Harunagi
Weather · World · Time — calm, cited, reader-first
Science & Sky · Mathematics

The coin flip hidden inside the whole numbers

Pick a whole number at random. The chance it is “clean” of square factors is exactly 6/π² — the circle constant, presiding over a question with no circle in it.

Written in-house · 18 August 2026 · 3 min read · evergreen
Plate 13 · Science & Sky · original Harunagi print春凪

Here is a game requiring no equipment. Pick a whole number at random — any size you like. Call it clean if no perfect square divides it: 10 is clean (2 × 5), but 12 is not (4 sits inside it), and 99 is not (9 does). Mathematicians say squarefree. Now the question: what fraction of all numbers are clean? The answer has been known for well over a century, and it is one of the loveliest facts in arithmetic: exactly 6/π², which is 60.79 percent. The circle constant π, uninvited, governing a coin flip buried in the whole numbers.

The route to it is short enough to sketch. A number avoids being divisible by 4 with probability 3/4; it avoids 9 with probability 8/9; it avoids 25 with probability 24/25 — and, by a miracle primes grant freely, these events behave independently. Multiply the chain across every prime, and Euler’s machinery turns the product into 1/ζ(2) — the reciprocal of a famous sum he solved in 1735: 1 + 1/4 + 1/9 + 1/16 + … = π²/6. Divide, and the odds of cleanliness come out 6/π². What looks like bookkeeping across infinitely many primes compresses to three symbols.

Then the ground falls away. Ask the same question not about all numbers but about the values of a single polynomial — say x⁴ + 1 as x runs along — and mathematics, which answered the first question in the nineteenth century, cannot answer this one. It is believed the same kind of density holds; it is proven only if one assumes the notorious ABC conjecture, whose own claimed proof has sat disputed for over a decade. And here the story turns strange and modern: Manjul Bhargava — who woke Gauss’s composition law from two centuries of sleep and won the Fields Medal for it — showed that some polynomials in forty variables of degree forty CAN be handled, even while x⁴ + 1, one variable, degree four, stands untouched. In arithmetic, difficulty does not grow with size; it hides where it pleases.

We print this at a weather-and-probabilities desk for a reason. The squarefree density is the purest specimen of a thing our whole trade depends on: certainty about proportions without knowledge of cases. No one can tell you whether the next number that comes to mind is clean, and no one can tell you whether it will rain on the fourteenth of next month — yet the fraction, over the long run, is knowable, and in arithmetic it is knowable exactly. The weather odds on our front page must earn their proportions from decades of verification. The integers hand theirs over by proof — six in ten, forever, with π keeping the books.

What is not yet establishedNo dice were harmed: “a number at random” is an idealization made precise by limits, and the 6/π² density is a theorem — unlike our weather percentages, which must re-earn their calibration every season. We find the kinship instructive and the difference honest.
Sources UCLA Mathematics Distinguished Lecture — Manjul Bhargava on squarefree values (2015) → Bhargava — The geometric sieve and the density of squarefree values of invariant polynomials (arXiv:1402.0031) → The Basel problem — Euler’s ζ(2) = π²/6 →

← More world news  ·  Harunagi front page  ·  Weather odds