A working atlas of the 186 ways a mind bends what it sees.

We forget intentions when attention splits — lapses of attention, not storage failure.

We attribute others' actions to character and our own to circumstances.

We prefer known risks over unknown ones — ambiguity feels like danger.

The first number we see shapes every number we judge after it.

A single vivid anecdote is mistaken for representative evidence of a broader truth.

We attribute human traits, intentions, and emotions to non-human systems and objects.

We treat novelty as evidence of superiority without proof it works better.

We treat what could happen as if it will happen, overweighting unlikely outcomes.

We reject a conclusion because the argument supporting it was flawed — even when the conclusion holds.

Recurring thoughts and worries quietly steer what we see.

We overweight opinions from authority figures regardless of the underlying evidence.

We over-trust automated outputs and under-scrutinise what algorithms tell us.

Things easy to recall feel more common, true, or important.

Corrective information sometimes strengthens the mistaken belief it was meant to fix.

We adopt beliefs and behaviours because others around us already have.

Vague, flattering statements feel uniquely accurate — so we trust the source.

Vivid cases override base-rate statistics when we judge what is likely or typical.

We judge arguments by whether we believe the conclusion — not whether the logic holds.

We catch bias in others while believing ourselves clear-eyed.

Teams spend disproportionate time on trivial issues because they are easy to understand.

Bizarre or surprising details are remembered better than ordinary ones — and mistaken for the norm.

We often miss large changes in our visual field when our attention is elsewhere.

Items appear more attractive in a group than when viewed individually.

We retroactively upgrade choices we have already made, remembering them as better than they were.

Random patterns look meaningful when we squint hard enough.

We invent plausible explanations for behaviour and memories without realising the story is constructed.

We notice and remember evidence that fits what we already believe.

We prefer testing one favoured hypothesis over comparing plausible alternatives.

Vivid combined scenarios feel more likely than simpler, more probable ones.

New evidence updates our beliefs more slowly than Bayesian logic would predict.

We recall things best in the environment we learned them in.

Debunked misinformation continues to influence reasoning even after people accept the correction.

Perception of a stimulus shifts depending on contrasting items seen immediately before or beside it.

People recognise and remember faces of their own racial group more accurately than others.

We mistake remembered information for original thought — source memory fails.

Once we know something, we can't imagine not knowing it.

We believe things are worse than before — past rose-tinted, present judged harshly.

A third, worse option makes the option next to it look brilliant.

Most users accept the default — make defaults a moral choice.

We blame victims more when we feel different from them — to maintain a sense of safety.

We prefer comparing easy attributes over hard ones — even when the hard choice matters more.

We spend smaller denominations more readily than the same value in larger units.

We sell winners too early and hold losers too long — anchored to original cost or effort.

Side-by-side comparison exaggerates minor differences between similar options.

Novices overestimate their skill; experts underestimate theirs.

We neglect how long an experience lasted — peaks and endings dominate memory.

We over-value what we built ourselves, however wobbly.

We overestimate how much others share our perspective, knowledge, and preferences.

We can't accurately predict how we or others will feel in a different emotional state.

We value things more highly simply because we own or possess them.

We increase commitment to a failing course of action because of prior investment.

We treat categories as having fixed essences — ignoring variation and context.

Expectations distort observation: we notice and interpret evidence to fit what we already believe.

Researchers unconsciously influence experiments toward expected outcomes through subtle cues and procedural choices.

We overestimate how much extrinsic rewards shape long-term motivation and underestimate intrinsic interest.

Negative emotion attached to memories fades faster than positive emotion.

We overestimate how much others share our opinions and behaviours.

We can remember events that never happened — confidently and in detail.

People overweight attributes that are prominent in discussion or comparison while neglecting others that matter in use.

People accept vague, general statements as uniquely accurate descriptions of themselves.

Same fact, different frame, different decision.

After noticing something once, we suddenly see it everywhere.

We perceive objects and patterns only in their familiar function, blocking novel uses and solutions.

We blame users' character; we excuse our own context.

We expect random sequences to self-correct — they don't.

Information we generate or assemble ourselves is remembered better than information we passively receive.

