Probability & Statistics
Basic Probability Intuition
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.
Context
Why this exercise
Probability is the area where human intuition fails most dramatically and most consequentially. Doctors mis-interpret diagnostic test results in ways that change treatment decisions; juries misjudge DNA-match probabilities in ways that change verdicts; investors overweight rare events in ways that destroy portfolios. The good news is that the failures are systematic — they follow predictable patterns identified by Kahneman, Tversky, Gigerenzer, and others — and can be corrected with deliberate practice. This exercise drills the foundational moves: thinking in base rates, distinguishing conditional from unconditional probabilities, and recognizing when your gut feeling is being hijacked by cognitive shortcuts.
Before you start
Probability theory in its modern form was developed by Pascal and Fermat in the 17th century to analyze games of chance, formalized by Pierre-Simon Laplace, and put on rigorous axiomatic foundations by Andrey Kolmogorov in 1933. The mathematical apparatus is straightforward: probabilities are numbers between 0 and 1, conditional probabilities follow the multiplication rule, independent events combine multiplicatively, and Bayes' theorem governs how to update beliefs based on new evidence. What makes probability hard for humans is not the math but the translation between the math and natural language: 'the probability that a person who tests positive actually has the disease' sounds like it should equal 'the probability that a person with the disease tests positive,' but they are different quantities related by Bayes' theorem and the prior base rate.
Several specific intuition failures recur often enough to be worth memorizing. Base-rate neglect, documented by Kahneman and Tversky, causes people to ignore prior probabilities when evaluating evidence — the famous mammogram problem produces wrong answers from most physicians because the base rate of breast cancer in screened populations is low enough that even a fairly accurate test produces more false positives than true positives. The conjunction fallacy, illustrated by the 'Linda the bank teller' problem, makes people judge a specific conjunction ('Linda is a bank teller AND a feminist activist') as more probable than a single conjunct ('Linda is a bank teller'), which violates a basic law of probability. The gambler's fallacy treats independent events as if they have memory ('red has come up five times, so black is due'). And the hot-hand fallacy treats independent random sequences as if they exhibit positive autocorrelation that they actually do not.
The procedural fixes are well-established. For diagnostic problems, draw out the natural-frequencies tree (Gerd Gigerenzer's preferred representation): imagine 1,000 people, compute how many have the disease at the base rate, how many test positive given the disease, how many test positive without the disease, and read off the answer by counting. For independence problems, ask whether knowing one event would change the probability of another; if not, they are independent and the gambler's-fallacy intuition is wrong. For Bayes problems, write down the prior, the likelihood, and the posterior explicitly. As you work the scenarios, practice translating verbal probability claims into structured mathematical form before trusting your intuition. For deeper treatment of how this connects to broader scientific reasoning, see Scientific Thinking.