The set-up: a town has 85 green taxis and 15 blue. A witness to a night-time accident says the taxi was blue, and tests show the witness gets colours right 80 percent of the time. How likely is it that the taxi was blue?Most people say 80 percent. The teacher draws 100 taxis as dots and counts. The witness would call 12 of the 15 blue ones blue, and would also call 17 of the 85 green ones blue. So 29 taxis get called blue and only 12 of them are. About 41 percent.The line I saved it for: the witness is not the only evidence. The town is evidence too, and it was there first.Generic clip, no creator noted. Saved on the tram home from lecture 4, on the day the launch should have happened, because the dots did in a minute what the textbook's medical test had not done in a chapter. It is the picture I wish 🎓 Lecture 3 - Probability and base rates had opened with. Filed beside 🔤 Base rate and 📐 Bayes' rule, under 00. Statistics for Product Decisions I Hub.
