Glossary: Every Term We Use, Explained

You do not need a math degree to understand our audits. Every term below is explained in plain language first, with the scientific definition underneath.

The Basics

RNG (Random Number Generator)

A digital coin-flipper. It generates unpredictable numbers that decide game outcomes — like the crash multiplier or which card you get.

Scientific: A deterministic or non-deterministic algorithm producing a sequence of values statistically indistinguishable from true randomness.

Provably Fair

A system where you can check after each round whether the casino cheated. Think of it like a sealed envelope: the result is locked in before you bet, and opened after.

Scientific: A cryptographic commitment scheme where the outcome is determined pre-bet via HMAC-SHA256. The server seed is revealed post-round for independent verification.

Server Seed

The casino’s secret ingredient. Generated before your bet, hidden during the round, revealed after so you can check.

Scientific: A cryptographic secret generated server-side, committed via its SHA-256 hash before the game round, disclosed post-round for verification.

Client Seed

Your ingredient. Generated by your browser. You can change it anytime to add your own randomness to the mix.

Scientific: A user-provided or browser-generated value participating in the HMAC computation, ensuring the server cannot unilaterally determine the outcome.

Nonce

A counter that goes up by 1 each round. Makes sure the same seeds never produce the same result twice.

Scientific: An incrementing integer (Number used ONCE) concatenated with the seed pair to guarantee unique outputs per round.

Hash / SHA-256

A digital fingerprint. Any input produces a unique code. Change one letter and the fingerprint is completely different. It cannot be reversed.

Scientific: A cryptographic hash function producing a fixed 256-bit digest. Preimage-resistant and collision-resistant.

HMAC-SHA256

The recipe that mixes server seed + client seed + nonce to produce the game result. One-way: you can verify it, but you cannot predict it.

Scientific: Hash-based Message Authentication Code using SHA-256. Combines a secret key with a message to produce a deterministic but unpredictable output.

House Edge

The casino’s built-in advantage. At 1% house edge, you lose about $1 for every $100 wagered over time. That is how casinos stay in business.

Scientific: Expected value of casino profit as percentage of wager: E[casino profit] / total wagered × 100.

Crash Point / Multiplier

The number that climbs until the game crashes. Cash out before the crash = you win. Miss it = you lose your bet.

Scientific: Game outcome derived from HMAC output, typically following an inverse power-law distribution with ~1% instant crash probability.

Test Results & Statistics

p-value

A score from 0 to 1 that answers: “How normal is this result if the game is truly random?” Above 0.01 = looks fine. Below 0.01 = suspicious.

Scientific: Probability of observing results at least as extreme as measured, assuming the null hypothesis (true randomness). Significance threshold: α = 0.01.

PASS / FAIL

PASS = the data looks random, nothing suspicious. FAIL = something is statistically off and needs investigation.

Scientific: p-value ≥ 0.01 = fail to reject null hypothesis (PASS). p-value < 0.01 = statistically significant deviation (FAIL).

Confidence Interval (CI 95%)

A range where the true answer almost certainly lies. “House edge 1.0%, CI: 0.82%-1.21%” means we are 95% sure the real number is in that range.

Scientific: Interval estimate where 95% of such intervals from repeated sampling contain the true population parameter.

Entropy

How unpredictable the data is. High entropy = impossible to guess the next number. Low entropy = patterns exist and something may be wrong.

Scientific: Shannon entropy: H = -Σ p(x) log2 p(x). Measures average information content per symbol.

Correlation

Does one round affect the next? Correlation near 0 = each round is independent (good). Far from 0 = rounds influence each other (bad).

Scientific: Serial autocorrelation (Lag-1): Pearson’s r between consecutive outcomes. Values near 0 indicate independence.

Sample Size

How many rounds we tested. 100 rounds proves nothing. 100,000 is solid. 462,000+ is extremely reliable. More data = more certainty.

Scientific: Number of independent observations (n). Statistical power increases with n. NIST recommends minimum 1,000,000 bits.

Variance

The natural ups and downs of gambling. Even in a perfectly fair game, you can have brutal losing streaks. That is variance, not cheating.

Scientific: Expected squared deviation from the mean. High-variance games produce extreme short-term swings despite fixed long-term expected value.

Distribution

The pattern of how results spread out. Low crash multipliers happen often, high ones are rare. If this pattern matches the math, the game is fair.

Scientific: Probability density function governing outcomes. Standard Crash: P(multiplier ≥ x) = 1/x for x ≥ 1 (inverse power-law).

Testing Frameworks

NIST SP 800-22

A toolkit of 15 tests created by the U.S. government to check if numbers are truly random. The same tests that gaming labs (GLI, BMM) use to certify slot machines in Las Vegas.

Scientific: Statistical Test Suite by the National Institute of Standards and Technology. 15 tests evaluating frequency, runs, spectral properties, entropy, complexity, and random walk behavior.

PractRand

A pattern hunter. Runs 70+ tests at once, designed to catch subtle tricks that other tests miss — especially in very large datasets.

Scientific: Battery of 70+ statistical tests by Chris Doty-Humphrey, optimized for detecting long-cycle correlations and bit-level dependencies.

TestU01

The academic gold standard. Created at a Canadian university, used in 1,000+ scientific papers. If an RNG passes BigCrush (106 tests), it is as good as it gets.

Scientific: Comprehensive C library by L’Ecuyer & Simard (Université de Montréal). SmallCrush (10), Crush (96), BigCrush (106 tests).

GLI-19

The rulebook that gaming labs follow to certify casino games. If a game does not pass GLI-19 tests, it cannot go on the casino floor. Section 4.2 requires NIST SP 800-22.

Scientific: Gaming Laboratories International Technical Standard for interactive gaming systems. Mandates NIST SP 800-22 for RNG certification.

SHA-256 Checksum

A unique fingerprint of our dataset. If anyone changes even one number, the fingerprint changes completely. Proves our data has not been tampered with.

Scientific: Fixed 256-bit digest for integrity verification. Any data modification produces a different hash.

Still confused? No problem.

You do not need to understand every test. These same methods certify PokerStars, bet365, and every slot machine in Las Vegas. We apply them to crypto casinos — and show our work. Full methodology →