Why So Many Numbers Start With One
The puzzle
Take the populations of every town in a country, or the lengths of every river, or the figures in a set of accounts. Look at the first digit of each.
The digits are not equally common: 1 appears about 30% of the time and 9 about 5%. Why?
Stuck?
One at a time, smallest first. Open only as many as you need — the whole point of the puzzle is the part you work out yourself.
Hint 1
Think about how long a growing quantity spends with each leading digit.
Hint 2
Going from 100 to 200 is a doubling; going from 800 to 900 is a 12.5% rise.
Show the answer Hide the answer
Because the data spans orders of magnitude and the leading digit is decided on a logarithmic scale. The probability of a leading digit d is log10(1 + 1/d), which is 30.1% for 1 and 4.6% for 9.
How to get there
Imagine a quantity growing steadily. To get from a leading 1 to a leading 2 it must double. To get from 9 to the next 1 it need only rise by about 11%. So it spends far longer with a leading 1 than with a leading 9.
Measured on a logarithmic scale, the stretch belonging to digit d runs from log(d) to log(d+1), a width of log(1 + 1/d). For d = 1 that is log10(2) = 0.301; for d = 9 it is log10(10/9) = 0.046.
It applies when data spans several orders of magnitude and has no natural scale - populations, river lengths, physical constants, accounting figures. It does not apply to heights or IQ scores, which cluster. Tax authorities use it to flag fabricated figures, because people inventing numbers spread the leading digits evenly.
The technique worth keeping
When a quantity spans orders of magnitude, reason on a logarithmic scale: equal ratios, not equal differences, are what is uniformly distributed. The leading digit then follows from the width of each interval in log space.