We forget information we believe we can look up later — transactive memory with technology.

We judge entire groups based on the traits or behaviour of a single salient member.

One impressive trait makes us assume the rest is excellent too.

Confidence calibration varies with task difficulty — overconfidence on hard tasks, underconfidence on easy ones.

Once we know the outcome, we believe we predicted it all along.

We believe a person or process on a successful streak is more likely to succeed again — independent of actual odds.

Humorous information is remembered better than non-humorous information.

We prefer a smaller reward now over a larger reward later.

We respond more strongly to a single identifiable victim than to large numbers of anonymous sufferers.

We believe we understand others better than they understand us — and better than they understand themselves.

We overestimate our ability to control outcomes that are actually determined by chance or external forces.

We attribute human-like intent and agency to automated systems, coincidences, and non-agentic processes.

We overestimate how clearly our emotions, intentions, and knowledge are visible to others.

We hold high confidence in predictions from weak evidence when those predictions feel coherent.

We perceive relationships between variables that are unrelated or only weakly related.

We tend to rate ourselves as above average on desirable traits and skills.

Repeated exposure makes a claim feel more credible — even when it's false.

We overestimate the intensity and duration of emotional impact from future events.

Automatic mental links between concepts steer perception before conscious thought.

We design for people who look, think, and click like us.

We miss unexpected things in plain sight when our attention is elsewhere.

We seek additional information even when it cannot improve the decision — mistaking activity for progress.

We treat small samples as representative, underestimating how much variance shrinks with scale.

We escalate commitment beyond rational limits — especially under competition, sunk costs, or public stakes.

We believe people get what they deserve — so we blame victims of bad outcomes to preserve a fair-world belief.

Groups systematically over-report performance above statistical norms — 'all the children are above average.'

Lesser options are preferred in isolation but rejected when compared directly to more complete alternatives.

Memory retellings flatten some details and exaggerate others with each repetition.

Deep, meaningful processing produces stronger memories than shallow visual exposure.

Longer lists increase recalled items but decrease recall proportion — more total, less per item.

Losing $10 stings more than winning $10 cheers — by about double.

Miller's 7±2 chunking limit is overgeneralised as a universal cap on menu items, choices, and list length.

We treat an entity as unknown or distinct when only its description changes — confusing labels with identity.

Retrieving one memory suppresses competing related memories — retrieval-induced forgetting.

We treat money and value differently depending on which mental account we assign it to.

Familiarity breeds preference — we like what we've encountered before.

We remember information but misattribute its source — wrong screen, brand, or context.

Post-event information silently rewrites what we remember.

Memory performance differs by presentation mode — auditory versus visual material is recalled differently.

People think in nominal monetary values rather than real purchasing power or relative cost.

Current mood biases recall toward mood-matching memories — sad moods retrieve sad episodes more easily.

Prior good deeds make us feel licensed to act less ethically or rigorously afterward.

We judge the morality or quality of actions by their results, not only the intentions and process behind them.

We overweight the expectation that anything that can go wrong will go wrong — planning and judging from worst-case certainty.

We turn messy data into clean stories — and trust them too much.

People assume others' actions are driven by self-interest and manipulation more than they actually are.

People believe they perceive reality directly and that those who disagree are uninformed, irrational, or biased.

Bad outcomes register louder and stick longer than good ones.

We disregard probability when making decisions under uncertainty, responding to vivid outcomes instead.

People recall less of what came before them when they are next to perform or speak in sequence.

We underestimate the likelihood and impact of disasters and disruptive change, assuming things will continue as usual.

We dismiss ideas, methods, or solutions that originated outside our group in favour of internal ones.

People modify behaviour when they know they are being observed.

Researchers' expectations unconsciously influence participants' behaviour and results.

We favour simple explanations over complex ones — even when simplicity misleads.

Harm caused by inaction is judged less harshly than equal harm caused by action.

We systematically overestimate positive outcomes and underestimate risks to ourselves.

People avoid information that threatens their self-image or comfort — figuratively burying their heads.

We see members of other groups as more alike than members of our own group.

