🔢 mathematics

A Rigorous Proof of a Ramanujan Machine Identity for $-\pi/4$ via Exact Recurrence Solving

This paper rigorously proves a Ramanujan Machine conjecture for the identity $-\pi/4$ by explicitly solving the associated second-order linear difference equation to derive a closed-form denominator sequence and evaluating the resulting limit via Abel summation and elementary integration.

Original authors: Chao Wang

Published 2026-04-08

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Chao Wang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are looking at a very strange, infinite staircase. Each step of this staircase is built from a specific recipe of numbers. If you climb this staircase forever, the height you reach isn't just a random number; it's a famous mathematical constant: negative pi over four ( $-\pi/4$ ).

For a long time, a super-smart computer project called the "Ramanujan Machine" looked at this staircase and said, "We are 99.9% sure this leads to $-\pi/4$ ." But in math, being "99.9% sure" isn't enough. You need a 100% proof.

This paper by Chao Wang is that proof. He didn't just guess; he built a bridge to show exactly why the staircase leads there. Here is how he did it, broken down into simple steps:

1. The Puzzle: A Weird Recipe

The staircase is a "continued fraction." Think of it like a Russian nesting doll where every layer has a new number inside.

The denominators (the bottom numbers) follow a simple, boring pattern: $-1, -4, -7, -10...$ (they just go down by 3).
The numerators (the top numbers) are the tricky ones. They jump around in a complex, quadratic pattern.

The Ramanujan Machine saw this pattern and guessed the final destination. Wang's job was to prove it.

2. The First Key: Cracking the Code of the Steps

To prove where the staircase ends, you first need to know the exact shape of every single step. Wang looked at the "denominator" steps (the bottom numbers) and realized they weren't random.

He found a secret formula (a "closed-form solution") that describes the size of the $n$ -th step perfectly.

The Analogy: Imagine trying to predict the height of a tree for every single year of its life. Instead of measuring it every year, Wang found a magic equation that tells you the height instantly for any year $n$ .
He proved this formula works for every single step using a method called "mathematical induction" (basically, proving that if it works for step 1, it must work for step 2, and so on forever).

3. The Second Key: The "Telescoping" Trick

Now that he knew the exact size of every step, he needed to see how the staircase behaves as you go higher and higher. Does it wobble? Does it stop?

He used a clever mathematical tool called a Wronskian identity.

The Analogy: Imagine you are walking up the stairs, but instead of just looking at one step, you look at the difference between your current step and the one before it. Wang discovered that these differences form a "telescoping" series.
What does that mean? Think of a telescope that collapses. When you add up all these tiny differences, most of the messy middle parts cancel each other out, leaving only the very beginning and the very end. This proved that the staircase doesn't wobble forever; it settles down to a specific, stable height (convergence).

4. The Final Key: Turning Numbers into a Picture

Now that he knew the staircase converges, he needed to prove the final height is exactly $-\pi/4$ .

He took the infinite sum of all those tiny differences and turned it into a calculus problem.

The Analogy: Imagine you have a pile of sand (the infinite sum). It's hard to measure the whole pile at once. Wang decided to melt the sand down into water and pour it into a specific-shaped mold (an integral).
He used a technique called Abel summation to rearrange the pile, and then used the Beta function (a special tool in calculus for measuring areas) to pour the "water" into a double-layered mold.

5. The Grand Finale: The "Aha!" Moment

Once the infinite sum was turned into a double integral (a picture of an area), he solved it using standard calculus moves:

He did a substitution (changing the perspective, like looking at a map from a different angle).
He used integration by parts (a way of breaking a hard area into a simple rectangle and a curve).

When he finished the math, the messy area calculation simplified down to a beautiful, clean result: $-\pi/4$ .

Summary

Chao Wang took a mysterious, computer-generated guess about a weird number pattern.

He found the exact formula for the steps.
He proved the steps settle down to a limit using a "telescoping" trick.
He turned the infinite sum into a geometric area and calculated it to be exactly $-\pi/4$ .

He didn't just show the computer was right; he showed why the universe of numbers arranged itself that way. It's a perfect marriage of pattern recognition and rigorous logic.

1. Problem Statement

The paper addresses a specific conjecture generated by the Ramanujan Machine project, an initiative dedicated to the automated discovery of mathematical constants. The project identified a non-canonical continued fraction representation for the constant $-\pi/4$ .

