You just made a confident call that was badly wrong.
- Most likely cause
- Overconfidence Effect
- Also check
- Dunning-Kruger Effect · Planning Fallacy
- Fix
- Apply CL-3 (Outside View Switch).
THE TOOLKIT
Everything you do with the map
The map names the traps. This is where you act on them — the protocols that correct them, the checklist that catches them before a decision, the shortcuts for when there's no time, the clusters that compound, and where each mechanism shows up beyond the lab.
Every countermeasure in the Lexicon reduces to one of these ten protocols. Entries reference them by name — learning them as a set is worth more than one at a time.
CL-1
The Pause Protocol
Stop before you act. The gap between stimulus and response is where judgment lives.
Most cognitive errors happen at speed. Inserting a deliberate pause before any significant decision, response, or judgment reduces the probability that System 1 is running the show. The pause does not need to be long - even five seconds changes the nature of the decision.
Trigger Strong emotion, time pressure, or an answer that came suspiciously easily.
Used across Affect Heuristic · Limbic Hijacking · Reactance Theory · Tazwell's Razor
CL-2
The Base Rate Check
Before estimating anything, ask what usually happens.
The most consistent improvement in judgment comes from anchoring estimates in actual frequencies rather than vivid examples or intuitions. Before assessing risk, likelihood, or outcome, ask: “What is the base rate for this in the relevant population?”
Trigger Any probability estimate, risk assessment, or prediction about an individual case.
Used across Availability Heuristic · Base Rate Fallacy · Inside View vs Outside View · Planning Fallacy · Representativeness Heuristic
CL-3
The Outside View Switch
Stop looking at your specific situation. Ask what usually happens to situations like this.
Kahneman’s most practical recommendation: when making predictions about your own situation, deliberately shift from the inside view (your specific details and plans) to the outside view (the base rate for projects, relationships, or decisions of this type).
Trigger Any planning, forecasting, or confidence judgment about your own project or situation.
Used across Planning Fallacy · Optimism Bias · Overconfidence Effect · Champion Bias · Survivorship Bias
CL-4
The Disconfirming Evidence Search
Find the best evidence against your current view before concluding.
Confirmation bias is best countered not by trying to be objective, but by actively and effortfully seeking disconfirming evidence. The search must be genuine, not performative. Ask: “What is the strongest evidence against my current position?” and “What would change my mind?”
Trigger Any research task, belief you hold strongly, or decision where you have a strong prior.
Used across Confirmation Bias · Belief Perseverance · Backfire Effect · Cherry-Picking · Opinion Shopping
CL-5
The Emotion Label
Name the emotion before acting on it.
Affect labelling - putting a feeling into words - reduces its intensity and restores prefrontal engagement. Identifying and naming the emotional state before making a decision separates the feeling from the judgment.
Trigger Strong or unexpected emotion before or during a decision.
Used across Affect Heuristic · Empathy Gap · Limbic Hijacking · Projection Bias · Tazwell's Razor
CL-6
The Pre-Mortem
Imagine it has already failed. Now work backward.
Before committing to a plan, assume it has failed and generate the most plausible explanation for that failure. This exploits the same imaginative mechanism that drives Planning Fallacy and Optimism Bias - but redirects it toward identifying vulnerabilities rather than reinforcing confidence.
Trigger Before any significant commitment - a project, investment, relationship, or strategy.
Used across Planning Fallacy · Optimism Bias · Overconfidence Effect · Commitment Bias · Plan Continuation Bias
CL-7
The Steel Man
Before refuting an argument, make it as strong as possible.
State the opposing position in its strongest, most charitable form before attempting to rebut it. This forces genuine engagement with the actual argument rather than a weakened caricature, and often reveals that the disagreement is smaller than it appeared.
Trigger Any argument, disagreement, or situation where you are about to respond to someone’s position.
Used across Straw Man · Motte-and-Bailey · Grice’s Razor · Rogerian Rhetoric · Gish Gallop
CL-8
The Environment Design
Change the environment rather than relying on willpower.
Willpower is finite and unreliable. Most self-regulation failures occur not from lack of effort but from being in the wrong environment at the wrong time. Structural changes - removing temptations, reducing friction for desired behaviours, automating defaults - produce more durable change than intention alone.
Trigger Any recurring self-regulation failure, or any new behaviour you want to make automatic.
Used across Ego Depletion Effect · Environmental Design · Willpower Fatigue · Default Effect · Habit Loop Effect
CL-9
The Source Trace
Follow the claim back to where it actually came from.
Most widely-shared statistics, findings, and assertions have travelled far from their original source and been distorted in transit. Before updating on any factual claim, ask: “Who originally measured this, and how?” Follow the citation chain to the primary source.
Trigger Any claim presented as established fact, especially if it is precise-sounding or widely shared.
Used across Woozle Effect · Illusory Truth Effect · Backfire Effect · Misinformation Effect · Cherry-Picking
CL-10
The Sunk Cost Reset
Ask only what the future holds. What you have spent is gone.
When evaluating whether to continue any investment of time, money, or emotion, deliberately set aside what has already been spent and ask: “If I were starting fresh today with no prior investment, would I choose this?” The answer - not the sunk cost - should drive the decision.
