The Unreasonable Depth of 1/π

Undiscovered Country, Essay 7

There’s a number that most people meet in middle school and forget by high school. It appears in the circumference of circles, the area of disks, the period of pendulums, and a hundred other places where the universe bends. You know it. 3.14159…

But turn it upside down. Look at 1/π instead. Something strange happens.

The Man Who Knew Infinity (and Its Reciprocal)

In 1914, a clerk from Madras named Srinivasa Ramanujan sent a letter to G.H. Hardy at Cambridge containing, among other miracles, this formula:

$$\frac{1}{\pi} = \frac{2\sqrt{2}}{9801} \sum_{k=0}^{\infty} \frac{(4k)!(1103 + 26390k)}{(k!)^4 \cdot 396^{4k}}$$

Read that again. Where does 9801 come from? Why 26390? Why 396⁴? These aren’t arbitrary. 9801 = 99². 396 = 4 × 99. And 99 relates to the class number of an imaginary quadratic field. Ramanujan hadn’t just found a fast formula for π — he’d found a tunnel between circle geometry and algebraic number theory that nobody knew existed.

Each term of that series adds roughly eight correct decimal digits of π. For comparison, Leibniz’s formula (1 - 1/3 + 1/5 - 1/7 + …) needs about five billion terms to get ten digits. Ramanujan’s formula gets ten digits from the first term alone.

Why 1/π and Not π?

Here’s the thing that doesn’t get enough attention: the reciprocal is more natural.

π itself is transcendental — proved by Lindemann in 1882, settling the ancient Greek problem of squaring the circle. You can’t construct π from rational numbers and roots. It lives in a wilderness beyond algebra.

But 1/π organizes. It submits to patterns. Ramanujan’s series, the Chudnovsky brothers’ refinement that broke every π computation record from 1989 onward, the BBP formula that lets you compute individual hexadecimal digits of π without computing the ones before them — these are all formulas for 1/π, not π.

It’s as though π is a locked room and the key fits better from the other side.

The Chudnovsky Algorithm

In 1988, David and Gregory Chudnovsky, working from a cramped apartment in New York with a homebrew supercomputer they’d built from mail-order parts, published this:

$$\frac{1}{\pi} = 12 \sum_{k=0}^{\infty} \frac{(-1)^k (6k)! (13591409 + 545140134k)}{(3k)!(k!)^3 \cdot 640320^{3k + 3/2}}$$

Each term: fourteen correct digits. This is the formula that has computed π to the current world record of 105 trillion digits. Not by brute force — by an almost obscene convergence rate that emerges from the arithmetic of a particular elliptic curve with complex multiplication.

640320³ = 262537412640768000. That number is within 0.00000000000075 of being an integer power of e^(π√163). That near-miss isn’t a coincidence. 163 is a Heegner number — one of exactly nine values of d for which the imaginary quadratic field ℚ(√-d) has class number 1. The Chudnovsky formula works because of a deep connection between these fields and modular functions. Ramanujan’s formula works for the same reason, with the Heegner number 58 (since 396 = 4 × 99, and 99² + 1 relates to discriminant -232 = -4 × 58).

Two formulas. Two Heegner numbers. One tunnel.

The Part Nobody Talks About

Here’s what gets me. Ramanujan didn’t know the theory of complex multiplication. He didn’t have access to the work of Weber, Kronecker, or Hilbert that would later explain why his formulas work. He derived them from raw pattern recognition, from staring at modular equations and seeing structure that took the rest of mathematics decades to formalize.

When Hardy asked how he came up with his results, Ramanujan said the goddess Namagiri revealed them in dreams. Mathematicians have spent a century trying to figure out what the goddess knew.

The Chudnovskys took thirty years of additional theory — the Borweins’ work on arithmetic-geometric means, Hilbert class fields, the theory of singular moduli — and still ended up in essentially the same tunnel Ramanujan had found by lamplight in Kumbakonam.

This happens more often than mathematicians admit. The formulas precede the understanding. The structure is there before anyone can explain why. You find the key, and then you spend decades figuring out what lock it opens.

163 and the Near Miss

Let me dwell on that near-integer for a moment, because it’s one of the most beautiful accidents in mathematics.

$e^{\pi\sqrt{163}} = 262537412640768743.99999999999925...$

Twelve nines after the decimal point. Not exactly an integer, but so close that when this was first computed, people suspected a hoax. It’s called Ramanujan’s constant, even though Hermite had noticed the near-miss decades earlier. The name stuck because it felt like the kind of thing Ramanujan would find.

The explanation is class field theory. When the class number is 1, the j-invariant of the corresponding elliptic curve is an integer, and e^(π√d) is almost (but not quite) a cube of that integer plus 744. It’s a consequence of the fact that there are only finitely many imaginary quadratic fields with unique factorization, and 163 is the largest discriminant where this happens.

Nine values. Exactly nine. A finite list that connects number theory, elliptic curves, modular forms, and the geometry of circles. All visible from the vantage point of 1/π.

What It Means

Mathematics has this reputation for being cumulative — each generation standing on the shoulders of the last. But the story of 1/π suggests something different. The structure was always there. Ramanujan saw it from one direction, the Chudnovskys from another. Neither view is complete, and neither is wrong. The formulas converge — both in the numerical sense and in the sense that every approach to 1/π seems to lead to the same few tunnels.

That’s the unreasonable part. Not that π has deep structure — everything in mathematics has deep structure if you look hard enough. The unreasonable part is that 1/π is more organized than π. That inverting the most famous constant in mathematics reveals a landscape more connected, more patterned, and more surprising than the constant itself.

Sometimes you understand something better by looking at what it isn’t.

The Undiscovered Country is a series about mathematical ideas that are beautiful, surprising, and underappreciated. Previous essays: The Ghost of e⁻², The Prime Conspiracy, The Drunkard’s Detour.