make it feel inevitable.

This matters because it makes us overconfident in our ability to predict the future, undervalues the role of luck and uncertainty, makes it harder to learn from mistakes (since we believe we “should have known”), and leads to unfair blame.

Homeostasis

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Homeostasis is a system’s tendency to resist change and return to its equilibrium state. Your body maintains a core temperature of 37°C — when you’re too hot, you sweat; when you’re too cold, you shiver. The system actively resists deviations from its set point.

Organisations, markets, and social systems exhibit homeostasis too. A company’s culture resists change even when leadership mandates it. A dieter’s body resists weight loss by lowering metabolism. A team’s performance reverts to its established norm after an unusually good or bad period.

Hyperbolic Discounting

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Humans systematically prefer smaller, sooner rewards over larger, later ones — and the preference isn’t proportional to the delay. The discount rate is hyperbolic, not exponential: we’re much more impatient about immediate versus slightly delayed rewards than about delayed versus even more delayed rewards.

Most people prefer $100 today over $110 tomorrow. But the same people would prefer $110 in 31 days over $100 in 30 days — even though the choice is structurally identical (wait one day for $10 more). The immediacy of “today” creates disproportionate urgency.

Hysteresis

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Hysteresis means that a system’s current state depends on its history, not just its current inputs. Even if you remove the cause of a change, the system doesn’t necessarily return to its original state. The path matters, and some changes leave permanent marks.

A rubber band stretched too far doesn’t return to its original shape. A reputation damaged by scandal doesn’t fully recover when the truth emerges. A relationship strained by betrayal doesn’t return to baseline when an apology is given.

Iatrogenics

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Iatrogenics is harm caused by the healer — damage inflicted by an intervention that was supposed to help. The term comes from medicine (iatrogenic illness = illness caused by medical treatment), but the principle extends to any domain where intervention can backfire.

Nassim Taleb argues that iatrogenics is one of the most underappreciated risks in modern life. We overintervene because we have a bias toward action: doing something feels better than doing nothing, even when doing nothing is the better choice.

Identity-Protective Cognition

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Identity-protective cognition means we evaluate evidence not by its quality but by whether accepting it would threaten the identity of our group. Dan Kahan’s research shows that people with high scientific literacy are actually more polarised on politically charged scientific topics — because they use their intelligence to construct better arguments for their side, not to find the truth.

The mechanism is subtle: we don’t consciously decide to reject evidence. We experience the threatening evidence as less credible, less relevant, or less well-conducted. Our identity filters the evidence before our analytical mind gets to evaluate it.

IKEA Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The IKEA Effect, named by Michael Norton, Daniel Mochon, and Dan Ariely, describes our tendency to overvalue things we’ve had a hand in creating. Participants in their studies valued self-assembled IKEA furniture 63% more than identical pre-assembled furniture. The effort invested created attachment that inflated perceived value.

The effect extends beyond furniture: code you’ve written, meals you’ve cooked, strategies you’ve developed, and ideas you’ve contributed to all receive an ownership premium in your evaluation — regardless of their objective quality.

Incentive-Caused Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Charlie Munger says you should never ask a barber whether you need a haircut. The answer will always be yes — not because the barber is dishonest, but because their livelihood depends on you saying yes. Their incentive structure makes objective advice impossible.

Incentive-caused bias operates unconsciously. The real estate agent genuinely believes now is a great time to buy. The surgeon genuinely believes surgery is the best option. The consultant genuinely believes you need more consulting. Their beliefs are sincere — and systematically distorted by their incentives.

Inversion vs Pre-Mortem

Mon, 01 Jan 0001 00:00:00 +0000

Inversion

Instead of asking how to succeed, ask what would guarantee failure — then avoid it.

Read the full model →

Pre-Mortem

Before starting, imagine the project has already failed. Then figure out why.

Read the full model →

What’s the difference?

Both models examine failure, but from different angles. Inversion asks ‘what would guarantee failure?’ as a general thinking tool applicable to any goal. A Pre-Mortem is a specific technique applied to a concrete plan — imagining the project has already failed and working backwards to identify causes.

Irreversibility

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Some decisions can be reversed cheaply; others can’t. Jeff Bezos calls these “two-way doors” and “one-way doors.” Two-way doors (reversible decisions) deserve fast action — if you’re wrong, you walk back through. One-way doors (irreversible decisions) deserve careful deliberation because there’s no going back.

The practical implication: match your decision-making effort to the reversibility of the decision. Most people apply the same level of deliberation to all decisions, which means they’re too slow on reversible ones and sometimes too fast on irreversible ones.