Trigger Any decision to continue an existing commitment - a project, relationship, investment, or position.
Used across Sunk Cost Fallacy · Commitment Bias · Escalation of Commitment · Disposition Effect
Run through before any significant decision — about 90 seconds. Each question names the mechanism it checks. Tick the ones that produce an uncomfortable answer; if more than 3 do, run a pre-mortem before you commit.
No time to navigate: match your situation to the closest card, then take the fix. Each points to the most likely cause and what else to check.
No card matches. Try fewer words, or scan the map.
Your team keeps agreeing with you and it feels wrong.
- Most likely cause
- Groupthink
- Also check
- Preference Falsification · Abilene Paradox
- Fix
- Assign someone to argue the opposite case before the next meeting.
You cannot stop a behaviour you know is harming you.
- Most likely cause
- Present Bias
- Also check
- Habit Loop Effect · Hyperbolic Discounting
- Fix
- Apply CL-8 (Environment Design) - change the environment, not the willpower.
You stayed too long in a job, relationship, or investment.
- Most likely cause
- Sunk Cost Fallacy
- Also check
- Commitment Bias · Loss Aversion
- Fix
- Ask - “If I were starting fresh today, would I choose this?”
Someone is winning an argument using tricks, not truth.
- Most likely cause
- Motte-and-Bailey
- Also check
- Straw Man Fallacy · Gish Gallop
- Fix
- Ask them to restate their actual position precisely before responding.
A policy is producing the opposite of what it was designed to do.
- Most likely cause
- Goodhart’s Law
- Also check
- Cobra Effect · Campbell’s Law
- Fix
- Ask who is gaming the metric and what incentive structure rewards them for it.
You feel emotionally activated by something online and want to act.
- Most likely cause
- Ragebait
- Also check
- Limbic Hijacking · Engagement Maximisation Capture
- Fix
- Apply CL-1 (Pause Protocol). Ask: was this content designed to activate me?
A statistic or study does not feel right.
- Most likely cause
- Selection Bias
- Also check
- P-Hacking · Survivorship Bias
- Fix
- Ask who was in the sample and who was left out.
You achieved your goal but feel nothing.
- Most likely cause
- Arrival Fallacy
- Also check
- Hedonic Treadmill · Impact Bias
- Fix
- This is normal. Do not set the next goal immediately - ask what you were actually chasing.
You are relying on an AI system for an important decision.
- Most likely cause
- Automation Bias
- Also check
- LLM Hallucination · AI Sycophancy
- Fix
- Apply CL-9 (Source Trace) to every specific claim before acting on it.
You changed your position after someone got angry at you.
- Most likely cause
- Conformity Bias
- Also check
- Boomerang Effect · Reactive Devaluation
- Fix
- Separate the argument from the emotional pressure. Would you accept this logic from a stranger?
Your prediction keeps being wrong in the same direction.
- Most likely cause
- Planning Fallacy
- Also check
- Optimism Bias · Inside View vs Outside View
- Fix
- Find the base rate for similar projects. Use that as your starting anchor.
You believe something strongly but cannot explain or defend it.
- Most likely cause
- Belief Perseverance
- Also check
- Motivated Reasoning · Cognitive Dissonance
- Fix
- Apply CL-4 (Disconfirming Evidence Search) - actively seek one strong counterargument.
Someone keeps redefining the terms mid-argument.
- Most likely cause
- Moving the Goalposts
- Also check
- Motte-and-Bailey · Equivocation
- Fix
- Write down the original claim before the conversation moves on.
A group of smart people made an obviously bad decision.
- Most likely cause
- Groupthink
- Also check
- Cascade Effects · Preference Falsification
- Fix
- Ask whether anyone in the room actually disagreed but did not say so.
Ten named clusters where several mechanisms activate at once and push the same direction — the assembled machines behind catastrophic decisions. Individual entries describe the components; this is the compound.
01
The Escalation Trap
Sunk Cost Fallacy+Commitment Bias+Loss Aversion+Optimism Bias
How they interact
Prior investment makes exit feel like loss. Loss aversion weights that loss at 2:1. Public commitment creates identity pressure to stay consistent. Optimism bias sustains belief that the next iteration will succeed. The four together produce rational-feeling persistence toward irrational ends - the subjective experience is not irrationality but determination.
Failure pattern
Continuing a failing project, relationship or strategy well past the point where neutral analysis would have exited. Classic in corporate strategy, military campaigns, failing startups and deteriorating relationships.
Structural countermeasure
Set exit criteria in writing before commitment is made. The criteria must include objective thresholds, not feelings and must be agreed before emotional investment forms.
Before committing further to a failing project, ask: which of these four mechanisms are currently active? Sunk cost (I have already invested too much to stop), commitment bias (I have publicly committed), loss aversion (stopping feels like losing), optimism bias (I believe the next iteration will succeed). If three or more are active simultaneously, treat this as a structural trap, not a strategic decision.
