4.5/5 - (48 votes)

In 1881, astronomer Simon Newcomb noticed something odd about his own library: the first few pages of a book of logarithm tables, the ones starting with the digit 1, were far more worn and smudged than any of the others. That small, physical observation turned out to describe a pattern hiding inside huge swaths of real-world data, decades before it got the name we use for it today.

The pattern everyone credits to the wrong man

Most people who have heard of it know it as Benford’s law. But Frank Benford, the physicist whose name it carries, was not the one who found it. Newcomb spotted the pattern in 1881, a full 57 years before Benford ever published on the subject.

Newcomb noticed that in his logarithm tables, the digit 1 turned up as a leading digit far more often than any other number. He wrote up the observation, and then the discovery mostly sat forgotten for over half a century.

What the pattern actually says

In enormous collections of real-world numbers, the very first digit is not evenly spread out. The digit 1 shows up as a leading digit about 30 percent of the time. The digit 9, by contrast, appears in that position in less than 5 percent of cases.

This shows up in electricity bills, stock prices, river lengths, city populations, and even fundamental physical constants. The reason comes down to scale: when values are spread across a logarithmic scale rather than a straight linear one, low digits simply have more room to lead. The pattern holds in other number bases too, not just the decimal system most people grew up with.

See also  5 famous correlations that have nothing to do with cause and effect

How Frank Benford stumbled onto it, independently

In 1938, physicist Frank Benford tested the same idea, apparently without knowing about Newcomb’s earlier note. He checked it against twenty completely unrelated sets of numbers, more than 20,000 data points in total: the surface areas of 335 rivers, the populations of 3,259 American towns, 104 physical constants, 1,800 molecular weights, thousands of entries from a math handbook, numbers pulled from Reader’s Digest, the street addresses of hundreds of scientists, and mortality rates.

The pattern held up across nearly all of it. That sheer breadth of testing is likely why history attached his name to the law instead of Newcomb’s brief observation.

From a worn bookshelf to fraud investigations

The law stayed a mathematical curiosity for decades. Then, in 1972, economist Hal Varian proposed something more practical: using it to catch fraud in official economic data. If numbers are fabricated by hand, they tend to follow a much more even, human distribution, not the lopsided one real data naturally falls into.

The idea has been put to real use since. Macroeconomic figures that Greece submitted to the European Union before adopting the euro were later flagged, in hindsight, as likely fraudulent using this exact test. In 2009, researcher Walter Mebane applied it to Iran’s presidential election results and found that the second digits in Mahmoud Ahmadinejad’s vote counts diverged sharply from what the pattern predicts, a signal widely read as evidence of ballot stuffing.

It is not a universal fraud detector

The pattern has real limits. It only holds for data that spans several orders of magnitude, so narrow ranges like IQ scores or human heights simply don’t follow it. And Mebane himself has cautioned against leaning on it too heavily at the level of individual polling stations, where the numbers involved are often too small and too similar in scale for the pattern to mean much.

See also  Guinness forbade his real name, so he signed his math as Student

A mathematician named Ted Hill finally gave the whole thing a rigorous proof in 1995, more than a century after Newcomb first noticed those worn pages, confirming why mixed, wide-ranging data sets drift toward this exact distribution.

A tired astronomer glancing at a well-thumbed book turned into one of the more unlikely tools now used to check whether a government is telling the truth about its own numbers.