Know Your Audience

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The same information, delivered identically, lands completely differently depending on who receives it. Effective communication starts with understanding the listener — their knowledge, motivations, concerns, and decision-making context — before crafting the message.

Herbert Clark’s “audience design” framework shows that skilled communicators continuously model their listener’s mental state and adapt in real time: what do they already know? What jargon will they understand? What do they care about? What will make them act?

Leverage

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Leverage means using a tool, system, or resource to amplify your output beyond what direct effort alone could produce. Naval Ravikant identifies four forms of leverage: labour (other people working for you), capital (money working for you), code (software working for you), and media (content working for you).

The last two — code and media — are what Naval calls “permissionless leverage.” You don’t need anyone’s approval to write software or publish content. A single person with code and media leverage can have the impact of a thousand-person organisation.

Loss Aversion vs Status Quo Bias

Mon, 01 Jan 0001 00:00:00 +0000

Loss Aversion

Losses hurt roughly twice as much as equivalent gains feel good — and this asymmetry distorts nearly every decision you make.

Read the full model →

Status Quo Bias

We prefer the current state of affairs simply because it’s familiar — even when alternatives are objectively better.

Read the full model →

What’s the difference?

Loss Aversion is the psychological tendency to feel losses roughly twice as strongly as equivalent gains. Status Quo Bias is the preference for the current state of affairs. They’re related but distinct: Loss Aversion explains why change feels threatening (potential loss outweighs potential gain). Status Quo Bias describes the resulting behaviour (defaulting to the current state).

Market for Lemons

Mon, 01 Jan 0001 00:00:00 +0000

How it works

George Akerlof showed that when buyers can’t verify quality before purchase, high-quality sellers withdraw from the market because they can’t get fair prices. Buyers, knowing this, lower their willingness to pay — driving out more quality sellers. The market degrades until only “lemons” remain.

The solution requires quality signals that are costly or difficult to fake: warranties, certifications, brand reputation, return policies, third-party reviews. These mechanisms reduce information asymmetry and allow quality sellers to differentiate themselves.

Mental Accounting

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Mental accounting, identified by Richard Thaler, describes how people create separate mental “accounts” for their money and treat funds differently depending on which account they’re in. A bonus feels like “free money” and gets spent more freely than salary. A tax refund feels like a windfall even though it’s money you overpaid. Casino winnings feel like “house money” and get risked more aggressively.

Economically, this is irrational — money is fungible, meaning a dollar is a dollar regardless of its source. But psychologically, we consistently violate this principle.

Mere Exposure Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Robert Zajonc discovered that simply being exposed to something repeatedly makes us like it more — even when we’re not consciously aware of the exposure. This applies to faces, words, melodies, brands, ideas, and foods. Familiarity breeds not contempt, but preference.

The effect is powerful because it operates unconsciously. You don’t decide to like something more because you’ve seen it before. Your brain simply processes familiar stimuli more fluently, and it misinterprets that processing ease as genuine preference.

Mimetic Desire

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Rene Girard’s theory of mimetic desire proposes that we don’t generate our desires independently — we borrow them from other people. We want things because we see others wanting them. The model (the person we’re imitating) isn’t just influencing our behaviour; they’re generating our desires.

This explains why desire is often competitive and conflictual. When two people want the same thing because they’re imitating each other’s desire, they become rivals — even if the object itself isn’t scarce. The rivalry, not the object, becomes the focus.

Mimetic Desire vs Bandwagon Effect

Mon, 01 Jan 0001 00:00:00 +0000

Mimetic Desire

We don’t want things independently — we want things because other people want them. Desire is borrowed, not original.

Read the full model →

Bandwagon Effect

The tendency to adopt beliefs and behaviours simply because many other people do.

Read the full model →

What’s the difference?

Mimetic Desire (Girard) says we borrow our desires from models we admire or compete with — we want things because specific people want them. The Bandwagon Effect says we adopt behaviours because many people are doing them — we follow the crowd. One is about imitating specific individuals; the other is about following masses.

Minimum Effective Dose

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The Minimum Effective Dose (MED) is the smallest input that produces the desired output. Water boils at 100°C — additional heat after that point doesn’t make it “more boiled.” It’s wasted energy.

Tim Ferriss popularised this concept: for any desired outcome, there’s a threshold of input below which nothing happens and above which additional input produces diminishing or zero returns. The skill is finding that threshold and stopping there.