02
The Ideological Bubble
Confirmation Bias+Availability Heuristic+In-Group Bias+Dunning-Kruger Effect
How they interact
Confirming evidence is selectively sought and selectively recalled. Vivid in-group examples load the mental sample. Out-group positions are attributed to character flaws rather than reasoned difference. Partial competence produces certainty that the position is correct and the opposing view is simply uninformed.
Failure pattern
A closed epistemological system where disconfirming evidence is categorised as bias, error or bad faith rather than data. Self-correcting mechanisms are eliminated. Classic in political communities, religious organisations, professional echo chambers and online tribes.
Structural countermeasure
Exposure to the strongest version of the opposing position, presented by its most credible advocate, before forming a final view.
Before reaching a conclusion on a contested topic, ask: when did I last encounter the strongest version of the opposing argument, presented by its most credible advocate? If the answer is "not recently" or "I am not sure what it is", the bubble is likely active.
03
The Overconfident Expert
Dunning-Kruger Effect+Einstellung Effect+Bias Blind Spot
How they interact
High competence produces genuine pattern recognition, which feels indistinguishable from certainty. The first solution found is also the most familiar solution. Bias blind spot prevents recognition that expert judgment is subject to the same distortions as novice judgment, just different ones.
Failure pattern
Expert dismissal of novel solutions, refusal to update on anomalous data and inability to recognise domain boundary conditions where prior expertise does not apply. Classic in medicine, finance, military strategy and scientific consensus maintenance.
Structural countermeasure
Explicitly identify which elements of the current problem differ from the reference class. Bring in outside expert review of the anomalous elements specifically.
Before applying a familiar solution to a new problem, ask: in what specific ways does this problem differ from the reference class my expertise was built on? If you cannot name three, the Einstellung Effect may be blocking search.
04
The Bad Hire
Halo Effect+Fundamental Attribution Error+Confirmation Bias
How they interact
A strong first impression creates a positive halo that reinterprets all subsequent information favourably. Shared background, communication style or appearance activates similarity preference. Positive attributes are attributed to character; any negative signals are attributed to situational factors. Post-decision information is filtered to confirm the initial impression.
Failure pattern
Systematically flawed hiring, promotion and partnership decisions that feel like good judgment. Classic in any hiring process, investment due diligence, political endorsement and academic admission.
Structural countermeasure
Structured interview with identical questions and scoring rubrics before any informal conversation. Blind evaluation of credentials and work samples before meeting the candidate.
Before making a hiring or partnership decision, ask: what did I know about this person before I met them and how has meeting them changed my evaluation? If meeting them increased your assessment substantially without new evidence about their work quality, the Halo Effect and similarity preference are likely active.
05
The Panic Decision
Availability Heuristic+Affect Heuristic+Action Bias+Negativity Bias
How they interact
A vivid negative event loads the mental sample with threat. Emotional arousal amplifies the perceived urgency. Action feels necessary and preferable to inaction regardless of expected value. Negative events receive disproportionate attentional weight. The four together manufacture a crisis from a normal adverse event and demand a response.
Failure pattern
Rapid, visible action taken in response to emotionally salient events that would benefit from slower, deliberate analysis. Classic in policy after disasters, product decisions after bad reviews, strategy changes after competitor announcements and medical decisions after scares.
Structural countermeasure
A mandatory waiting period before responding to any high-emotion event. Explicit base-rate consultation: how often do events of this type warrant the response being considered?
Before responding to a high-visibility adverse event, ask: what is the base rate for events of this type requiring the response I am considering? And: what decision would I make if this event had occurred three months ago, quietly, without media attention? If the answers differ, Availability Heuristic and Action Bias are driving the response.
06
The False Consensus
Groupthink+Social Proof
How they interact
Each individual conforms to what they believe others think. Nobody expresses their private reservations. Social proof signals that the group is aligned. Groupthink pressure actively suppresses dissent. The result is a group decision that nobody privately endorses but that feels unanimous.
Failure pattern
Catastrophic group decisions that feel unanimously endorsed but reflect no individual’s actual judgment. Classic in corporate boards, military command structures, scientific peer review and political cabinets.
Structural countermeasure
An anonymous pre-vote on key assumptions before discussion. A designated devil’s advocate role with genuine authority. Dissent explicitly rewarded in the group’s stated values, not just its aspirational ones.
Before treating group agreement as genuine consensus, ask: did each person in this group form their view before hearing others or did the group converge during discussion? If views formed during the meeting, collect them anonymously before discussion in the next high-stakes decision.
07
The Rationalised Fraud
Motivated Reasoning+Normalisation of Deviance+Commitment Bias
How they interact
Small rule violations go unaddressed and become the new normal. Motivated reasoning generates post-hoc justification for each increment. Moral disengagement separates actions from their consequences through abstraction or diffusion. Commitment to the outcome makes reversal feel more costly than continuation.
Failure pattern
Gradual escalation from minor ethical compromise to major violation, with each step feeling justified and the aggregate invisible. Classic in financial fraud, institutional cover-up, political corruption and research misconduct.
Structural countermeasure
Zero-tolerance thresholds for first violations, not a graduated response. Explicit naming of the normalisation pattern in organisational culture so individuals can recognise it while the deviation is still small.
