Brandolini's
Law
The effort needed to refute a misleading claim can greatly exceed the effort needed to produce it. The asymmetry comes from evidence retrieval, source evaluation, causal explanation, uncertainty, distribution, and the lingering influence of a memorable first account.
definitions
denominator
time window
comparison
uncertainty
causality
context
correction
distribution
Price the claim. Then price a responsible answer.
The model below makes the work visible. It is not a measured universal ratio. Verification effort rises when claims combine multiple assertions, inaccessible evidence, causal language, and high correction standards; reach and emotional novelty affect the response window.
A bundled, high-pull claim with weak source access demands triage. Correct the consequential core first; do not mirror every detail.
One vivid claim
Context collapses
Sources + experts
Smaller audience
An asymmetry of burden, not a scientific constant.
In January 2013, Italian software practitioner Alberto Brandolini formulated what became known as the "bullshit asymmetry principle": refuting nonsense may require an order of magnitude more energy than producing it. The phrase is an aphorism. "Order of magnitude" communicates a striking imbalance; it is not a validated ten-to-one coefficient.
The useful insight is procedural: truth-seeking bears burdens that careless communication can skip. A responsible answer needs evidence, accurate scope, uncertainty, and a causal account. It may also need to reach people who saw the original claim and update a story already encoded in memory.
A correction competes not only with a sentence, but with its speed, simplicity, emotion, repetition, and place in an audience's existing model of the world.
False certainty can skip expensive steps.
Records may span papers, databases, archives, experts, languages, and unavailable primary material.
A short assertion can bundle facts, definitions, time ranges, causal implications, and value judgments.
A credible response distinguishes known, unknown, misleading, unsupported, and false rather than replacing one overclaim with another.
Misinformation can continue to influence reasoning after correction, a pattern studied as the continued influence effect.
The original travels through the audience that found it compelling. A later correction may not reach the same people or receive the same ranking.
Sources, groups, and prior attitudes affect acceptance; factual detail alone may not resolve distrust or motivated reasoning.
Refute propositions, not rhetorical fog.
"The new system doubled failures."
The shareable surface.
What counts as failure?
Which system, period, population, denominator, baseline, and data source?
Did the system cause the change?
Or did reporting, volume, case mix, exposure, or measurement change?
Dates, quantities, quotations, events, and document contents.
Terms such as "failure," "crisis," "normal," or "effective" need operational meaning.
Confounders, selection, timing, and alternative mechanisms matter.
Forecasts need assumptions, ranges, base rates, and evaluation dates.
Evidence can inform consequences but cannot by itself choose every moral objective.
A responder may need to expose the structure rather than chase endless supporting details.
Lead with fact, explain the gap, supply a replacement model.
Give readers a usable conclusion before repeating the false claim.
Identify the misleading element without amplifying unnecessary language or imagery.
Name the wrong denominator, missing context, altered image, unsupported inference, or unreliable source.
A correction is more durable when it provides a coherent accurate explanation, not only a negation.
Link primary records, disclose uncertainty, date the assessment, and distinguish evidence from inference.
"No. This is fake."
"The image is from a 2018 drill, not yesterday's event. The original archive record and matching frame are linked here; current official imagery shows a different location."
The cheapest correction may happen before exposure.
Teach manipulation techniques.
Brief warnings and examples can help people recognize impersonation, false dilemmas, emotional manipulation, or conspiracy-style incoherence.
Slow impulsive sharing.
Prompts to read, inspect a source, or confirm an old article's date can interrupt low-attention transmission.
Preserve where media came from.
Signed metadata, source history, and accessible originals make later verification faster, though provenance does not prove every claim true.
Leave the page to check the page.
Professional fact-checkers investigate the source, reputation, and external coverage before trusting polished on-page cues.
Reuse verified work.
Structured claim reviews, source packs, and correction archives reduce repeated search costs across teams.
Prepare before predictable events.
Preapproved experts, monitoring, and clear escalation shorten the gap between a high-impact claim and a careful answer.
