From player action to authoritative outcome.

RNG

Node.js crypto.randomInt() drives a Fisher–Yates draw without replacement.

State

A NestJS-owned in-memory session is keyed by a cryptographic UUID.

Game logic

The service validates 2–10 selections, draws 20 numbers, counts hits, and applies the configured multiplier.

RTP / testing

Combinatoric probability helpers calculate theoretical RTP for each selection count.

Demo credits and sessions are simplified; there is no production wallet, durable ledger, or certification layer.

RNG

Node.js crypto.randomBytes() provides weighted, server-side symbol selection.

State

The server retains the complete card; the client receives only symbols it has revealed.

Game logic

Nine weighted symbols are sealed when the card is created, then match groups resolve against the paytable.

RTP / testing

The RTP test runner samples large card volumes and compares observed return with the configured model.

The balance is a portfolio-only credit simulation; sessions are in memory and do not represent financial settlement.

RNG

Node.js crypto.randomBytes() produces an unbiased house move after the player commits.

State

The service owns virtual balance, current wager, round result, statistics, and session history.

Game logic

The engine resolves the move pair, applies a win/tie/loss multiplier, and returns the updated balance.

RTP / testing

The three equally likely house moves and explicit multipliers produce a 95% theoretical return.

Balances are virtual and sessions are ephemeral; authentication, ledgering, and responsible-gaming controls are outside this demo.

RNG

No RNG is used; movement and collision behavior are deterministic.

State

Position, movement count, elapsed time, and completion state live entirely in the browser.

Game logic

Keyboard, touch, or on-screen input advances the player only when the next bounding box avoids every wall.

RTP / testing

A fixed maze and deterministic step size make routes repeatable and interaction behavior easy to verify.

This is deliberately a client-side rendering experiment, not a server-backed wagering game.

RNG

Node.js crypto.randomInt() selects an independent stop on each calibrated reel strip.

State

A NestJS session owns bet configuration and the previous spin while the UI maintains demo-only credits.

Game logic

The scanner handles blanks, exact triples, mixed bars, cherries, and Double Diamond wild multiplication.

RTP / testing

Deterministic scanner tests verify one-wild 2× substitution against the original Go payout semantics.

The 95.109723% figure belongs to the imported reel set; demo credits are not a wallet or durable ledger.

RNG

Every reel stop is sampled server-side using Node.js crypto.randomInt().

State

The game session owns the bet-per-line, active line count, screen, wins, and last result.

Game logic

Only selected paylines are evaluated; three matching symbols settle from the imported paytable.

RTP / testing

A focused test proves inactive lines cannot contribute wins to a spin response.

The calibrated 95.504233% source RTP assumes its exact reel set and play configuration, not a short demo sample.

RNG

Five calibrated strips are sampled independently by the backend CSPRNG.

State

The engine returns a reel-major screen and normalized line/scatter win records with exact coordinates.

Game logic

Left-aligned line wins, scatter counts, and consecutive filled-column multipliers are evaluated separately.

RTP / testing

The parity suite verifies a five-column filled screen applies the original 5× multiplier.

This demonstrates game-math separation and result contracts; it does not implement regulated settlement.

RNG

The service uses cryptographic reel-stop selection against a source-calibrated strip set.

State

One isolated session owns configuration and the previous result; every win includes its line and hit positions.

Game logic

The scanner evaluates left and right runs without overlap, applies wild multipliers, then settles scatters independently.

RTP / testing

A deterministic screen proves simultaneous left- and right-paying combinations preserve the Go behavior.

The 95.353086% reel model is presented as technical evidence, not certification or a financial guarantee.

RNG

A cryptographic Fisher–Yates shuffle draws 20 unique numbers from 80.

State

The server owns validated selections, bet, board state, previous draw, and payout multiplier.

Game logic

Hit count indexes the exact Fire Keno row for the selected-number count.

RTP / testing

A deterministic two-hit draw verifies the imported 6× multiplier and final payout.

The 92.028857% source RTP is mathematical metadata; the UI balance is deliberately local and virtual.

RNG

The same reusable crypto sampler draws unique numbers without moving authority into the browser.

State

Selections and the last draw remain isolated inside a short-lived NestJS session.

Game logic

The game supports one-number play and unusual zero-hit returns for selected higher-count rows.

RTP / testing

A deterministic one-hit case verifies the original 3.8× single-number payout.

The 95.616967% value is the source game RTP; virtual credits exist only to make result deltas visible.