Before accepting a small rule exception as a one-time accommodation, ask: what would this decision look like if it became the new baseline? If the answer is uncomfortable, Normalisation of Deviance has begun.
08
The Stagnant Strategy
Status Quo Bias+Loss Aversion+Commitment Bias+Sunk Cost Fallacy
How they interact
The current strategy requires no action to continue. Changing it feels like a loss. Prior investment argues for continuity. Public commitment to the current direction creates identity pressure. The four together make strategic pivots feel costly, risky and personally threatening even when the evidence for change is overwhelming.
Failure pattern
Organisational and personal persistence with failing strategies long after the signal for change is clear. Classic in corporate strategy, political platforms, personal career choices and institutional reform.
Structural countermeasure
Scheduled strategic reviews with pre-committed criteria for revision. The criteria must be set before the current strategy is enacted, not evaluated against it retrospectively.
Before rejecting a strategic pivot, ask: if this were a new organisation with no prior strategy, would the current approach be chosen today? If not, identify specifically which of the four mechanisms - status quo, loss aversion, commitment bias, sunk cost - is providing the resistance.
09
The Forecasting Disaster
Planning Fallacy+Optimism Bias+Overconfidence Effect
How they interact
Your specific plan feels uniquely tractable. Base rates for similar projects are ignored in favour of the inside view. Optimism inflates every estimate upward. High subjective confidence is then assigned to the inflated estimate. The four together produce systematically optimistic forecasts delivered with high certainty.
Failure pattern
Every project that comes in late and over budget. Every strategic plan that underestimates time, cost and competitive response. Classic in construction, software development, new product launches and personal financial planning.
Structural countermeasure
Reference-class forecasting as a mandatory first step. Estimate based on what usually happens to projects of this type before adjusting for specific factors. Never adjust the reference-class estimate upward by more than 20% on the basis of inside-view factors alone.
Before finalising a project estimate, ask: what is the reference class for this type of project and what is the median outcome? Then ask: what specific evidence justifies adjusting this estimate above the median? If the adjustment is more than 20% upward and the evidence is primarily inside-view, the cluster is active.
10
The Narrative Trap
Narrative Fallacy+Hindsight Bias+Outcome Bias
How they interact
Random outcomes acquire explanatory stories. The story makes the outcome feel inevitable in retrospect. Decisions are judged by their outcomes rather than by the quality of the reasoning that produced them. Causal attribution assigns credit and blame to agents rather than to the distribution of possible outcomes they faced.
Failure pattern
Learning the wrong lessons from both success and failure. Good processes that produced bad outcomes are abandoned. Bad processes that produced good outcomes are institutionalised. Classic in post-mortem analysis, performance review, historical interpretation and investment strategy evaluation.
Structural countermeasure
Separate process review from outcome review. Evaluate the decision against the information available at the time it was made, not against the information that became available after. Score the quality of reasoning, not the quality of result.
Before drawing a lesson from a success or failure, ask: were there other projects that followed the same process and produced different outcomes? If yes, the lesson may be about the distribution of outcomes, not about the specific choices made. Separate process evaluation from outcome evaluation explicitly.
The mechanisms were mostly discovered in psychology and economics labs. They operate everywhere — 80 highest-transfer concepts across 8 domains, each with a line on how it shows up. Unrevealed names render as plain text until the Lexicon publishes them.
AI Systems & Machine Learning
- Confirmation BiasTraining data reflects the biases of those who curated it; evaluation benchmarks reflect the assumptions of those who designed them. Models inherit the confirmation structure of their training pipeline.
- Survivorship BiasModels are evaluated on benchmarks built from available data. Failure modes that produce no data - silent errors, out-of-distribution collapses - are invisible in evaluation.
- Goodhart's LawAny metric used to optimise a model ceases to be a good measure of what it was intended to measure. RLHF reward hacking is Goodhart's Law in action.
- Dunning-Kruger EffectAI systems show high confidence in domains where their training data is thin or miscalibrated - the same metacognitive blindness that characterises low-competence human performance.
- Sunk Cost FallacyOrganisations continue investing in underperforming model architectures because prior investment makes pivoting feel more costly than continuing.
- Planning Fallacy"Six months to production-ready" is a planning fallacy until proven otherwise - AI deployment timelines are subject to the same systematic optimism as all complex technical projects.
- Availability HeuristicAI safety priorities are shaped by recent, vivid failure modes rather than by statistical frequency of actual risk. Post-incident investment patterns follow availability, not expected harm.
- Framing EffectModel capability claims are highly sensitive to how benchmarks are framed and which comparators are chosen. The same performance looks different depending on what is held constant.
- In-Group BiasAI research communities develop shared assumptions, preferred methods and evaluation frameworks that function as in-group norms - systematically undervaluing approaches from outside the group.
- Overconfidence EffectCapability claims consistently outpace demonstrated generalisation. The confidence-accuracy gap in AI capability assessment mirrors the documented gap in human probability calibration.
Military Strategy & Geopolitics
- GroupthinkMilitary intelligence failures follow the false-consensus pattern: dissenting analysis suppressed, alternatives not explored, unanimity mistaken for accuracy.
