I read it as the GP saying there's often confusion about assuming behavior about individuals as opposed to making predictions based on trends of large groups of individuals.
For some population, you can safely state that some portion of them will contract appendicitis. You cannot make that same assertion about an individual person. This likewise carries to specific behaviors (theft, charity, becoming a pet owner, etc).
The same is true of say washing machine motors. You can predict that ~ 10% of them will fail in a certain amount of time (and even go into how they fail), but that doesn't tell you much about a specific motor. Or sports events, if you say there's a 25% chance of team A winning a single match; the results of the match don't support or refute your prediction... you'd have to run many matches to see if your prediction was empirically correct.