Moral Hazard

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Moral hazard occurs when someone takes more risk because they don’t bear the full consequences of that risk. Insurance is the classic case: once you’re insured against theft, you might be less careful about locking your door. The insurance doesn’t make you dishonest — it changes your incentive structure.

The pattern appears whenever risk and consequences are separated. Employees spending company money are less careful than spending their own. Governments bailing out banks create moral hazard — banks take bigger risks knowing they’ll be rescued if things go wrong.

Naive Realism

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Naive realism is the belief that you perceive reality objectively and without bias, and therefore anyone who disagrees with you must be uninformed, irrational, or biased. It’s the mother of all cognitive biases because it makes you blind to all your other biases.

The belief has three tenets. First: “I see the world as it really is.” Second: “If others had access to the same information, they’d agree with me.” Third: “If they still disagree, they must be biased, stupid, or acting in bad faith.”

Negativity Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Humans are wired to give more weight to negative information than to positive. One insult outweighs ten compliments. One bad review drives away more customers than ten good reviews attract. One traumatic experience can override years of positive ones. This asymmetry served our ancestors — noticing threats was more survival-critical than noticing opportunities.

John Gottman’s research on marriages found that stable relationships require a ratio of roughly 5 positive interactions for every 1 negative interaction. Below that ratio, the relationship deteriorates — because the negative interactions carry more weight.

Network Effects

Mon, 01 Jan 0001 00:00:00 +0000

How it works

A network effect occurs when a product or service becomes more valuable to each user as the total number of users grows. A telephone is useless if you’re the only person who has one. With two people, it has one possible connection. With ten people, it has 45. With a million, it has nearly 500 billion.

Network effects create powerful feedback loops: more users → more value → more users → more value. This often produces winner-take-all markets where the largest network captures nearly all the value, even if competitors have superior technology.

Network Effects vs Critical Mass

Mon, 01 Jan 0001 00:00:00 +0000

Network Effects

Some products become more valuable as more people use them — creating winner-take-all dynamics.

Read the full model →

Critical Mass

Some processes need a minimum threshold of input before anything happens — then they suddenly accelerate.

Read the full model →

What’s the difference?

Network Effects describe how a product becomes more valuable as more people use it. Critical Mass is the threshold at which the network effect becomes self-sustaining — where growth generates more growth without external push. Network effects are the mechanism; critical mass is the tipping point.

Niche Construction

Mon, 01 Jan 0001 00:00:00 +0000

How it works

In biology, organisms don’t just adapt to their environments — they modify their environments to better suit themselves. Beavers build dams. Humans build cities. This is niche construction, and it’s one of the most powerful strategic principles available.

Applied to life and work, niche construction means proactively shaping your environment to amplify your strengths rather than passively accepting the conditions you find yourself in. Instead of asking “how do I fit in here?”, ask “how do I reshape this to fit me?”

Occam's Razor vs First Principles Thinking

Mon, 01 Jan 0001 00:00:00 +0000

Occam’s Razor

The simplest explanation that fits the evidence is usually the correct one.

Read the full model →

First Principles Thinking

Break any problem down to its fundamental truths, then build your reasoning up from there.

Read the full model →

What’s the difference?

Occam’s Razor starts with existing explanations and prefers the simplest one. First Principles ignores existing explanations entirely and rebuilds understanding from fundamental truths. Occam’s Razor is a filter for choosing between theories. First Principles is a method for generating new theories.

Optionality

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Optionality means structuring situations so you have the right but not the obligation to take action. Options are asymmetric by nature — you capture the upside if things go well, and your downside is limited to what you paid for the option.

In finance, options are formal contracts. In life, optionality is a way of thinking about decisions. Taking a job at a company with many possible career paths preserves more options than a narrow specialist role at a dead-end company. Learning a versatile skill preserves more options than learning a niche one.

Paradox of Choice

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Barry Schwartz’s research shows that increasing the number of options doesn’t always increase satisfaction. Beyond a threshold, more options lead to decision paralysis (unable to choose), regret (constantly wondering if you chose wrong), and elevated expectations (with so many options, surely the perfect one exists).

The distinction is between “maximisers” (who want the absolute best option) and “satisficers” (who want the first option that meets their criteria). Satisficers make faster decisions and report higher satisfaction, despite spending less effort choosing.

Pareto Principle vs Bottleneck / Theory of Constraints

Mon, 01 Jan 0001 00:00:00 +0000

Pareto Principle

Roughly 80% of effects come from 20% of causes. Find the vital few and ignore the trivial many.