- Sunk Cost FallacyMilitary campaigns and geopolitical confrontations follow the sunk cost, commitment and loss aversion cluster with predictable regularity - exit costs are perpetually reframed as more costly than continuation.
- Planning FallacyMilitary operations are consistently underestimated in duration, cost and complexity. Historical base rates for similar operations are routinely ignored in favour of inside-view optimism.
- Fundamental Attribution ErrorAdversarial behaviour is attributed to character and ideology while ignoring the strategic situation, institutional pressures and information available to the other party.
- Availability HeuristicStrategic threat assessment is dominated by the most recent and vivid threat, producing systematic over-preparation for the last conflict and under-preparation for the next one.
- Normalisation of DevianceRules of engagement, treaty obligations and operational constraints are progressively relaxed through accumulated precedent until the original constraint is unrecognisable.
- Narrative FallacyPost-war historical accounts construct coherent causal narratives from contingent sequences of events. Counterfactual analysis consistently shows the narratives overstate causal necessity.
- Loss AversionNuclear deterrence, territorial disputes and face-saving dynamics are primarily driven by loss aversion rather than rational gain-seeking.
- In-Group BiasAlliance cohesion and coalition fragility are partially explained by in-group preference dynamics. Cross-cutting identities reduce the effect; shared threat amplifies it.
- Social ProofDiplomatic consensus mechanisms are subject to social proof cascades where visible alignment produces further alignment independent of underlying national interests.
Medicine & Healthcare
- Omission BiasPhysicians consistently over-treat relative to under-treat, partly because acts that cause harm are judged more harshly than omissions of equivalent harm.
- Availability HeuristicDiagnostic anchoring on recently seen cases produces systematic errors in base-rate estimation. A physician who recently treated a rare condition over-diagnoses it in subsequent patients.
- Planning FallacyClinical trial timelines, hospital construction projects and healthcare IT implementation consistently run over schedule and over budget at rates comparable to other complex projects.
- Status Quo BiasEvidence-based treatment protocols take 15 to 20 years to achieve widespread adoption after publication - the current treatment is the default; evidence for a better alternative must overcome status-quo inertia.
- Loss AversionPatients accept more risk to avoid certain loss than to achieve equivalent gain. Treatment framing (survival rates versus mortality rates) exploits this asymmetry and should be presented in both forms.
- Confirmation BiasOnce a working diagnosis is formed, subsequent evidence is interpreted through it and disconfirming evidence is weighted down.
- Overconfidence EffectPhysician confidence in diagnoses is systematically higher than diagnostic accuracy in controlled studies. Autopsy studies show discordance rates higher than confidence levels predict.
- Fundamental Attribution ErrorPatient non-compliance is attributed to character rather than to situational factors - cost, access, competing demands, side-effect profiles - that often better predict adherence.
- GroupthinkMedical consensus can suppress anomalous findings and minority clinical experience. H. pylori (Barry Marshall) and hand-washing (Semmelweis) illustrate the cost of expert consensus operating as groupthink.
- Normalisation of DevianceHealthcare-associated infection rates, medication error rates and near-miss events are subject to normalisation of deviance: rates that would have been alarming become baseline through gradual recalibration.
Product Design & Technology
- Default EffectDefault settings determine the majority of user behaviour. The choice to set a privacy, notification or data-sharing default is a policy decision that shapes outcomes more than any individual user choice.
- Loss AversionTeams resist removing features even when removal would improve the product, because removal feels like loss and addition feels like gain.
- Availability HeuristicProduct priorities are shaped by the most recent, vocal or visible user complaints rather than by systematic evidence of user need frequency or value.
- Framing EffectThe same price presented as "saving £5" versus "avoiding a £5 charge" produces different conversion rates - pricing, trial offers and upgrade prompts are all framing decisions.
- Sunk Cost FallacyTechnical debt accumulates because refactoring existing code feels like discarding past work - the sunk cost of the current architecture drives continued investment in a known-suboptimal foundation.
- Confirmation BiasUser research is distorted when teams formulate and test hypotheses in the same study. Separate discovery from validation research.
- Dunning-Kruger EffectNew users overestimate their ability to use complex products after brief exposure. Onboarding that surfaces full complexity too early triggers the competence gap; delaying it produces the overconfidence that leads to abandonment.
- Endowment EffectUsers who have invested time configuring a product resist switching to better alternatives - switching cost is partly psychological, not just technical.
- Planning FallacyProduct roadmaps are planning-fallacy artefacts. The reference class for features at similar complexity should anchor estimates before any inside-view adjustment.
- Social ProofUser review systems, download counts and "X people are viewing this" signals are social-proof mechanisms, most influential for unfamiliar products and least for products where the user has strong prior preferences.
Finance & Investing
- Loss AversionInvestors hold losing positions longer than winning ones across all documented portfolio studies - the disposition effect is loss aversion applied to portfolio management.
- Overconfidence EffectActive fund managers' confidence in above-average performance is uncalibrated: most active funds underperform passive indices net of fees over any ten-year period.
