π contains everything — normal numbers

What just happened?

We searched for short digit strings in the expansion $\pi = 3.14159265\ldots$ — and found them: “14” at position 1, “159” at position 3, and even six nines “999999” at position 762 (the so-called Feynman point). Some longer strings (such as “2718”) do not fit in the first 1000 digits — but they probably wait somewhere further on.

A “normal” number — what is it?

A normal number is one in whose infinite decimal expansion all digits and all finite digit sequences appear with the same, “fair” frequency — as if the digits were the results of dice rolls. The concept was introduced by the French mathematician Émile Borel in 1909. In particular, each of the ten digits occurs with frequency:

each pair of digits with frequency $\tfrac{1}{100}=1\%$, each triple $\tfrac{1}{1000}$ and so on. (The panel on the right shows the digit distribution in the first 1000 digits of $\pi$ — each hovers around 10%.)

A fascinating consequence

If a number is normal, then in its expansion we will find EVERY finite sequence of digits — your date of birth, any phone number, and even (after encoding letters as digits) the text of any book, the complete works of Shakespeare, or this very page! Since every finite sequence appears somewhere, so does any arbitrarily long — but finite — string. The infinite, non-periodic expansion “holds everything”.

Is π normal? It is an open puzzle

More than 100 trillion digits of $\pi$ have been computed and their distribution looks perfectly “normal” (each digit ~10%, each pair ~1%, etc.) — but there is still NO proof that $\pi$ is normal. It is one of the many mysteries of this constant. Interestingly, mathematicians have proved that normal numbers are “almost all of them”, and have even constructed explicit examples — such as Champernowne's constant $0.123456789101112\ldots$ — but for “natural” constants like $\pi$, $e$ or $\sqrt{2}$ normality remains unsettled.

An important caveat: “π contains everything” is a property of normal numbers — and for $\pi$ normality is a (strongly supported but unproven) conjecture, not a theorem. And note: normality does not mean that $\pi$ is “random” — $\pi$ is fully determined, its digits are merely distributed statistically evenly.

Why this matters for your exams:

It is an extension of the topics of decimal expansions and irrational numbers. Learn to distinguish: a terminating or repeating expansion ⇔ a rational number; an infinite, non-repeating one ⇔ an irrational number. “Normality” goes a step further — it says not only that the expansion does not end, but that the digits are distributed perfectly evenly. This trains intuition about infinity and the difference between “proved” and “very likely”.

Remember: in the expansion of a normal number every finite sequence of digits appears (each digit with frequency $\tfrac{1}{10}$). For $\pi$ normality is a conjecture — checked on trillions of digits, but still unproven.