Biased shuffle. Cards that sit next to each other in a fixed list land in the same hand more often. All 62 live hands matched the verifier. Baccarat Player Pair returns 77.14% against 80.81% on a fair shoe.
Crypto Casino Fairness Tracker
Fairness watch
Every published investigation into games at crypto casinos, in one place. Each row shows what the casino claimed, what the testers found, how the casino answered and whether players got money back.
A provably fair badge did not catch these problems. At BC.Game and Winna every tested bet still matched the casino's own verifier. The faults sat in the payout maths, the shuffle or the seed chain. I update a row when a casino answers or a fix is re-tested.
Lower return at every cash-out target. No public commitment for the seed chain was found until a public seeding event on 4 November 2026.
Towers
The 5% cut was taken again on every row. A cash-out at row 10 returned 59.87%. The pattern goes back to September 2023.
Limbo
A configuration error set the RTP to 97.5%. The fairness proof for Limbo, Dice and Plinko could not be reproduced. Dice and Plinko paid what they promised.
Plinko
Probability tables changed from 17 December 2025 to 10 March 2026. Every bet still replayed from its seeds and the verifier kept passing.
What happened at each casino
The five cases run from the newest report to the oldest. Each card shows what the casino promised, what the testers measured and how the casino answered. The bars turn that gap into dollars per $1,000 wagered. Refund figures come from the casinos unless a card says otherwise.
Case file
- Casino
- BC.Game
- Report
- Money back
- 1,916,155.68 USDT
- Reply
- Answered 2 days after the report
- Screenshot
- Open the game screen
Case 1 of 5
BC.Game Video Poker
The claim
BC.Game lists Video Poker with a 2% house edge and a 98% return. The game page still showed both figures on 19 November 2026.
What the testers found
The deal builds 52 shifted copies of one hash and sorts them. Cards that sit next to each other in a fixed list then land in the same hand about 30% more often. Under best play the return falls to about 96.4%, an edge of about 3.6%. The testers simulated 600 million hands.
All 62 live hands matched BC.Game's own verifier. Baccarat Player Pair uses the same deal and returns 77.14% against 80.81% on a fair shoe. The card list goes back to April 2020 in BC.Game's published code.
What is not known
The report does not establish intent or total player losses.
The answer
On 19 November 2026 the BC.Game CEO, Ako, confirmed the shuffle bias in a public post. Video Poker and single-player Baccarat were taken down. He said the casino's simulations matched the testers' figures: about 96.4% on the 98% Video Poker paytable, and about 77.2% on Baccarat Player Pair against 80.8% under a fair shuffle. He put the fault on an in-house shuffle from 2018 and a 2020 change to the deck's starting order. The seeds and the verifier, he said, still work.
He reported 1,779,578.29 USDT to 576,782 Video Poker accounts and 136,577.39 USDT to 142,267 Baccarat accounts, as a bonus with no wagering requirement that can be withdrawn at once. Later that day he said the sums had been credited automatically. The testers confirmed that the amount on their own test account was accurate, and said they will re-check both games when they return.
When I opened Video Poker that day, the game showed Under Maintenance. Nobody outside has checked the individual payments, and the testers have not yet re-tested a new shuffle.
On this site
BC.Game is on my No KYC casinos list. My BC.Game review carries the same investigation next to its Originals.
Case file
- Casino
- Gamdom
- Report
- Money back
- $339,113 + $250,000 credits
- Reply
- Fix re-tested 8 days after the report
- Screenshot
- Open the game screen
Case 2 of 5
Gamdom Crash
The claim
Gamdom advertises a 99% return on Crash, a 1% edge. Its Zero Edge offer promises a 100% return on the first $5,000 a day.
What the testers found
The published code returned 98.33% at a 1.5x cash-out, 98% at 2x and 97.21% at 10x. The edge reached 2.98%. On 99,119 public bets the real edge was about 2.12%. Zero Edge paid 99% at 2x.
The testers found no advance commitment for the seed chain. The chain also changed in October 2026 without notice. Their estimate of the extra cost: about $1.12 per $100 wagered, or about $1.58 million from January to October 2026.
The answer
Gamdom changed Crash. The testers checked the new verifier and 29 rounds in a row, and Crash now returns 99% at every target. Rounds from 4 November 2026 link to a public seed commitment.