- Planning FallacyFinancial projections in business plans and investment memoranda are subject to the same systematic optimism as all planning. Reference-class forecasting outperforms inside-view financial models.
- Availability HeuristicPost-crisis investment behaviour is dominated by the most recent crisis. Post-2008 risk models overweight the specific failure mode of 2008; the next crisis will be different.
- Anchoring BiasIPO pricing, acquisition multiples and analyst price targets all show anchoring effects that produce insufficient adjustment from prior reference points.
- Confirmation BiasAnalysts and investors seek confirming evidence for positions they already hold and discount disconfirming signals, amplified by the social dynamics of investment teams.
- Narrative Fallacy"Markets fell because of X" narratives are routinely constructed from events that occurred after the movement began.
- Social ProofFinancial markets are subject to information cascades: price movements that attract attention generate further price movements through social proof rather than fundamental information.
- Sunk Cost FallacyContinued investment in underperforming portfolio companies happens because prior investment makes exit feel more costly than continuation.
- Hyperbolic DiscountingShort-termism in investment decisions is partly explained by hyperbolic discounting: quarterly earnings pressure, fund manager career risk and institutional incentives all reinforce present bias over long-horizon value creation.
Law & Justice
- Anchoring BiasSentencing anchors - prosecutorial recommendations, prior sentences for similar crimes - produce documented effects on judicial decisions even after controlling for case characteristics.
- Fundamental Attribution ErrorCriminal sentencing overweights dispositional factors and underweights situational ones (poverty, trauma, systemic disadvantage) in ways that predict documented sentencing disparities.
- Confirmation BiasProsecutorial tunnel vision - continuing to pursue a case despite accumulating disconfirming evidence - follows the confirmation-bias pattern. High-profile wrongful-conviction cases consistently show this structure.
- Framing Effect"Assault" versus "altercation," "terrorist" versus "militant" - legal and political language shapes judgment in ways that go beyond the underlying facts.
- Outcome BiasLegal decisions that produced bad outcomes are judged more harshly than equally reasonable decisions that happened to produce good outcomes.
- In-Group BiasMock jury research and real case data both show in-group and out-group dynamics affecting verdict and sentencing outcomes in ways that correlate with defendant-juror demographic overlap.
- Hindsight BiasNegligence cases hinge on what a reasonable person "should have known" before the harm occurred. Hindsight bias systematically inflates assessments of prior foreseeability, disadvantaging defendants.
- Authority BiasExpert witness testimony is weighted by apparent authority and presentation quality as well as evidential value - the same testimony delivered with different credibility signals produces different jury responses.
- Overconfidence EffectEyewitness confidence is systematically uncalibrated: high-confidence identifications are only modestly more accurate than low-confidence ones yet are given substantially more weight by juries.
- GroupthinkJury deliberation dynamics follow groupthink patterns in some conditions: early consensus formation, pressure on dissenting jurors and failure to adequately examine minority viewpoints.
Education & Learning
- Desirable DifficultyConditions that make learning harder in the short term - spaced practice, interleaving, retrieval practice - produce better long-term retention than conditions that feel easier.
- Dunning-Kruger EffectStudents who believe they understand material stop studying it before genuine understanding is achieved - the peak of overconfidence at early competence is particularly consequential in education.
- Confirmation BiasStudents seek confirming examples of their understanding and avoid challenging problems that might reveal gaps. Instructors who only call on confident hands reinforce this dynamic.
- Overconfidence EffectMetacognitive calibration - knowing what you know and don’t know - is itself a learnable skill that standard education rarely teaches explicitly, producing cohorts with poor self-assessment accuracy.
- Availability HeuristicCurriculum design is subject to availability bias: topics that generate vivid examples, media coverage or passionate teachers receive more time than their importance to the domain warrants.
- Narrative FallacyEducation organised around heroic narratives overstates causal agency and understates structural, collaborative and contingent factors.
- Present BiasImmediate social rewards - peer approval, entertainment - consistently outcompete delayed academic rewards in moment-to-moment attention allocation.
- Fluency EffectStudents who feel they understand material during a well-delivered lecture perform no better on tests than students who struggled with worse-delivered content.
- Intrinsic MotivationExternal evaluation systems - grades, rankings, prizes - systematically undermine intrinsic motivation in domains where the external reward is salient.
- Social ProofAcademic norms, citation practices and research fashions are subject to social-proof cascades: topics that attract citations attract more researchers regardless of their actual importance.
Organisational Leadership
- GroupthinkStrategic planning processes, board decisions and leadership consensus are subject to groupthink dynamics. The countermeasure - adversarial collaboration, anonymous pre-voting, a designated devil’s advocate - must be structural, not aspirational.
- Normalisation of DevianceOrganisational culture degrades through accumulated small exceptions. Performance standards and ethical boundaries that are not actively maintained erode gradually through the rationalised-fraud pattern.
- Fundamental Attribution ErrorLeaders attribute poor performance to individual capability and motivation and systematically underweight system design, resource constraints and role clarity - producing personnel solutions to system problems.
- Loss AversionOrganisational risk appetite is shaped by loss aversion. New initiatives must overcome the perceived risk of visible failure; the cost of not innovating is diffuse and attributed to market forces rather than decisions.
