Originally posted by T1000
This one's been giving me a headache. Not a puzzle or a question, just interesting. Feels counterintuitive ... I mean, [b]30%! Harrumph.
Not really - Benford's law makes the assumption that the logarithm of the numbers is in some sense uniformly distributed, so you'd expect small numbers to appear much more often than large ones.
What is surprising is that this is a reasonable assumption to make - 'real-life' distributions can be modelled as all sorts of things, and yet it seems that many statistics one might collect are fairly evenly spread over a certain range of orders of magnitude.