Skip to content
BY PROVABLYFAIR

Independent proof that Duel.com’s certified games work as advertised.

Duel.com

Certified

Official website ·duel.com (opens in a new tab)

Independent proof that Duel.com’s certified games work as advertised.

PF-2026-DUEL · Framework v1.1

CERTIFICATION ISSUED
Passed
Next review · Dec 2026 · Unannounced

Certified since 8 Jun 2026

Bets placed & verified61,155
Simulated rounds266M
Original games in scope11
Player checks
—Loading…
AUDIT RESULTPassed

8 Jun 2026

Next review · Dec 2026

Unannounced review window

Bets placed & verified61,155
Simulated rounds266,000,000
Games in scope11 original games
Player checks
—Loading…
CertificationInitial audit completed
10 original games
Re-auditFramework v1.1
10 original games
Limbo addedAudited under Framework v1.1
11 games in scope

Published audit record

  1. Certification

    Initial audit completed

    10 original games
  2. Re-audit

    Framework v1.1

    10 original games
  3. Limbo added

    Audited under Framework v1.1

    11 games in scope
View full record (opens in a new tab)

If a result feels wrong

  1. 01Verify bet
  2. 02Understand result
  3. 03Dispute result
160 integrity checks · 4 disclosed findings · 0 fairness failuresEvidence & methodology → (opens in a new tab)

PF-2026-DUEL · Framework v1.1. Audit totals describe the published test datasets; fully matching player checks join public activity without bet IDs or usernames. Full findings and supporting evidence remain in the technical report.

01Per-game resultsGame results

Find your game. Inspect the captured bets, mathematical return and simulated rounds in the published audit.

11 games validatedLatest update
GameVerified betsRTP · ProvenRTP · SimulatedResultEvidence
Plinko8,10099.9%99.917%27M roundsPass
Dice6,70099.9%99.922%98M roundsPass
Mines7,05099.9%99.904%24M roundsPass
Keno5,10099.9%99.970%40M roundsPass
Beef6,45099.2%99.147%4M roundsPass
Crash1,10099.9%99.872%5M roundsPass
Castle Roulette1,350100.0%100.007%5M roundsPass
Video Poker5,40099.9%Strategy-dependent10M roundsPass
Blackjack6,00099.4296%99.4670%10M roundsPass
Groomer’s Van6,20599.9720%100.06%30M roundsPass
Limbo7,70099.9%99.911%13M roundsPass

RTP · Proven99.9%
RTP · Simulated99.917%27M simulated rounds
Bets captured
8,100
RNG type
HMAC-SHA256

61,155 verified live audit bets in the published record. RTP figures retain the report’s precision; consult each report for strategy and assumptions.

02What this certificate covers

Game fairness, clear boundaries and ongoing evidence.

Audited

  • 11 original games in the published scope
  • Seed commitments, nonce & randomness
  • Live bets independently replayed
  • Payout calculations & game RTP
  • Adversarial fairness & input checks

Live casino capture · March–May 2026; Limbo added October 2026

Ongoing verification

Quarterly unannounced spot checks revisit the certified games.

Player-submitted bets run against the audited code. Saved result summaries add fresh evidence to this casino’s public record.

Audited

  • 11 original games in the published scope
  • Seed commitments, nonce & randomness
  • Live bets independently replayed
  • Payout calculations & game RTP
  • Adversarial fairness & input checks

Live casino capture · March–May 2026; Limbo added October 2026

Not covered

  • Solvency, fund custody & withdrawals
  • Licensing & regulatory compliance
  • Platform & infrastructure security
  • Third-party games outside the audit
  • Rakeback, rebates & promotions
  • High-stake scaling edge testing
  • Maximum-win cap enforcement

Ongoing verification

Quarterly unannounced spot checks revisit the certified games.

Player-submitted bets run against the audited code. Saved result summaries add fresh evidence to this casino’s public record.

03Straight answers

Meaning

The same committed inputs produce the same result.