Gamdom says it paid $339,113.41 plus $250,000 in reload credits to 1,293 accounts. Nobody outside has checked the individual payments.
Disclosure
Before the report, people connected to Gamdom asked the testing firm about a paid audit for another brand. The firm declined and no fees were paid.
On this site
My Gamdom review carries the same Crash investigation next to its Originals.
Case file
- Casino
- 500 Casino
- Report
- Money back
- $1,741,242
- Reply
- Fixed 1 day after the report
- Screenshot
- Open the game screen
Case 3 of 5
500 Casino Towers
The claim
500 Casino advertised a fixed 5% house edge on Towers in all three modes.
What the testers found
The 5% cut was taken again on every row. The edge grew from 9.75% on row 2 to 40.13% on row 10, where a cash-out returned 59.87%.
The evidence came from the game code, 33 games and 53 tiles in October 2026, and 27 winning records from November 2024. The pattern goes back to September 2023. The testers modelled $3.5 million to $21.6 million in extra player losses over three years.
The answer
500 Casino confirmed the error and fixed Towers on 30 October 2026. The testers re-checked all 30 multipliers and 10 wins, and all matched a 5% edge.
500 Casino says it refunded $1,741,242 to more than 35,800 players across Towers and HiLo, with 10% extra. The figure comes from the casino.
My check
On 19 November 2026 the Easy grid paid 1.42x on row 1 and 54.78x on row 10. Both are the fair multiple less 5% once.
Case file
- Casino
- BetFury
- Report
- Money back
- Not stated
- Reply
- Confirmed by email weeks later
- Screenshot
- Open the game screen
Case 4 of 5
BetFury Limbo
The claim
BetFury listed Limbo at a 99.02% return.
What the testers found
The testers placed 240,000 bets on Limbo, Dice and Plinko. Limbo returned 97.33% over 60,000 bets. Dice and Plinko paid what they promised.
The fairness proof failed on all three games. The seed already holds the result, and there is no published formula to rebuild it.
The answer
By email in July and August 2026, BetFury confirmed a configuration error that set the Limbo RTP to 97.5%. It reported a correction and contacted the test account.
Nobody outside has checked the time frame, the fix or how complete the compensation was.
What the game shows
The Limbo panel lists RTP 99.02%. At a 2x target it offers a 49.51% win chance, and 2 times 49.51% is 99.02%. The panel matches the advertised rate. Only a new test can show what the game pays.
Case file
- Casino
- Winna
- Report
- Money back
- Offered via live chat
- Reply
- Fixed 10 days before the report
- Screenshot
- Open the game screen
Case 5 of 5
Winna Plinko
The claim
Winna lists Plinko at a 99% return. The game page still showed 99% on 19 November 2026.
What the testers found
From 17 December 2025 to 10 March 2026, fixed probability tables cut the return to about 98%. All 27 setups across three risk modes were hit.
Every bet still replayed from its seeds and the verifier passed. The change sat in the step that turns a random number into a slot.
The answer
On 10 March 2026 Winna published an incident report and called it a configuration error. It offered refunds of the difference to 99%, claimed through live chat, and added automated RTP tests.
A re-check on 14 March 2026 found the per-pin logic restored. That re-check was not a full audit.
How to check a casino original yourself
Every fault on this page was visible in the game itself, if you knew where to look. These five checks take a few minutes and use only what the casino shows you.
- Find the promised rateOpen the game info or look at the bet panel. BC.Game prints House Edge 2.00% and RTP 98% under Video Poker. BetFury prints RTP 99.02% at the foot of the Limbo panel.
- Do the panel mathsIn Limbo and Dice, the target multiplier times the win chance should equal the RTP. On BetFury Limbo, 2 times 49.51% is 99.02%. In Towers and Mines, each row should pay the fair multiple less the edge once. On 500 Casino Easy, row 10 pays 54.78x: the fair 57.67x less 5%.
- Look for the seed hash before you betThe casino should show a hash of its server seed before the round. Without it, the verifier only proves that a result matches seeds picked after the fact. The testers found no advance commitment for Gamdom Crash before 4 November 2026.
- Verify a few bets, then know the limitRun two or three bets through the casino's verifier. A pass shows the result came from the seeds. It does not show the payout maths is fair. At BC.Game and Winna every tested bet passed while the return was low.