We judge a decision's quality by its outcome rather than by the information and process available at the time.

We systematically overestimate the accuracy of our beliefs, predictions, and abilities.

More options feels like freedom but produces paralysis and regret.

We perceive familiar patterns — especially faces and agency — in random or ambiguous stimuli.

Presenting some items from a list as cues impairs recall of the remaining uncued items.

We remember the peak moment and the ending — most of the rest blurs.

Safety improvements can lead to riskier behaviour that partially or fully offsets the benefit.

We overestimate the likelihood and impact of negative events relative to positive or neutral ones.

Images are remembered far better than the words next to them.

Expectations of benefit produce real improvements in experience and performance independent of the intervention's objective power.

We underestimate how long, costly, and risky our plans will be.

Older adults and retrospective recall overweight positive information relative to negative in memory.

After committing to a choice, people inflate benefits and downplay flaws to reduce cognitive dissonance.

We overvalue benefits of innovation and undervalue adoption costs, risks, and legitimate status quo advantages.

Information that requires more cognitive effort to process can be remembered better — contrary to fluency intuitions.

We assume tomorrow's user will want what today's designer wants.

We respond to partial certainty as if it were complete — depending on how choices are framed.

Tell people they can't have something and they want it more.

We devalue proposals and features because of who proposed them, not what they contain.

After noticing something recently, we believe it increased in frequency or is novel — when only our attention changed.

We overestimate our ability to resist temptation and impulses in future tempting situations.

Forbidden or discouraged actions become more attractive — reactance-driven reversal.

Rhyming statements are judged as more accurate and truthful than non-rhyming equivalents.

People adjust behaviour when perceived risk changes — often offsetting safety gains.

Past experiences sweeten in memory — interviews lie a little.

Few left, ending soon — pressure cuts deliberation short.

Expectations filter what users actually see on a screen.

We maintain consistency with prior beliefs, statements, and choices despite new contradictory evidence.

Information related to the self is processed more deeply and remembered better.

We credit successes to our skill and failures to bad luck or outside forces.

New evidence that contradicts established norms or identity is reflexively rejected.

First and last items in a list outlive everything in the middle.

Recall of lists follows serial position patterns — order and position strongly affect what is remembered.

We evaluate ourselves and our work by comparing to others — distorting judgment and motivation.

We respond in ways that make us look good to others — distorting surveys and interviews.

People confuse where they learned information — source memory fails while content memory persists.

We learn better in spaced doses than in a single cram session.

We overestimate how much others notice our appearance, errors, and performance.

The current setup feels safe even when a better one is one click away.

Group labels stand in for individuals — convenient and wrong.

We judge the whole as less than the sum of its parts — or misestimate combined value and probability.

People accept vague or general statements as personally meaningful and accurate.

Appending irrelevant suffix items to a list impairs recall of the final list items.

Memory and judgment are shaped by leading questions and suggestions — people incorporate suggested details into recall.

We finish what we started — even when starting over is cheaper.

Studying only what survived hides the lessons of what didn't.

We tend to defend existing social systems and status quo — even when they disadvantage us.

We misremember the timing of events — recent ones seem more distant, distant ones seem more recent.

Active retrieval strengthens memory more than passive review of the same material.

We believe media and persuasion affect others more than they affect us.

We misjudge time savings — overweighting small gains at high baseline speed, underweighting gains at slow speeds.

People strongly feel they know a word or fact but cannot retrieve it at that moment.

We view ourselves as flexible and others as having fixed, stable traits.

We attribute negative out-group behaviour to disposition and positive in-group behaviour to situation — distorting research and support.

We tend to want one unit of consumption — a portion, serving, or package — regardless of its size.

The odd-one-out is the one users remember.

Perceived intensity of a stimulus is proportional to the logarithm of its physical magnitude — we notice relative change, not absolute.

We perceive familiar routes as shorter and faster than unfamiliar ones of equal objective length.

Incomplete tasks stay active in memory — finished ones fade quickly.

We prefer eliminating a small risk entirely over significantly reducing a large one.

We perceive situations as zero-sum — one party's gain must be another's loss — when positive-sum outcomes are possible.