Read the full model →

Bottleneck / Theory of Constraints

Every system has one constraint that limits overall throughput. Improving anything else is waste until you fix the bottleneck.

Read the full model →

What’s the difference?

The Pareto Principle says a small number of inputs produce the majority of outputs. The Theory of Constraints says every system has exactly one bottleneck that limits total throughput. Pareto tells you where the leverage is. Bottleneck tells you where the constraint is. They often point to the same place — but not always.

Path Dependence

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Path dependence means that the outcome depends not just on current conditions but on the specific sequence of past events that led here. The path taken — not just the destination — shapes what’s possible. Different starting points or different sequences of steps can lead to radically different endpoints, even with identical resources and conditions.

The QWERTY keyboard layout is the canonical example. It was designed in the 1870s to prevent typewriter key jams by separating frequently used letter pairs. The mechanical constraint disappeared decades ago, but QWERTY persists because switching costs (retraining billions of typists) exceed the marginal benefit of a better layout. We’re stuck on a path chosen for reasons that no longer exist.

Peak-End Rule

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Daniel Kahneman discovered that when people evaluate past experiences, they don’t calculate the average pleasantness or total duration. They judge the experience based on two moments: the peak (most intense point) and the end. Everything else is largely ignored by memory.

This means a painful medical procedure that ends with a period of gradually decreasing pain is remembered as less painful than a shorter procedure that ends at peak pain — even though the longer procedure involved more total pain. Memory doesn’t track totals. It tracks peaks and endings.

Planning Fallacy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Daniel Kahneman and Amos Tversky identified the planning fallacy: people systematically underestimate the time, cost, and risk of future actions while overestimating their benefits. This happens even when they have direct experience of past projects taking longer than expected.

The mechanism: when planning, we focus on the specific scenario we envision (the “inside view”) rather than the base rate of similar projects (the “outside view”). We imagine everything going right and build our estimate around that ideal scenario.

Planning Fallacy vs Pre-Mortem

Mon, 01 Jan 0001 00:00:00 +0000

Planning Fallacy

We systematically underestimate how long tasks will take, even when we have direct experience of past overruns.

Read the full model →

Pre-Mortem

Before starting, imagine the project has already failed. Then figure out why.

Read the full model →

What’s the difference?

The Planning Fallacy is the bias: we systematically underestimate time, cost, and risk. The Pre-Mortem is the antidote: imagine failure before starting, which surfaces the risks that optimism hides. The fallacy is the disease; the pre-mortem is one of the best treatments.

Power Laws

Mon, 01 Jan 0001 00:00:00 +0000

How it works

A power law distribution is one where a small number of items account for a disproportionate share of the total. Unlike a normal (bell curve) distribution where most values cluster around the average, power law distributions have fat tails — a few extreme values dominate.

This appears everywhere: a tiny fraction of earthquakes cause the vast majority of damage, a few authors sell most of the books, a small number of venture investments produce nearly all the returns, and a handful of your habits drive most of your outcomes.

Premature Optimisation

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Donald Knuth wrote that “premature optimisation is the root of all evil” in programming — but the principle extends far beyond code. Premature optimisation means perfecting something before you’ve confirmed it should exist at all.

The pattern: you spend weeks polishing a feature, process, or plan to peak efficiency, only to discover that the whole thing was solving the wrong problem. The optimisation effort is wasted because the foundation was wrong.

Price Discrimination

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Price discrimination means charging different prices to different customers for the same product, based on their willingness to pay. Airlines do it (business class vs economy for the same flight), software companies do it (student pricing, enterprise pricing), and movie theatres do it (matinee discounts, senior pricing).

For price discrimination to work, the seller needs to segment customers by willingness to pay, prevent resale between segments (a student can’t buy the cheap version and sell it to a business), and have some market power (competitive markets make discrimination harder).

Principal-Agent Problem

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The Principal-Agent Problem occurs whenever one person (the agent) acts on behalf of another (the principal), but their interests aren’t perfectly aligned. The principal wants the agent to act in the principal’s best interest, but the agent has their own goals — and the principal can’t perfectly monitor what the agent does.

This isn’t about bad people. It’s about misaligned structures. A real estate agent technically works for you, but earns commission based on the sale price. A CEO technically works for shareholders, but is motivated by their own compensation, reputation, and job security. A doctor technically works for the patient, but practices within a system that rewards certain treatments over others.

