What's happening?
The surprise
Take a big pile of numbers from real life, such as populations, country sizes or river lengths, and look only at the first digit of each one. You'd expect each digit from 1 to 9 to show up about equally, around 11% of the time. It doesn't. About 30% start with 1, about 18% with 2, and only about 5% start with 9. This is called Benford's Law.
What you're looking at
The left page draws truly random numbers. Its bars stay about level, near the dashed 11% line. The right page draws numbers from real life. Its bars lean down from 1 to 9 and follow the gold curve, which is the pattern Benford's Law predicts. The box under each chart says which pattern the numbers match so far. With only a few numbers, either side can wobble, so draw a few hundred.
Where it came from
In 1881 an astronomer named Simon Newcomb noticed that in library books of number tables, the pages for numbers starting with 1 were dirty and worn, and the pages for 9 were almost clean. People simply looked up numbers starting with 1 more often. In 1938 a physicist named Frank Benford checked over 20,000 numbers, from rivers to baseball stats, and found the same pattern. The law got his name.
Why it happens
Think of things that grow, like a town or a savings account. To go from 1,000 to 2,000, a town has to double, which takes a long time, so its population spends a long time starting with 1. To go from 9,000 to 10,000, it only has to grow 11%, so it passes through the 9s quickly. When numbers span many sizes, from tiny to huge, they pile up on the low digits. It even works in any unit: switch from miles to kilometers and the pattern stays.
Why the random side doesn't follow it
Picking a number from 1 to 999,999 where every number is equally likely gives each first digit the same share: exactly 111,111 numbers start with each digit. So truly random numbers are flat. Real life isn't random in that way.
Catching cheaters
When people make up numbers, they spread them out evenly, the way they think random looks. Real accounts don't look like that. Tax auditors and fraud examiners use Benford's Law to spot books that don't fit the curve. It has been used to question economic figures and election results. A bad fit isn't proof of cheating, though. It's a reason to look closer.
When it doesn't work
Numbers that stay in a narrow range don't follow it: adult heights, ages, prices that all end up around $9.99, or assigned numbers like phone numbers and zip codes. Cutoffs bend it too. The world-cities list only includes places with 1,000 or more people, so it has even more 1s than the law predicts.
And the Fibonacci numbers?
The famous series 1, 1, 2, 3, 5, 8, 13, 21… follows Benford's Law almost perfectly, and so do the powers of 2. Try them on the right page.
Where the randomness comes from
Right now the left page uses your browser's built-in random number maker. It's good, but it's really a very clever math formula. Soon this lab is plugging in a quantum random number generator: a small USB stick from Crypta Labs that measures tiny, unpredictable particles of light. When it's connected, the dot at the bottom turns from yellow to green.