08-17-2026, 06:01 PM
Can π generate itself?
Summary
The paper “Can $\pi$ Generate Itself? A Monte Carlo Analysis of 314 Trillion Digits” by Alessandro Razeto and Nicola Rossi investigates a fascinating self-referential question: can the decimal digits of $\pi$ themselves be treated as pseudorandom data and used to estimate $\pi$? Using the record dataset of 314 trillion digits of $\pi$ computed by the end of 2025, the authors convert sequences of these digits into numerical samples and apply the standard Monte Carlo method for estimating the area of a circle.
Their statistical analysis finds that the available digits display a remarkably high degree of randomness, providing empirical support for random-like behavior beyond simple tests of digit frequencies, while emphasizing that true independence cannot hold because $\pi$ is a deterministic number. After optimizing how the digits are mapped into Monte Carlo points, the experiment successfully reconstructs $\pi$ to approximately $\pi \approx 3.141593$.
The result does not prove that $\pi$ is normal or that its digits are genuinely random; rather, it provides an unusually large-scale empirical demonstration that the known digits behave sufficiently like random numbers for $\pi$ to be used, in a playful sense, to numerically “generate itself.” (arxiv.org)
Read the paper on arXiv
Summary
The paper “Can $\pi$ Generate Itself? A Monte Carlo Analysis of 314 Trillion Digits” by Alessandro Razeto and Nicola Rossi investigates a fascinating self-referential question: can the decimal digits of $\pi$ themselves be treated as pseudorandom data and used to estimate $\pi$? Using the record dataset of 314 trillion digits of $\pi$ computed by the end of 2025, the authors convert sequences of these digits into numerical samples and apply the standard Monte Carlo method for estimating the area of a circle.
Their statistical analysis finds that the available digits display a remarkably high degree of randomness, providing empirical support for random-like behavior beyond simple tests of digit frequencies, while emphasizing that true independence cannot hold because $\pi$ is a deterministic number. After optimizing how the digits are mapped into Monte Carlo points, the experiment successfully reconstructs $\pi$ to approximately $\pi \approx 3.141593$.
The result does not prove that $\pi$ is normal or that its digits are genuinely random; rather, it provides an unusually large-scale empirical demonstration that the known digits behave sufficiently like random numbers for $\pi$ to be used, in a playful sense, to numerically “generate itself.” (arxiv.org)
Read the paper on arXiv
┌────────────────────────────────┐
│ KONSTANTINOS MICHAILIDIS │
└────────────────────────────────┘
│ KONSTANTINOS MICHAILIDIS │
└────────────────────────────────┘