Prisoner's Dilemma

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Two suspects are arrested and interrogated separately. Each can cooperate (stay silent) or defect (betray the other). If both cooperate, both get light sentences. If both defect, both get heavy sentences. If one defects while the other cooperates, the defector goes free and the cooperator gets the heaviest sentence.

The dilemma: regardless of what the other person does, defecting is individually rational. But if both follow this logic, both end up worse off than if they’d cooperated. Individual rationality leads to collective irrationality.

Probabilistic Thinking vs Bayesian Updating

Mon, 01 Jan 0001 00:00:00 +0000

Probabilistic Thinking

Think in likelihoods, not certainties. Assign probabilities to outcomes instead of assuming binary results.

Read the full model →

Bayesian Updating

Start with your best guess, then update it proportionally as new evidence arrives.

Read the full model →

What’s the difference?

Probabilistic Thinking is the broad practice of thinking in likelihoods rather than certainties. Bayesian Updating is the specific method for adjusting those likelihoods as new evidence arrives. Probabilistic Thinking is the philosophy. Bayesian Updating is the mechanics.

Process vs Outcome Thinking

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Process-outcome thinking separates the quality of a decision from the quality of its result. Good decisions can produce bad outcomes (a well-researched investment that loses money due to unforeseeable events). Bad decisions can produce good outcomes (a reckless gamble that happens to pay off).

Over time, good processes produce better outcomes than bad processes. But any single outcome is a noisy signal about process quality. Judging decisions solely by outcomes leads to learning the wrong lessons — reinforcing lucky bad processes and abandoning unlucky good ones.

Reactance

Mon, 01 Jan 0001 00:00:00 +0000

How it works

When people feel their freedom or autonomy is being threatened — told they can’t do something, or must do something — they often react by doing the opposite, even if compliance would benefit them. This isn’t stubbornness for its own sake: it’s a deep psychological need to protect perceived freedom of choice.

The stronger the perceived threat to freedom, the stronger the reactance. Banning something makes it more desirable. Mandating something makes it less appealing. This is why “reverse psychology” works — by appearing to restrict a choice, you can actually increase desire for it.

Reading the Room

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Before acting in any situation, assess the actual dynamics at play: who are the stakeholders? What are their incentives? Where does the power sit? What are the unwritten rules? The same action produces completely different results in different contexts.

Reading the room isn’t manipulation — it’s due diligence. A surgeon who doesn’t assess the patient before operating is reckless, not brave. A leader who doesn’t assess the organisational dynamics before intervening is equally reckless.

Redundancy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Redundancy means building backup capacity into systems so that the failure of any single component doesn’t cause total failure. Nature does this extensively: you have two kidneys, two lungs, two eyes. Each carries more capacity than strictly necessary, providing a buffer against failure.

In engineering, redundancy is a design principle: critical systems have backup power, backup communications, backup data storage. In personal and professional life, redundancy means not having single points of failure — not depending on one income source, one key relationship, one critical skill, or one tool.

Regression to the Mean

Mon, 01 Jan 0001 00:00:00 +0000

How it works

When something performs exceptionally well or badly, the next measurement is likely to be closer to the average. This isn’t because of any corrective force — it’s pure statistics. Extreme results require an unusual alignment of factors, and that alignment is unlikely to repeat.

This creates one of the most common reasoning errors: mistaking regression to the mean for a causal effect. A sports team has a terrible season and fires the coach. Next season, they improve. Was it the new coach? Maybe — but regression to the mean alone would predict improvement after an unusually bad year.

Removal as Strategy

Mon, 01 Jan 0001 00:00:00 +0000

How it works

When facing a decision about how to improve something, the instinct is Via Positiva — add something new. A better tool, a new process, an additional team member. Via Negativa asks the opposite: what could you remove to improve the situation?

Subtraction is underrated because addition feels productive while removal feels like giving up. But removing friction, waste, complexity, and distraction often produces more improvement than any addition could.

Satisficing vs Maximising

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Herbert Simon coined “satisficing” to describe a decision strategy where you choose the first option that meets a set of predefined criteria, rather than exhaustively searching for the optimal choice. The opposite strategy — maximising — involves evaluating every possible option to find the absolute best.

Research by Barry Schwartz and others shows that maximisers, despite spending more time and effort on decisions, tend to be less satisfied with their choices than satisficers. The reason: maximisers always wonder if they could have found something better. Satisficers feel good about their choice because it met their criteria.

