[who_would_fardels_bear]

If you are slightly faster than all other runners you will tend to win every race ... not just slightly fewer.

If your business model is slightly better than that of all your competitors you will eventually get most of the customers ... not just a few more.

If you look at undirected things such as average heights of humans then you will get average bell-curve results.

If you look at directied things such as people wanting to pay the lowest price or companies wanting to run at the lowest cost, then you will often get non-linear feedback.

If taller people were generally more desirable and it was relatively easy to control your height, then not even height would not follow a bell-curve.

[driftless2]

"slightly faster"

A point I have been making for years. For a long time people have assumed that dominant athletes, for example, are far superior to their competitors when in fact, like Tiger Woods, they are only maybe a few percentage points better if that. But those few percentage points translate into dominance. Nevertheless even Woods doesn't win every tournament. If it can be measured accurately that I am five percent better than my opponent in let's say tennis, I will thrash him virtually every time. Even being just one percent better would allow me to win better than two thirds of my matches.

[Myrddin]

It's mostly a function of sample size and an adequately random sampling method. If you impose any kind of selectivity, you are certain to skew the distribution. Las Vegas (or any gambling establishment) understands how to play the odds in a subtle manner. The overall odds are always in favor of the house. They allow enough "black swans" to keep the rest of the fools filling their pockets.