Benford’s Law

October 26, 2010

Benford’s Law, which was discovered by Simon Newcomb in 1881 and rediscovered by Frank Benford in 1938, states that, for many sets of numbers that arise in a scale-invariant manner, the first digit is distributed logarithmically, with the first digit 1 about 30% of the time, decreasing digit by digit until the first digit is 9 about 5% of the time. Stated mathematically, the leading digit d ∈ {1 … b-1} for a number written in base b will occur with probability Pd = logb(1 + 1/d). Thus, in base 10, the probabilities of the first digit being the number 1, 2, … 9 are 30.1%, 17.6%, 12.5%, 9.7%, 7.9%, 6.7%, 5.8%, 5.1% and 4.6%.

Benford’s Law is counter-intuitive but arises frequently in nature. It is also frequently used in auditing. Make a list of the amounts of the checks that a bookkeeper has written in the past year; if more that 5% begin with the digit 8 or 9, the bookkeeper is likely an embezzler. More important, precinct results of the disputed Iranian elections a year ago displayed anomalous first-digit counts, suggesting vote fraud.

Recently, Shriram Krishnamurthy issued a programming challenge on the Racket mailing list, asking for smallest/tightest/cleanest/best code to calculate the first-digit percentages of a list of numbers; he also challenged readers to apply the function to data in a comma-separated values file. He didn’t give a source, but did mention that he was interested in the littoral area (in acres) of the lakes of Missesota; sample data appears on the next page.

Your first task is to write a function that calculates the first-digit percentages of a list of numbers. Your second task is to calculate the first-digit percentages of the data on the next page. When you are finished, you are welcome to read or run a suggested solution, or to post your own solution or discuss the exercise in the comments below.


Pages: 1 2 3