Reference page · Last updated 2026-05-25

Randomness glossary, in plain English

If you've come from /fairness or /how-it-works and hit a term that doesn't make immediate sense, this page is for you. No math background assumed.

CSPRNG

Cryptographically Secure Pseudo-Random Number Generator. A randomness source designed to resist prediction even by attackers with significant computational power. Used in TLS keys, password generators, and — at NamesOnWheel — every wheel spin. Browsers expose it as crypto.getRandomValues.

PRNG

Pseudo-Random Number Generator. A deterministic algorithm that produces numbers that look random but are predictable from the seed. Math.random in browsers is a PRNG. Fine for non-security uses; not fine for picks an audience will scrutinize.

Modulo bias

The subtle bias introduced when you compute random32() % n and n doesn't evenly divide 2^32. Some indices land slightly more often. Removed by rejection sampling.

Rejection sampling

An algorithm that draws random values and rejects any that fall outside a fair range, then re-draws. Used to convert a 32-bit random integer into a uniform integer in [0, n) without modulo bias.

Chi-square test (χ²)

A statistical test for whether a set of observed counts matches an expected distribution. We use it on /fairness to verify that 100,000 spins produced uniform counts across segments. The 'p-value' tells you the probability of the result happening by chance under a fair generator.

P-value

Probability of observing a result at least as extreme as what you saw, assuming the null hypothesis (here: 'the wheel is fair'). Above 0.05 = no evidence the wheel is biased. Between 0.05 and 0.95 is the comfortable 'looks fair' zone.

Uniform distribution

Every outcome equally likely. A fair coin produces uniform 50/50. A fair 6-sided die produces uniform 1/6. Our wheel produces uniform 1/N for N entries with equal weights.

Weighted random

A pick where some outcomes are more likely than others. We support this with the *N suffix syntax (e.g. 'Alice *3' makes Alice 3x as likely as a default-weight entry).

Entropy

A measure of unpredictability. CSPRNGs draw entropy from hardware sources (mouse movements, keyboard timings, dedicated CPU instructions like RDRAND). High entropy = hard to predict.

Seed

The initial state of a PRNG. Same seed produces the same sequence. CSPRNGs continuously add entropy so there's no single 'seed' that defines the entire output.

Entropy pool

A buffer in the OS that mixes entropy from many sources. CSPRNG calls drain from it. Modern OSes keep this pool well-fed at all times.

Service Worker

Browser feature that caches assets so a website can run offline. NamesOnWheel uses a service worker to make the entire site usable without internet after first visit.

Progressive Web App (PWA)

A website that can be installed as an app from the browser, with offline support and home-screen icon. NamesOnWheel installs as a PWA so you can open it like an app.

Schema markup

Structured data (JSON-LD) embedded in a page so search engines and AI assistants can understand its content. We ship Organization, WebSite, WebPage, Person, FAQPage, HowTo, BreadcrumbList — all the formats Google uses for AI Overviews citation.