Why randomness feels less random than it really is
Humans are excellent pattern-finders, which is useful until noise starts wearing a convincing disguise.
Read articleStatistics, uncertainty & scientific instinct
Bayesic Instinct is a home for sharp, accessible and occasionally strange writing about statistics: probability, prediction, scientific evidence, public data, noisy measurements and the ways numbers shape how we understand the world.
Editorial promise
The site is built around a simple idea: most interesting questions involve uncertainty. The goal is not to make numbers look clever. The goal is to make uncertainty readable.
Latest essays
Starter essays are included so the site feels alive immediately. Replace, expand or link these cards as the blog grows.
Humans are excellent pattern-finders, which is useful until noise starts wearing a convincing disguise.
Read articleA simple explanation of why rare events can produce surprising conclusions, even with good tests.
Read articleA practical guide to statistical significance without the ritualistic fog.
Read articleGood forecasts are maps of uncertainty, not magic claims about the future.
Read articleWhy causal questions need structure, assumptions and often more humility than the chart suggests.
Read articleSmall design choices can make the same dataset feel boring, alarming or decisive.
Read articleInteractive statistics
The interactives section includes a configurable coin-flip simulator, a Bayes test calculator, and an uncertainty playground. They are intentionally lightweight so they work on normal hosting.
Open interactivesChange the true probability, number of flips and animation speed. Watch the observed rate wobble and settle.
Adjust prevalence, sensitivity and specificity to see how rare events can fool intuition.
See how sample size changes the width of a simple uncertainty interval.
Topic map
Common mistakes, misunderstood methods and suspicious charts.
Short data-led investigations about sport, health, economics, culture and science.
Priors, likelihoods, evidence updates and better intuition under uncertainty.
Simple browser tools that let readers play with assumptions.
Newsletter
No spam. Just interesting statistics, explainers and interactive experiments.