A PRNG (Pseudo‑Random Number Generator) Validator applies statistical tests to a sequence of numbers to check if they appear random and uniformly distributed. It helps developers, cryptographers and statisticians assess the quality of random number generators before using them in simulations, games or security.

The tool runs four fundamental tests:

• Mean Test – checks if the average matches the expected midpoint.
• Chi‑Square Goodness‑of‑Fit – verifies uniform distribution across bins.
• Runs Test – detects patterns by counting sign changes around the median.
• Autocorrelation (lag‑1) – measures dependency between consecutive numbers.

Together they provide a reliable quick check of randomness.

Each test returns a p‑value. If the p‑value is greater than the significance level α (default 0.05), the sequence passes that test – the null hypothesis of randomness cannot be rejected. A green badge indicates a pass; a red badge suggests possible non‑random behaviour. A single failure does not necessarily mean the generator is bad; look at the overall pattern.

While the calculator works with as few as 20 numbers, we recommend at least 100 values for trustworthy results. Smaller samples can produce unstable p‑values, especially for the Chi‑Square test. The tool will warn you if your sample seems too small.

Yes. Uncheck Auto range and manually set the minimum and maximum values (e.g., 1 and 6 for a six‑sided die). The same tests apply – just be aware that with a discrete distribution the Chi‑Square approximation may be slightly conservative.

These four tests provide a basic sanity check, but they are not a complete cryptographic evaluation. For security‑critical applications you should use dedicated test suites like NIST SP 800‑22 or Dieharder. Our tool is ideal for initial debugging, education, and non‑crypto use cases.