4 Exercises · Probability & Statistics
The math of uncertainty — applied to the decisions you actually make.
Probability and statistics are where human intuition fails most predictably, from misreading medical test results to seeing patterns in noise. This category trains base-rate thinking, Bayesian updating, and the identification of statistical fallacies in the contexts where they actually cause harm.
Probability and statistics are the formal language of uncertainty, and these exercises translate that language into the everyday reasoning you actually need. You already think probabilistically without noticing — when you hear a friend tested positive for a rare disease, or a candidate is 'leading' in the polls, or a stock is 'down 30% from its peak,' you're doing probability, often badly. The exercises drill the moves that separate accurate judgment from gut feeling: base rates, conditional probability, sample size, regression to the mean, and Bayesian updating.
The single most important idea is the base rate — how common something is before you see any specific evidence. People reliably ignore the base rate in favor of vivid case detail, which produces confident errors in the wrong direction. Beginner exercises cover simple calculations and reading common formats; intermediate ones add conditional probability and the famous rare-disease test puzzle; advanced ones cover Bayesian reasoning and statistical illusions. If math intimidates you, relax — the emphasis is on concepts, not arithmetic.
Stakes
Probability literacy measurably improves health, money, and policy decisions. Doctors and patients who reason in natural frequencies ('10 out of 1,000') make systematically better choices than those reasoning in percentages, because the frequency framing makes the base rate impossible to ignore. Practice these and you reach that framing automatically.
It pays off professionally too. Anyone who interprets data — analysts, product managers, clinicians, journalists — is constantly making probability judgments, and the gap between someone calibrated about uncertainty and someone who isn't is enormous over a career. A few base-rate errors in high-stakes decisions can dwarf every other reasoning weakness combined.
Hazards
The reasoning errors these exercises specifically train against.
Given specific evidence about one case, people fixate on the case and ignore how common the thing actually is. A 95%-accurate test for a 1-in-10,000 disease produces mostly false positives — yet most people guess a positive result means the patient probably has it.
The chance of cancer given a positive test isn't the same as the chance of a positive test given cancer. Swapping those two is the technical heart of base-rate neglect, and the exercises drill the difference until it's automatic.
A 70% success rate from 10 tries is far weaker evidence than a 60% rate from 1,000. People treat percentages as comparable regardless of the sample behind them, which produces huge misjudgments of how strong the evidence is.
Extreme results tend to be followed by less extreme ones, with no causal change at all. Crediting that natural drift to an intervention is the source of many false beliefs about coaching, medicine, and personal habits.
Method
Each exercise presents a probabilistic scene — a medical test, a poll, an A/B result, an investment claim — and asks for the right reading. The wrong answers reflect the canonical errors: base-rate neglect, sample-size blindness, misreading conditionals. The explanations rephrase the problem in natural frequencies, which makes the right answer obvious, and show how the original framing produced the mistake.
Most exercises can be reasoned qualitatively, with rough estimates instead of precise math. The goal is calibrated intuition — glancing at a claim and immediately sensing whether it's plausible — not numerical fluency. That comes later; the conceptual moves come first.
4 Exercises
Start anywhere, finish at your own pace.
Confront the scenarios where human intuition about probability fails most dramatically, from emergency rooms to courtrooms to casinos. These puzzles expose systematic flaws in how our brains estimate likelihood, teaching you to recognize when your gut feeling is being hijacked by cognitive shortcuts. Mastering these foundations will change how you evaluate risk in medical decisions, financial choices, and everyday life.
Develop the skills to interrogate statistical claims you encounter in news headlines, pharmaceutical ads, corporate earnings reports, and political campaigns. These exercises train you to spot the specific techniques -- from axis manipulation to cherry-picked comparisons to survivorship bias -- that make misleading numbers look convincing. You will learn to ask the right questions before accepting any statistical claim at face value.
Master the framework that doctors, intelligence analysts, and data scientists use to update beliefs rationally when new evidence arrives. These exercises build your intuition for Bayes' theorem through scenarios involving criminal investigations, medical diagnosis, A/B testing, and geopolitical analysis. You will learn when evidence should dramatically change your mind versus when it should barely shift your confidence.
Identify the most dangerous statistical fallacies that lead to wrongful convictions, failed policies, wasted research funding, and medical harm. These advanced scenarios test whether you can spot subtle errors involving Simpson's paradox, the prosecutor's fallacy, multiple comparisons, selection bias, and expected value traps that regularly fool judges, journalists, scientists, and executives.
In the wild
Deeper look
Probability is where human intuition fails most consistently and most expensively. Doctors misread test results in ways that change treatment; juries misjudge match-probability evidence in ways that change verdicts; investors overweight rare events and underweight cumulative ones in ways that wreck portfolios. These aren't signs of stupidity — they're predictable outputs of a mind built for small-sample reasoning that now has to handle large-scale data and complex risk.
One robust finding shapes the whole category: the same problem that most physicians get wrong when it's phrased in percentages, most of them get right when it's phrased in natural frequencies. The fix — translate into frequencies, draw the tree, apply Bayes' rule explicitly — is what produces reliably correct answers. And the same handful of moves span wildly different domains: the base-rate translation that fixes the mammogram problem also fixes misread DNA evidence in court; the cumulative-probability calculation that stops a mayor from dismissing a 2% annual flood risk also stops a homeowner from dismissing the odds an appliance fails over its lifetime.
FAQ
Sources
Primary sources and reputable references for the concepts covered above.
On natural-frequency framing and how to reason clearly about probabilistic risks.
Scholarly treatment of Bayesian reasoning as a theory of belief revision.
The classic introduction to common statistical deceptions; still directly relevant decades after publication.
On calibrated probabilistic thinking applied to forecasting, drawing on the largest study of prediction accuracy ever conducted.
