How these tools are checked
Arctic Lantern Analytics was established to provide transparent, unadorned numerical instruments for students, educators, and laboratory researchers who require verified algorithmic behavior without commercial bloat or deceptive claims.
Our Testing Pipeline
Prior to publishing any computational routine, our workbench evaluates the implementation against standard empirical test suites for randomness:
- NIST SP 800-22 Test Suite: We subject sample bit-streams to frequency (monobit) tests, block frequency evaluation, runs tests, and discrete Fourier transform spectral checks to ensure no repeating harmonic structures exist.
- Dieharder Battery: We run continuous distributions through Marsaglia's Dieharder test suite to measure collision resistance, parking lot distribution spacing, and squeeze tests.
- Chi-Square Goodness-of-Fit: For integer generators bounded between [a, b], we draw 10,000,000 samples across varying interval sizes, confirming that observed bin tallies match theoretical counts within critical p-value thresholds (0.01 < p < 0.99).
Disclosed Algorithms and Code Audits
We do not rely on proprietary or closed-source calculation layers. All frontend calculations are executed via readable vanilla JavaScript accessible in browser developer tools. We explicitly state when a routine taps operating system hardware noise versus when a standard pseudo-random fallback is invoked.
| Evaluation Parameter | Standard Specification | Verification Frequency |
|---|---|---|
| Bit Frequency Parity | p-value > 0.01 via Monobit | Quarterly automated test |
| Rejection Efficiency | Discards < 2.5% of draws on standard spans | Per-build regression audit |
| Boundary Handling | Inclusive Min & Max boundary inclusion | Unit test automated suite |
| Bessel Corrected Variance | n - 1 sample variance compliance | Continuous analytical check |
Operational Philosophy
We reject the sensationalization of randomness. Numerical generation is an exact discipline governed by probability axioms. We make no claims of producing "lucky" combinations, and we actively educate users on the mathematical certainty of house edges and independent probability.
If you encounter an anomaly in our tool outputs, please report an unexpected result to our verification team.