In many collections of data, a given data point has roughly a 30% chance of starting with the digit 1.