Scenario Planning

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Scenario planning, developed at Royal Dutch Shell in the 1970s, is the practice of imagining multiple plausible futures — not to predict which one will happen, but to prepare for several simultaneously. Instead of a single forecast (“the market will grow 8%”), you develop three to four coherent scenarios representing different possible futures.

The power of scenario planning isn’t prediction — it’s mental preparation. When you’ve already imagined and planned for multiple futures, you respond faster and better when reality unfolds, because nothing is entirely unexpected.

Scope Insensitivity

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Scope insensitivity (also called scope neglect) is our failure to respond proportionally to the magnitude of a problem. In a classic study, people were willing to pay $80 to save 2,000 birds, $78 to save 20,000 birds, and $88 to save 200,000 birds. A 100x difference in scope produced virtually no difference in response.

The mechanism: we respond to the mental image of the problem, not its scale. One bird covered in oil creates a vivid emotional response. Ten thousand birds covered in oil creates roughly the same response, because the mental image doesn’t scale.

Selection Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Selection bias occurs when the sample you’re observing isn’t representative of the full population because of how it was selected. You’re drawing conclusions from a filtered subset without realising the filter exists.

The most common form: you only see the data that made it through a selection process, not the data that was filtered out. This is closely related to survivorship bias, but broader — any non-random filtering of observations creates selection bias.

Signal vs Noise

Mon, 01 Jan 0001 00:00:00 +0000

How it works

In any stream of information, a small fraction is signal — data that actually changes your understanding or should change your behaviour — and the rest is noise — random variation that looks meaningful but isn’t.

Nassim Taleb demonstrates this mathematically: if you check a volatile investment daily, you’ll see roughly 50% positive and 50% negative days, generating enormous emotional noise. Check it annually, and you’ll see a much clearer signal of the long-term trend. The underlying reality hasn’t changed — only how much noise you’re exposed to.

Single Point of Failure

Mon, 01 Jan 0001 00:00:00 +0000

How it works

A single point of failure (SPOF) is any component whose failure causes the entire system to fail. The most dangerous single points of failure are the ones you haven’t identified — the dependencies so embedded in your system that you don’t notice them until they break.

Every critical system should be audited for SPOFs: one key employee, one supplier, one server, one revenue source, one communication channel, one decision-maker. If any single element’s failure would be catastrophic, you have a design problem.

Skin in the Game

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Nassim Taleb’s “Skin in the Game” principle states that people make better decisions when they bear the consequences of those decisions. When decision-makers are insulated from downside risk, their judgment becomes unreliable — not because they’re dishonest, but because their incentives are misaligned.

The principle has four dimensions. First, symmetry: people who get the upside should also bear the downside. Second, filtering: systems where participants have skin in the game naturally eliminate incompetence over time. Third, trust: you can trust someone’s advice more when they bear a cost for being wrong. Fourth, ethics: it’s morally wrong to transfer risk to others while keeping the reward.

Specialisation vs Generalisation

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Specialists develop deep expertise in a narrow domain. Generalists develop broad competence across many domains. The optimal strategy depends on environmental stability: in stable, predictable environments, specialists outperform (a narrow but deep advantage wins consistently). In volatile, unpredictable environments, generalists outperform (broad adaptability beats narrow expertise when conditions change).

David Epstein’s research in Range showed that in “kind” learning environments with clear rules and fast feedback (chess, golf, classical music), early specialisation works. In “wicked” environments with ambiguous rules and delayed feedback (business, geopolitics, most real-world domains), generalists who sampled widely before committing outperformed early specialists.

Spotlight Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Thomas Gilovich and colleagues demonstrated that we consistently overestimate how much other people notice and remember about our appearance, behaviour, and mistakes. In one study, students were asked to wear an embarrassing T-shirt (featuring Barry Manilow) to a room full of peers. The wearers estimated that 50% of people in the room noticed the shirt. The actual number: 25%.

The spotlight effect occurs because we’re the centre of our own world. We notice everything about ourselves, so we assume others do too. They don’t — they’re busy being the centre of their own world.

Start Here

Mon, 01 Jan 0001 00:00:00 +0000

What are mental models?

Mental models are thinking frameworks — repeatable patterns of reasoning that help you understand problems, make decisions, and avoid mistakes. They’re the tools your brain uses to process the world. Everyone already has mental models. The question is whether yours are deliberate and effective, or unconscious and full of blind spots.

The world’s best thinkers — Charlie Munger, Elon Musk, Daniel Kahneman, Ray Dalio — don’t just know more facts. They have better models. ThinkingKit is a free library of the most powerful ones, with interactive tools to help you actually use them.

