The Most Counterintuitive Number in Daily Life
A quick test — answer with your gut. A condition affects 1 person in 1,000. There's a screening test for it that's quite good: it catches 100% of true cases, and it's wrong only 5% of the time on healthy people. You test positive. How likely is it you actually have the condition? Nearly everyone — in famous studies, including a majority of physicians — answers something near 95%. The real answer is about 2%. Not a trick, not a typo: a positive result from a good test for a rare condition usually means you're fine, and the reasoning gap behind that sentence causes real-world panic, real-world overtreatment, and real-world bad decisions far beyond medicine.
The missing ingredient: how rare the thing is
The intuition failure has a name — base rate neglect — and a simple shape: when judging what a test result means, people stare at the test's accuracy and completely ignore how common the condition was to begin with. But the base rate is doing most of the work: among 1,000 people tested, only 1 actually has the condition — while the 5% error rate, applied to 999 healthy people, produces about 50 false alarms. Your positive result is one ticket in a stack of ~51 positives, of which exactly 1 is real. That's the 2%. The test isn't bad; the thing is rare, and rare things stay improbable even after decent evidence points at them. (You'll recognize the family resemblance to the numbers topic's per-capita rule and the odds topic's rare-events chapter — this topic is their missing sibling, and arguably the most practically consequential of the three.)
Where this one skill pays
The same arithmetic governs an astonishing range of alarms: medical screening results (the mammogram callback, the PSA number, the prenatal screen — Part 3 walks the real cases), fraud alerts and security warnings, background checks and drug tests, spam filters and content flags, the smoke detector's opinion of your cooking — and, stretched slightly, everyday inference itself: what one bad review, one symptom, one alarming headline should actually do to your beliefs (Part 5). The promise of this topic, in the shelf's tradition: no formulas. The entire skill is a counting picture you can draw on a napkin — Part 2 teaches it — plus three questions to ask of any alarm (Part 6). By the end, "positive" will mean to you what it means to the statisticians: time for the follow-up question, not time for the panic.