- Confirmation BiasLeadership teams develop shared mental models that function as confirmation filters - signals that challenge the prevailing strategy are systematically downweighted.
- Planning FallacyStrategic plans consistently underestimate time, cost and competitive response and overestimate adoption rates and team capacity - planning fallacy operating at scale.
- Decision FatigueLeadership decision quality degrades through the day and week. High-stakes decisions scheduled at the end of full agendas are made by depleted decision-makers applying default responses.
- Status Quo BiasOrganisational change resistance is not primarily psychological - it is structural. Incentive systems, role definitions and performance metrics all embed the status quo.
- Sunk Cost FallacyContinued investment in failing products, strategies and personnel is driven partly by sunk-cost reasoning - the longer and more publicly a commitment has been made, the more expensive exit feels.
- Escalation of CommitmentLarge organisational projects follow the escalation-trap pattern with high regularity. Pre-committed exit criteria, set before the project begins, are the only structural countermeasure with documented effectiveness.
Every framework has edges where its own logic breaks down. A document that doesn't name those edges isn't honest — it's advocacy. Read this before deploying any concept here in a high-stakes conversation.
01
Premature Labelling
What it is Naming a bias before establishing that the underlying judgment was actually wrong.
Failure pattern
A person makes a fast, confident decision. An observer with access to this Lexicon labels it "System 1 error" or "availability heuristic." The decision turns out to be correct - the fast judgment was drawing on calibrated pattern recognition. The label was applied to the process before the output was evaluated.
Why it matters
Premature labelling treats the presence of a fast or emotionally resonant judgment as sufficient evidence of bias. It confuses the appearance of a cognitive pattern with the operation of a cognitive error. Expert intuition in high-familiarity domains looks identical to bias from the outside.
Countermeasure
Before labelling a judgment as biased, ask: what is the evidence that the judgment was wrong, not just that it was fast or confident? Bias is a claim about accuracy, not about process.
02
The Bias Bias
What it is The systematic tendency to see cognitive biases operating even in situations where the heuristic produced a correct result.
Failure pattern
A person fluent in this Lexicon begins labelling every confident belief as overconfidence, every risk aversion as loss aversion, every fast decision as System 1 error. The framework becomes a hammer; everything looks like a bias. The person becomes less accurate in their judgments, not more, because they are now second-guessing calibrated intuitions alongside uncalibrated ones.
Why it matters
Kahneman noted this explicitly: teaching bias awareness without teaching when fast judgment is correct produces worse outcomes than no training. The framework’s value is selective application, not universal application.
Countermeasure
For each suspected bias, ask: is there evidence the judgment was wrong or only evidence that it was produced by a process associated with errors? Calibrated fast judgment is not bias. Keep a record of your post-bias-labelling decisions to check whether your accuracy improved or declined.
03
Overcorrection Spiral
What it is Applying countermeasures so aggressively that the correction produces a worse error class than the original bias.
Failure pattern
A person learns about optimism bias and begins systematically adjusting all estimates downward. Their forecasts become pessimistic. Their risk models become overly conservative. They pass on opportunities with genuine positive expected value. The correction overshoots by as much as the original bias.
Why it matters
Every countermeasure has a correct calibration range. Outside that range, the correction is itself a bias in the opposite direction. Availability heuristic overcorrection produces base-rate anchoring that ignores genuine recent signals. Loss aversion overcorrection produces reckless risk tolerance. Dunning-Kruger overcorrection produces impostor syndrome.
Countermeasure
Track the direction of your corrections over time. If you are consistently adjusting in one direction, the correction has become a new bias. The goal is calibration, not reversal.
04
Label Weaponisation
What it is Using bias vocabulary to dismiss legitimate disagreement rather than to understand it.
Failure pattern
In a meeting, one person argues against a proposal. Another person, fluent in this Lexicon, responds: "That sounds like loss aversion" or "You might want to check your negativity bias on this." The label is used as a rhetorical move to close down the argument rather than engage with its content. The concern raised may have been valid. It never gets examined.
Why it matters
This is the most socially damaging failure mode of the framework. It converts tools for self-reflection into weapons for dismissing others. It is epistemically dishonest because it substitutes a process claim (your reasoning looks biased) for a content claim (here is why you are wrong). It is also socially corrosive - people stop raising concerns when they expect to be labelled rather than heard.
Countermeasure
The bias framework applies to your own reasoning first. When applying it to another person’s reasoning, engage with the content of their argument before or instead of the process. "I notice this might be loss aversion and here is why I think the risk is actually manageable" is legitimate. "That’s just loss aversion" is not.
05
False Calibration
What it is Treating knowledge of bias names as equivalent to reduced susceptibility to the biases named.
Failure pattern
A person reads this Lexicon and concludes they are now better at avoiding cognitive biases. Their confidence in their own reasoning increases. But the research is consistent: knowing the name of a bias does not reduce its operation. The person is now more confident in reasoning that has not actually improved.
Why it matters
This is the most insidious failure mode because it is invisible to the person experiencing it. The knowledge feels like protection. It is not. Kahneman spent his career studying cognitive biases and reported no improvement in his own performance on the tasks he had spent decades studying.
