Archive
Everything we have published, newest first — 100 articles.
August 2026
Base rate neglect: why a vivid detail beats a background number
Survivorship bias is why advice from people who succeeded is mostly noise
Hindsight makes a coin flip look like a plan
More options can make a choice worse, but the evidence is messier than the…
Opportunity cost is the price that never appears on the tag
Absolute and relative risk describe the same event and feel nothing alike
Anchoring survived the replication wars, but its explanation did not
Confirmation bias is a search problem before it is a belief problem
Why money already spent keeps getting a vote
July 2026
A decision can be good and still turn out badly
Reversible and irreversible choices deserve different amounts of thought
Very small probabilities behave badly in the mind
A test that is right most of the time can still be mostly wrong
Expected value stops being good advice when the loss is permanent
Risk and uncertainty are not the same problem
What a control group is actually for
Regression to the mean makes useless interventions look like they worked
Before you call it causation, three other explanations have to fail
Why a shelf of famous psychology findings quietly stopped being true
Analysis has too many degrees of freedom, which is why preregistration…
June 2026
A habit is a cue and a routine, and the cue is the part you can move
What the research on habit formation actually supports
If-then plans are the least glamorous thing that keeps working
Friction decides more than motivation does
A streak is a motivation device with a failure mode built in
The outside view asks how things like this usually go
What comes to mind easily gets treated as common
The same option described as a loss or a gain stops being the same option
One strong impression contaminates every judgement that follows
Confidence tracks how easily an answer arrives, not how often it is right
Removing your own options is sometimes the strongest move available
When the clock is the binding constraint, cut analysis before you cut…
Knowing when to stop looking is a separate decision from what to pick
May 2026
Most decisions are posed as yes or no when they were never that shape
A group decides worse than its members when everyone talks first
The best and worst performers are usually just the smallest ones
Making something safer changes how people use it
Risks that look independent tend to arrive together
Two risks can’t be compared until they share a denominator
The average outcome is often one that nobody actually gets
A prediction that cannot be scored was never a prediction
When two experts disagree, the disagreement is the information
A single number is a claim about precision it usually cannot support
April 2026
Every finding comes attached to the people it was found in
The thing being measured is rarely the thing anyone cares about
Experience only teaches when the feedback has three properties
Practice explains less of expert performance than the famous number implied
Goals set on outcomes leave you nothing to do on Tuesday
Paying for something people already wanted to do can backfire
Switching between tasks costs more than the interruption itself
Character is the first explanation offered and the least examined
Motivated reasoning and plain ignorance produce the same wrong answer
Ask someone to explain a mechanism and the understanding evaporates
Ten thousand and a million feel the same when a decision is being made
March 2026
The idea that a good deed licenses a bad one has thinner support than its…
A bright line and a rule of thumb fail in opposite directions
Doing nothing feels safer than acting, and it is not the same as being safer
The future gets a discount, and the rate is not constant
Decision fatigue is a plausible idea resting on contested evidence
Anticipated regret is a real input and a poor master
Why the other queue moves faster and your friends have more friends
Every alarm sets a trade between two kinds of error
February 2026
A risk that is negligible for you is certain for someone
The hot hand was debunked and then the debunking was corrected
A near miss is data, and it is almost always read as reassurance
A natural experiment borrows its randomness from circumstance
A single story is evidence of possibility, not of frequency
A theory that explains everything afterwards predicts nothing beforehand
Pooling studies fixes the sample size and not the bias
When absence of evidence is evidence of absence, and when it is not
Running an experiment on yourself without fooling yourself
January 2026
Keeping a change going is a different problem from starting one
Spacing and self-testing are the findings that survived
Identity is a popular lever with a thin evidence base
Habits appear to spread between people, and proving it is very hard
Round numbers and precise ones are read as two different kinds of claim
A reason and a rationalisation are indistinguishable from the inside
December 2025
A model that was right once stops being right without announcing it
Repetition makes a statement feel true, including one you know is false
The famous curve about incompetence may be mostly a measurement artefact
Deciding for someone else produces a different answer from deciding for…
Many preferences are assembled at the moment of choosing rather than looked…
A crude model usually beats an expert combining the same information by feel
Some trade-offs get refused rather than answered, and the refusal is doing…
A forecast and a plan are different documents and should not share a number
A more detailed scenario is always less likely and usually more convincing
A probability about a single event needs a reference class before it means…
November 2025
Every precaution introduces a risk of its own somewhere else
A one-in-a-million figure was calculated, not observed
Why some risks frighten more than their numbers warrant, and when that is…
The wording of a question is part of the answer it produces
A relationship measured across groups can reverse inside them
A second opinion is worth exactly what its independence is worth
A number can become a fact by being cited often enough
Noise in a measurement is not neutral, and it bends results in a known…
October 2025
Expertise transfers between domains far less than the people holding it…
A checklist protects against forgetting, not against being wrong
Sleeping on a decision helps, though not for the reason usually given
The feedback about you is filtered long before it reaches you
Skills decay at very different rates, and which ones will is predictable