I’m not sure how I’ve not stumbled across Benford’s Law before, but I haven’t — it sure seems like the kind of mathematical trivia that’s right in my happy place. The law states that given a list of numerical data from real-world sources (i.e., baseball statistics, street addresses, Dow Jones averages, tax return amounts), the first digit of a number in the list is more likely to be 1 than any other digit (specifically, a 30.1% probability), and there are specific probabilities for each other digit as well. The law can be used to look for fraudulent sets of data — for example, if tax return data doesn’t follow the probabilities specified by the law, it has a much higher likelihood of being falsified.

Rex Swain republished the Times article along with some enlightening charts that help illustrate the law, and of course, Wikipedia has more info. And finally, there’s a Java tool that can help you analyze your own data sets against Benford’s Law… just be forewarned that data that’s truly uniformly distributed won’t adhere to the law.