Countermeasure
Treat this collection as a recognition tool, not an immunity protocol. The test of improved reasoning is not whether you can name the bias - it is whether your predictions are better calibrated, your decisions produce better outcomes and your estimates track reality more accurately over time. Measure outcomes, not vocabulary.
06
Framework Paralysis
What it is Running every decision through so many diagnostic lenses that no decision gets made or that decision quality degrades from excessive self-monitoring.
Failure pattern
A person internalises this Lexicon and begins applying multiple diagnostic checks to every decision. Simple choices take disproportionate time. High-pressure decisions are delayed past their useful window. The person experiences decision fatigue not from volume but from self-interrogation.
Why it matters
The framework is a cost-benefit tool. Applying it to low-stakes, reversible, high-familiarity decisions consumes resources without commensurate benefit. The cognitive overhead of systematic debiasing is itself a resource expenditure that should be proportional to decision stakes.
Countermeasure
Apply the Pre-Decision Recognition Checklist to high-stakes, irreversible, novel decisions. Apply fast satisficing to low-stakes, reversible, familiar decisions. Calibrate the depth of diagnostic work to the cost of being wrong, not to the availability of the framework.
07
The Sophistication Trap
What it is Higher cognitive sophistication producing better rationalisation of predetermined conclusions rather than better reasoning toward correct ones.
Failure pattern
A person fluent in this Lexicon and in general reasoning uses their sophistication to construct elaborate post-hoc justifications for conclusions they reached emotionally. The reasoning is technically more sophisticated than naive rationalisation. It is still rationalisation. The sophistication trap is that the more tools a person has for reasoning, the more tools they have for motivated reasoning.
Why it matters
Intelligence correlates with motivated reasoning, not inversely. More capable reasoners produce more convincing rationalisations. The framework gives sophisticated users more sophisticated tools for this failure mode.
Countermeasure
The test is: at what point in the reasoning process did you form your conclusion? If the conclusion preceded the reasoning, the reasoning is likely rationalisation regardless of its quality. Practise steel-manning as a deliberate check on whether you can make the opposing case as well as your own.
08
Expertise Mismatch
What it is Applying a bias framework developed in WEIRD laboratory settings to cultural, professional or experiential contexts where it does not accurately predict behaviour.
Failure pattern
A consultant trained in Western behavioural economics advises an organisation in a high-context, collectivist cultural environment. They apply loss-aversion framing, individualist incentive structures and autonomy-based decision architectures that work in the culture they were trained in. They do not work here. The framework is not wrong - it is mismatched.
Why it matters
The universality ratings in this collection (high/moderate/low) flag this risk entry by entry. But the risk of framework-level mismatch - applying the whole structure to a context it was not designed for - is not captured by individual entry ratings. The framework as a whole has cultural and contextual assumptions embedded in its architecture.
Countermeasure
When applying this framework outside your home cultural and professional context, treat all Universality: Moderate and Universality: Low entries as hypotheses requiring local validation rather than established mechanisms. Seek local expertise before deploying countermeasures.
09
The Retrospective Fallacy
What it is Using the bias framework primarily for post-hoc explanation, which is easy and unfalsifiable, rather than for prospective prediction, which is hard and testable.
Failure pattern
After a decision fails, a person reviews it and identifies the biases that caused it. The analysis feels rigorous. But post-hoc bias identification is almost always possible - the framework is flexible enough to explain any outcome after the fact. The explanations feel insightful but do not improve future decisions because they were never tested against prospective prediction.
Why it matters
Post-hoc explanation has zero predictive value if it is not also prospective prediction. A bias framework that is only used retrospectively is a rationalisation framework with more sophisticated vocabulary. The Prediction fields added to entries in this collection exist specifically to address this failure mode.
Countermeasure
Before any significant decision, make explicit predictions: which mechanisms do I expect to be active and what specific errors do I predict those mechanisms will produce? Record them. Evaluate them after. Only this process produces genuine calibration improvement.
10
Therapeutic Intellectualisation
What it is Using the cognitive framework to name and categorise emotional experiences rather than actually process them.
Failure pattern
A person experiencing grief labels it as "negativity bias" and attempts to correct for it. A person experiencing legitimate anger at an injustice labels it as "emotional reasoning" and discounts it. A person with genuine safety concerns labels their anxiety as "availability heuristic overactivation" and suppresses it. The cognitive framework is being used to bypass emotional processing rather than to supplement it.
Why it matters
Emotions carry information. Some of that information is distorted by the mechanisms described in this collection. But some of it is accurate, adaptive and necessary. The framework can distinguish between these only imperfectly. Using it to dismiss all emotional experience as cognitive error removes a functional information source.
Countermeasure
Cognitive bias tools apply to judgment and decision-making, not to emotional experience. If you are using this framework to talk yourself out of an emotional response, the question to ask is not "which bias is this?" but "what is this emotion telling me and is that information accurate?" Seek qualified support for emotional processing; this Lexicon is not a substitute.
The map is where these get applied — 631 named traps, each with its correction.
Open the map →