The conjectured identity is:
$-\frac{\pi}{4} = \cfrac{1}{-1 + \cfrac{1}{-4 + \cfrac{-2}{-7 + \cfrac{-9}{-10 + \cfrac{-20}{-13 + \ddots}}}}}$
where the partial denominators are linear ( $b_n = -(3n-2)$ ) and the partial numerators follow a piecewise quadratic pattern ( $a_1=a_2=1$ , and $a_n = -(n-1)(2n-5)$ for $n \geq 3$ ).

While algorithmic methods can verify such identities numerically to high precision, they do not provide analytic proofs. The complexity of the numerator sequence $a_n$ for $n \geq 3$ obscures any direct connection to standard hypergeometric series, making a rigorous analytical proof challenging.

2. Methodology

The author employs a three-stage analytical approach to prove the identity, moving beyond heuristic asymptotic arguments:

A. Exact Solution of the Recurrence Relation

The convergents of the continued fraction, $p_n/q_n$ , satisfy a second-order linear difference equation:
$\begin{pmatrix} p_n \\ q_n \end{pmatrix} = b_n \begin{pmatrix} p_{n-1} \\ q_{n-1} \end{pmatrix} + a_n \begin{pmatrix} p_{n-2} \\ q_{n-2} \end{pmatrix}$
The core of the proof involves solving this recurrence explicitly for the denominator sequence $q_n$ . The author derives a closed-form expression for $q_n$ using strong induction and double factorial notation.

B. Wronskian Telescoping Argument

To establish convergence and relate the partial quotients to the limit, the author utilizes the discrete Wronskian ( $W_n = p_n q_{n-1} - p_{n-1} q_n$ ).

A recurrence relation for the Wronskian is derived: $W_n = -a_n W_{n-1}$ .
This allows the difference between consecutive convergents ( $p_n/q_n - p_{n-1}/q_{n-1}$ ) to be expressed as a telescoping series term $D_n$ .
The convergence of the infinite series $\sum D_n$ is proven using the Ratio Test, showing the terms decay sufficiently fast (ratio $\to 1/2$ ).

C. Abel Summation and Integral Evaluation

Once the limit is expressed as an infinite series, the author applies Abel summation (summation by parts) to transform the series into a more tractable form.

The series is converted into a sum involving the Euler Beta function.
Using integral representations of the Beta function and geometric series, the sum is transformed into a double integral.
The inner integral is evaluated using an arctangent substitution, followed by a change of variables ( $y = \sqrt{x/(2-x)}$ ).
The final single integral is solved via integration by parts, yielding a closed-form result.

3. Key Contributions and Results

Theorem 3.1: Closed-Form Denominator

The paper establishes the exact solution for the denominator sequence $q_n$ :
$q_n = (-1)^n (2n - 3)!! (n^2 + n - 1)$
This formula holds for all $n \geq 0$ under specific conventions for double factorials ( $(-1)!!=1, (-3)!!=-1$ ). This explicit form is crucial for the subsequent steps.

Theorem 4.1: Convergence and Limit

The paper proves that the continued fraction converges absolutely to $-\pi/4$ .

Convergence: Proven via the ratio test on the telescoping terms $D_n$ .
Limit Identification: The limit $L$ is reduced to the expression:
$L = -\frac{1}{2} - \frac{1}{2} I$
where $I$ is an integral evaluated as $I = \frac{\pi}{2} - 1$ .
Substituting this back yields:
$L = -\frac{1}{2} - \frac{1}{2}\left(\frac{\pi}{2} - 1\right) = -\frac{\pi}{4}$

4. Significance

Analytical Rigor: The paper provides the first rigorous analytic proof for this specific Ramanujan Machine conjecture, moving beyond numerical verification.
Methodological Novelty: It demonstrates a powerful technique for handling continued fractions with complex, non-standard numerators by combining exact recurrence solving with discrete Wronskian identities and integral transforms.
Bridging Fields: The work connects the automated discovery of constants (Ramanujan Machine) with classical analysis techniques (difference equations, special functions, and integral calculus), offering a template for proving other machine-generated identities that lack obvious hypergeometric structures.

In summary, Chao Wang successfully validates the Ramanujan Machine's conjecture by deriving an exact closed-form for the recurrence denominators, proving absolute convergence, and evaluating the resulting series through a sophisticated chain of integral transformations to arrive at the exact value of $-\pi/4$ .