Evaluating Cryptographic Randomness: an In-Depth Study of the CryptRndTest package in R

Accepted November 2025

Abstract

Random number generation plays a fundamental role in many mathematical and computational fields, such as simulation and cryptography, where it is essential for safeguarding private data. Although usually generated by deterministic algorithms, these numbers are designed to exhibit statistical properties indistinguishable from true randomness, hence the term pseudo-random. Assessing the quality of pseudorandom number generators (PRNGs) requires rigorous statistical evaluation, often performed using comprehensive test sets. In this study, we perform an in-depth analysis of the CryptRndTest test suite, a collection of eight statistical tests specifically designed to evaluate the randomness of results produced by cryptographically secure PRNGs. Our research emphasises key attributes of these tests, in particular their statistical independence and sensitivity, to provide a comprehensive understanding of their effectiveness in validating cryptographic randomness.