| A Short Thought
A very small difference among average values can turn into a big difference in the "tails" of the normal distribution. For instance: If the average 100 yard dash for a guy with dark hair is 12.3 seconds, and a redhead is 12.1 seconds, then a random dark haired guy is maybe 45 percent likelier to beat a random redhead in a race. But if I knew the winner broke 10 seconds flat, I could probably lay 5-to-1 odds that it was the redhead that won. (It is for this reason that people form stereotypes when there is only a modicum of truth to them.) | 