The revealed seed is checked against its saved commitment, then the game is replayed independently. Changing a committed input breaks that check.

Method
  • 1,065 seed cycles checked across 8 direct-HMAC client-seed games.
  • Player-controlled client seeds checked where applicable.
  • Nonce continuity and deterministic outcomes verified.
  • 2,450 drand-backed rounds checked on Crash and Castle Roulette.
Sample

61,155 / 61,155 live bets replayed · 0 mismatches

Limit

A matching hash proves consistency with the saved commitment. Keep that commitment before betting; the hash alone does not prove when it was published.

A displayed result that disagrees with the committed math fails verification. This check concerns the game result, not the account balance or withdrawal.

Quarterly spot checks and saved player verifications extend the audit record. Certification remains limited to the published games and versions.

04The complete public record

Findings, fixes and every certification event stay visible.

Disclosed findings
4
Games with findings
3
Fairness failures
0

05Certification history

Every milestone is part of the record.

  1. Mar–May 2026

    Original 10-game audit

    Anonymous live-bet capture and independent testing.

  2. Initial audit

    Findings issued

    4 findings disclosed across 3 games.

  3. Certified

    PF-2026-DUEL

    Certified
  4. Re-test under Framework v1.1

    The original 10 games re-tested under the updated framework.

  5. Limbo added

    Limbo added as the 11th game in the audit scope.

  6. Next review window

    Unannounced

Certification requirements 5/5 passed

  1. 01Passed
    Seed, nonce and determinism
  2. 02Passed
    RNG and entropy model
  3. 03Passed
    Live vs verifier match
  4. 04Passed
    Game logic and RTP
  5. 05Passed
    Fairness integrity
ControlEvidenceResult
01Seed, nonce and determinism1,065 seed cycles intact8 direct-HMAC client-seed gamesPass
02RNG and entropy model8 HMAC-SHA2562 HMAC + drand · 1 HMAC → PCG32Pass
03Live vs verifier match61,155 / 61,155 live bets matchedPass
04Game logic and RTP11 game RTP proofs266M simulated roundsPass
05Fairness integrity160 checks156 passed · 4 disclosed findingsPass

Certification requirements

5 requirements. Each backed by evidence.

5 / 5
Requirements passedFramework v1.1

View full methodology (opens in a new tab)
INDEPENDENT REPRODUCTION

07Don’t trust the certificate. Check it.

Every audit is built so the result can be independently reproduced.

Verify your bet

Check a result using the published audit’s verification tool.

Verify your bet

Re-run the audit

Open a game repository and its dataset. Follow its documented setup and reproduction steps.

Have an AI review the code

Open a GitHub review prompt in your AI. Replace [Game] with your game, or tell the AI which game when it asks.

OPEN SOURCE AUDIT

Run the audit on your machine.

Inspect the code. Replay the bets. Challenge the findings.

Duel · Reproduction guidehttps://github.com/orgs/ProvablyFair-org/repositories?q=duel

1. Choose a game repository.
2. Follow the repository’s README.
3. Reproduce the published evidence.

Verifier code · Captured bets · Simulation harness

11 game repositories · 61,155 verified bets

Browse all 11 repositories ↗ (opens in a new tab)
OPEN SOURCE AUDIT

Reproduce the audit

Clone the code. Replay the bets.
Challenge the findings.

duel-plinko
$ git clone https://github.com/ProvablyFair-org/duel-plinko$ cd duel-plinko# Follow the repository README# Replay the published evidence
  • Verifier code & captured bets
  • Simulation harness & findings
Open the audit repositories (opens in a new tab)

11 game repositories · 61,155 verified bets
Setup and runtime: see each repository’s README

Dispute a result

Something doesn’t add up? Start with the bet and the published evidence.

An independent second look.

Compare the result against the audited game. Keep the bet reference and supporting evidence together.

Compare your result with the published game report and verifier.

Understand a verification result