Take a set of numbers collected from “the wild”. You might take company profits, city population statistics, street numbers or odometer readings. How many of the numbers in your set start with “1″? how many with “2″? with “3″?, “4″? …
Amazing as it sounds, these leading digits are NOT evenly distributed but follow a pattern (30% “1″s, down to less than 5% “9″s). This pattern is so reliable, it’s been used in fraud detection, to trap companies cooking their books: the invented numbers did not follow Benford’s Law.
Follow the link for a fascinating explanation of why this is so.
Cute video featuring nine ways to visualise 1 in a million. Interesting that five of the nine ways involve the use of an intermediate grouping. For example, 100 people have an average of 100,000 hairs on their heads. 10 hairs represents 1 in a million. 4 of the comparisons are direct, eg 1 word in all of the text of the Harry Potter novels.