The Presence of Ruin: Why You shouldn’t Drive a Motorbike

The blind spot of statistics

Suppose you’re interested in a new pill against headaches. You set up a double-blind experiment, get a few thousand participants, separate them into two groups (real pill and placebo) and run the test. Then you analyse the results and of all of the people who didn’t receive a placebo, 60% report their headache gone after the pill.

Now you’re the one having a headache. Is it right to assume that there is a 60% chance this pill would heal your pain?