I am not missing out anything. Those numbers are imaginary data from a "rate this mans looks on a scale from 1 to 100" experiment. The subjects are then given 10 pictures and are asked to rank them. If all test subjects gave rankings like what I said (1, 2, 3, 4, 5, 6, 7, 8, 50, 100 (that means, 8 bad looking, one 'meh' and one perfect)), then 80% of those men are below average looking. If those men were picked from a statistically valid pool of men, then one could infer that 80% of all men are below average looking.
There is nothing wrong with having 80% of a population below average in some property. Another example is this: say there is a group of six people and five of them earn 20 000$ pr year, while the last one earns 100 000$. What is the average income among those people, and how many % are below that average?
A lot of people do confuse the median and the average though, and it is not really that weird