Randomness tests

Quick sanity checks for a random sequence: chi-square, runs test, autocorrelation, and normality.

Runs locally in your browser. Input data is not uploaded. These tests do not prove cryptographic security.

Other languages: ja | en | es

How to use (3 steps)

  1. Paste a bit sequence (0/1) or a list of numbers (whitespace/CSV).
  2. Choose settings and click Run tests.
  3. Review p-values and charts, then copy a settings-only URL or download a report.

Check bias and structure

Randomness test tool

Chi-square checks uniformity, runs checks switching, and ACF checks simple dependence (not a full test suite).

Tip: You can drag & drop a .txt/.csv file onto the sequence box.

Samples

Settings

Tests to run

Results

Chi-square


          

Runs test


        

Autocorrelation


          

Normality (Jarque–Bera)


        

Frequently asked questions

Does passing these tests prove true randomness?
No. These are simple sanity checks. Passing does not prove cryptographic security, and failing can happen by chance or due to mismatched assumptions.
Is my input uploaded to a server?
No. Everything runs locally in your browser.
Why can chi-square fail for normal-distributed data?
This chi-square test checks uniformity over a range. A normal distribution is not uniform, so it can fail by design.
How large should my sample be?
Larger samples are more stable. For chi-square, keep expected counts per bin sufficiently large (a common rule of thumb is at least 5).
Which test should I use for normal-distributed samples?
Use the normality check (Jarque–Bera) to test whether your numbers look consistent with a normal distribution. The chi-square test on this page checks uniformity, not normality.

Related tools

Related calculators