Status Quo Bias

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Status quo bias is the preference for the current state of affairs. People tend to treat the existing situation as a baseline and view any change as a potential loss. This is closely linked to loss aversion — the psychological tendency to weigh losses more heavily than equivalent gains.

The practical effect: we stay in jobs longer than we should, keep subscriptions we don’t use, hold investments past their optimal point, and resist organisational changes that would objectively improve things.

Strawman vs Steelman

Mon, 01 Jan 0001 00:00:00 +0000

How it works

A strawman argument weakens an opponent’s position to make it easy to attack. A steelman argument strengthens an opponent’s position to ensure you’re engaging with the best version of their reasoning. Strawmanning wins arguments. Steelmanning wins truth.

The practical test: could the person you’re arguing against recognise your representation of their position as fair? If they’d say “yes, that’s what I believe,” you’ve steelmanned. If they’d say “that’s not what I said at all,” you’ve strawmanned.

Streetlight Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

An old joke captures this bias perfectly: A man is searching for his keys under a streetlight. A friend asks where he dropped them. “Over there in the dark alley,” he replies. “Then why are you looking here?” “Because the light is better here.”

The streetlight effect describes our tendency to focus investigation where it’s easiest to look, rather than where the answer is most likely to be found. We measure what’s measurable, study what’s accessible, and solve the problems we know how to solve — all while the actual answers may lie in harder-to-reach places.

Survivorship Bias vs Base Rate Neglect

Mon, 01 Jan 0001 00:00:00 +0000

Survivorship Bias

We study the winners and forget the losers — which distorts our understanding of what actually causes success.

Read the full model →

Base Rate Neglect

We ignore how common or rare something is in general, and focus too much on the specific case in front of us.

Read the full model →

What’s the difference?

Survivorship Bias means you’re only seeing the winners and missing the losers. Base Rate Neglect means you’re ignoring how common or rare the outcome is in general. Both involve missing data, but different kinds: survivorship hides who failed; base rate neglect hides how many tried.

Symbiosis

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Symbiosis describes relationships where two organisms benefit from their association. In mutualism, both parties gain. Clownfish get protection from anemones; anemones get food scraps and cleaning from clownfish. Neither thrives as well alone.

Applied to business and life, symbiosis means identifying partnerships where both parties gain more together than either would alone — creating value that neither could create independently.

Case study: How Apple and app developers created a trillion-dollar symbiotic ecosystem

When Apple launched the App Store in 2008, it created a symbiotic relationship with millions of developers. Apple provided the platform, distribution, and payment infrastructure. Developers provided the apps that made iPhones valuable. Each party’s success reinforced the other’s.

Systems vs Goals

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Scott Adams, the creator of Dilbert, distinguishes between goals and systems. A goal is a specific outcome you want to achieve: “lose 20 pounds,” “write a book,” “get promoted.” A system is a process you follow regularly: “eat meals I’ve prepared myself,” “write 500 words every morning,” “learn one new skill each quarter.”

Goals are binary — you’ve either achieved them or you haven’t — which means you spend most of your time in a state of failure. Systems are ongoing — you either followed your system today or you didn’t — which means you succeed every time you execute the process, regardless of the final outcome.

Technical Debt

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Ward Cunningham coined “technical debt” as a metaphor for the accumulated cost of expedient decisions in software development. Like financial debt, it has an initial benefit (faster delivery) and an ongoing cost (interest payments in the form of reduced development speed, increased bugs, and harder maintenance).

Some technical debt is strategic — deliberately chosen to ship faster with a plan to pay it down later. Some is unintentional — the result of poor decisions, insufficient knowledge, or evolving requirements. The most dangerous kind is invisible — debt you don’t know you have until it comes due.

The Lindy Effect

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The Lindy Effect, popularised by Nassim Taleb, states that for non-perishable things (ideas, technologies, books, institutions), the longer something has survived, the longer it’s expected to survive. A book that has been in print for 100 years is likely to remain in print for another 100. A restaurant that has been open for 50 years is more likely to survive another 50 than a restaurant that opened last year.

The Peter Principle

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Laurence Peter observed that in hierarchical organisations, people are promoted based on their performance in their current role. This continues until they reach a role they can’t perform well — at which point promotions stop. The result: over time, every position tends to be filled by someone who is incompetent at it.

The mechanism is simple. A great salesperson gets promoted to sales manager. A great engineer gets promoted to engineering lead. But the skills that made them great at doing the work (selling, coding) are completely different from the skills needed to manage people doing the work. Being good at the previous job doesn’t predict being good at the next one.