- Keep your proofSave bet IDs, seeds and a dated screenshot of the game rules. A refund can come months later. Winna asked players to claim through live chat.
Warning signs from these cases
500 Casino took 5% again on each Towers row. By row 10 the house kept 40.13%.
Compare each row with the fair multiple.Gamdom Crash returned 98.33% at 1.5x and 97.21% at 10x, against a promised 99%.
Test more than one target or mode.At BetFury the seed already held the result, with no published formula to rebuild it.
If you cannot repeat the calculation, the badge proves little.Gamdom switched its Crash seed chain in October 2026 without notice. Winna ran lower Plinko tables for almost three months.
Save the rules page with a date.Two days after the report, BC.Game took Video Poker and single-player Baccarat down to replace the shuffle.
Wait for the testers' re-test before you play a returned game.How to read the tracker
- Fixed, re-tested
- The testers checked the game again after the fix and it matched the promised rate.
- Odds restored
- A re-check found the old odds back in place. It was not a full audit.
- Fix claimed
- The casino says it fixed the game. Nobody outside has checked.
- Refund unverified
- The refund figure comes from the casino. Individual payments were not checked.
- No answer
- The casino has not answered in public. The row shows how many days have passed.
- The bars show what the house keeps from $1,000 of bets at the promised edge and at the edge the testers found. The arithmetic is mine, from the testers' figures.
Where the cases come from
Each case comes from a published investigation by ProvablyFair.org, an independent testing firm. I read the full report, the data notes and any answer from the casino before a case goes up. The firm charges casinos for audits and certification, and says it funds these investigations itself. Its independence and funding page sets out the rules.
Three investigations are not on this tracker. Winna VIP Slide and Moonroll Plinko are about bets that could not be verified at all, with no advertised rate to compare. Luck.io had a technical review with no single finding to track. The Winna and Moonroll cases are covered in the guide to provably fair games. For how these in-house games compare at the casinos I review, see my original games page.
Closing note
Fairness tracker: a closing note from Alex
I keep this tracker because a provably fair badge does not show the return. 5 cases are on it today, and the latest report is dated 17 November 2026.
I do not run these tests myself. Every figure comes from a published investigation and the casino's answer. When a casino answers or the testers re-test a game, I update the row.
What these five cases taught me
- A verifier pass is only half the checkAt BC.Game and Winna every tested bet matched the verifier while the return was low. The verifier shows the result came from the seeds. It does not check the payout maths.
- Most answers came within days500 Casino fixed Towers one day after the report. BC.Game answered in two days. Gamdom's fix was re-tested eight days after the report.
- Refund figures are the casino's wordEvery refund figure on this page comes from the casino. Nobody outside has checked the payments to players.
- Rules change quietlyGamdom switched its Crash seed chain without notice. Winna ran lower Plinko tables for almost three months. A dated screenshot is your own record.
Three things I do before I play an original
- Read the RTP on the game panel and check it with the panel maths.
- Look for the server seed hash before the first bet.
- Start small, verify two bets and save the bet IDs.
Play with a limit A fair game still has a house edge. Set a budget before you log in and stop when you reach it.
Responsible gambling guideQuestions players ask
Where do these cases come from?
From investigations published by ProvablyFair.org, an independent testing firm. Each case on this page links to the full report with the data and the scripts. I read every report and the casino's response before a row goes up.
Are the refund figures checked?
No. Every refund figure on this page comes from the casino. The testers re-tested the fixed games at 500 Casino and Gamdom, and confirmed the amount paid to their own test account at BC.Game. Nobody outside has checked the individual payments.
I played one of these games. What should I do?
Save your bet history, the bet IDs and the game rules now. Then write to support, name the investigation and ask how your share of any refund was worked out. Keep the reply.
Does a fix pay back past losses?
Not on its own. A fix changes the game from that day on. Money for earlier rounds needs a separate refund decision by the casino.
Is a casino that is not listed safe?
Not necessarily. The testers cannot look at every game. A casino missing from this page has simply not been investigated in public.
ProvablyFair.org charges casinos for certification. Duel and LIQD are certified clients, and this site has partner links with both. A partner deal does not change a row on this tracker. How the site makes money
Alex JovanovicCrypto Casino Reviews