Different media create different verification burdens.
Twelve words, several archives.
A quote card names a famous scientist but provides no date. Verification requires searching primary texts, variants, quote histories, and reliable editions. Absence from search is not absolute proof it was never said.
RESPONSE: label unsupported; document the search boundaryA true number with a false frame.
The plotted value may be accurate while a truncated axis, cherry-picked start date, changed denominator, or missing comparison drives the visual conclusion.
RESPONSE: reproduce the data with full definitions and contextAnecdote outruns population evidence.
A vivid personal story can be published instantly. Assessing general risk may require diagnosis quality, base rates, comparison groups, study design, adverse-event surveillance, and uncertainty.
RESPONSE: acknowledge the experience; separate case from ratePixels are only one evidence stream.
Visual artifacts can help but are not decisive. Source provenance, earliest upload, location, shadows, weather, reverse search, and independent records form a stronger verification chain.
RESPONSE: show the chain of custody, not detector confidence aloneNot every bad claim deserves a full public rebuttal.
Build an evidence pack and response threshold without unnecessarily amplifying the claim.
Correct the consequential core, use trusted messengers, link evidence, and update as facts change.
Archive if useful; direct attention to higher-value communication.
Use a short factual clarification, community note, or reference page.
Triage should consider expected harm, reach, audience vulnerability, source credibility, novelty, evidence readiness, and whether responding will increase exposure. Silence, monitoring, prebunking, private contact, and public correction are distinct options.
The law can itself become a shortcut.
"Claims I dislike are bullshit."
The label is not evidence. Decompose the claim and apply the same sourcing standard to allies and opponents.
"Refutation is always harder."
A direct primary record, reproducible test, or obvious contradiction can make some corrections cheap.
"A long answer is therefore true."
Length and effort do not guarantee accuracy. Sophistry can also be expensive and elaborate.
"Ten times is a measured constant."
The original phrase is rhetorical; costs vary by claim, evidence, standards, audience, and distribution.
"Debunk everything."
Public correction can amplify obscure claims and exhaust limited verification capacity.
"Facts alone always persuade."
Trust, identity, values, source relationships, and alternative explanations affect belief revision.
A 2013 aphorism met an older science of correction.
Publicly formulates the asymmetry principle after reading about fast and slow thinking and watching a contentious political discussion.
Experimental psychology documents how retracted misinformation can remain available in memory and continue shaping inference.
Research moves from merely repeating corrections toward warnings, explanations, alternative accounts, trusted sources, and audience-aware design.
Inoculation interventions, lateral reading, provenance, and platform design broaden attention from correction to prevention.
Sources and further reading.
The aphorism's origin is distinguished from empirical research on misinformation and correction.
- Alberto Brandolini (2013) - Original public formulationThe January 2013 post that introduced the named asymmetry principle.Primary historical source
- Stephan Lewandowsky et al. (2012) - Misinformation and Its CorrectionA major review of continued influence, memory mechanisms, worldview effects, and evidence-based correction strategies.Psychological Science in the Public Interest, 13(3)
- Ullrich Ecker et al. (2022) - The psychological drivers of misinformation belief and its resistance to correctionA comprehensive review of cognitive, social, and affective drivers and practical intervention principles.Nature Reviews Psychology, 1, 13-29
- Briony Swire-Thompson, Joseph DeGutis & David Lazer (2020) - Searching for the Backfire EffectA review cautioning against overgeneralizing claims that corrections routinely strengthen misinformation.Journal of Applied Research in Memory and Cognition, 9(3), 286-299
- Jon Roozenbeek et al. (2022) - Psychological inoculation improves resilience against misinformation on social mediaLarge-scale evidence on brief prebunking videos that teach common manipulation techniques.Science Advances, 8(34)
- Digital Inquiry Group - Civic Online ReasoningResearch-informed materials for lateral reading, source investigation, and evaluating online claims.Stanford-originated education research program