The Scientific Method

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The scientific method is the most reliable process humans have developed for separating truth from opinion. Its power comes from a simple structure: form a hypothesis, design a test that could prove it wrong, run the test, observe what happens, and update your beliefs accordingly.

The critical element is falsifiability — you must be able to specify what evidence would disprove your hypothesis. A theory that explains everything explains nothing. If no possible observation could prove you wrong, you’re not doing science — you’re defending a belief.

Thinking Toolkit Matcher

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Most mental model resources give you a long list and leave you to figure out which ones apply to your situation. The Toolkit Matcher flips this: you describe what you’re dealing with, and it recommends the models most likely to help.

Step 1. Tell it what kind of challenge you’re facing — a decision, a problem to solve, a system to understand, a risk to manage, or an argument to evaluate.

Tragedy of the Commons

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Garrett Hardin described the Tragedy of the Commons using a parable: shepherds sharing a common pasture each have an incentive to add one more sheep (the benefit is private), even though overgrazing destroys the pasture (the cost is shared). Each individual acts rationally, but the collective outcome is catastrophic.

The pattern appears whenever a shared resource has no mechanism for limiting individual consumption. Each actor’s rational self-interest leads to the destruction of the resource that everyone depends on.

Unknown Unknowns

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Donald Rumsfeld, in a 2002 press briefing, distinguished three categories of knowledge: known knowns (things we know we know), known unknowns (things we know we don’t know), and unknown unknowns (things we don’t know we don’t know).

Known unknowns can be researched and planned for. Unknown unknowns cannot — by definition, you can’t plan for something you’re not aware of. This is where the real danger lives: in the territory beyond the edge of your map, where you don’t even know the map has edges.

Via Negativa vs Minimum Effective Dose

Mon, 01 Jan 0001 00:00:00 +0000

Via Negativa

Improve by removing what’s harmful rather than adding what might help. Subtraction often beats addition.

Read the full model →

Minimum Effective Dose

Find the smallest input that produces the desired result. More is not always better — sometimes it’s waste.

Read the full model →

What’s the difference?

Via Negativa improves by removing what’s harmful. Minimum Effective Dose finds the smallest input that produces the desired result. Both are about subtraction and efficiency, but Via Negativa focuses on eliminating negatives while MED focuses on finding the threshold of positives.

Via Positiva vs Via Negativa

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Via Negativa (the negative way) means improving by removing what’s harmful rather than adding what might be beneficial. Via Positiva (the positive way) means improving by adding something new. Nassim Taleb argues that Via Negativa is generally more reliable than Via Positiva — it’s easier to identify what’s clearly harmful than to predict what addition will help.

Removing a bad habit is more reliable than adding a good one. Eliminating toxic food is more impactful than adding supplements. Cutting unnecessary meetings is more productive than adding productivity tools.

Wicked Problems

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Horst Rittel and Melvin Webber coined “wicked problems” to describe a class of problems fundamentally different from “tame” ones. Tame problems (like solving an equation) have clear formulations, definitive solutions, and objective tests for correctness. Wicked problems have none of these.

Characteristics of wicked problems: there is no definitive formulation (the problem changes as you understand it better), there is no stopping rule (you can always do more), solutions are not true or false but better or worse, every attempt to solve the problem changes the problem, and there is no opportunity to learn by trial and error (every intervention counts).

Winner's Curse

Mon, 01 Jan 0001 00:00:00 +0000

How it works

The Winner’s Curse occurs in competitive situations — auctions, bidding wars, hiring competitions — where the winner tends to have overpaid. The logic: if many informed people are bidding, the one who bids highest likely overestimated the value more than anyone else. Winning is itself evidence of overvaluation.

The more bidders there are, the stronger the curse. With many competitors, the winning bid is increasingly likely to be an outlier on the high end of estimates.

Zero-Sum vs Positive-Sum

Mon, 01 Jan 0001 00:00:00 +0000

How it works

Zero-sum thinking assumes that every interaction has a fixed pie — that one person’s gain must be another’s loss. Sometimes this is true: in a poker game, every dollar you win comes from another player. But in many situations, the pie can grow. Trade, cooperation, innovation, and technology all create positive-sum games where everyone can gain.

The most consequential thinking error is applying zero-sum logic to positive-sum situations — treating negotiations as battles, viewing colleagues as competitors, or opposing trade because “if they gain, we lose.”