What happened:: Betpanda claims their original games (dice, turbo crash, etc.) are provably fair. Their provably fair is cleverly broken in a way that makes verification difficult when verifying normal bets, and impossible when verifying autobet.
In every other casino that claims provably fair - literally every single one - the system works like this: server seed stays the same, client seed stays the same, and the nonce increments with each bet. You can verify millions of rolls independently.
In Betpanda the nonce stays the same.
The server seed changes on every bet.
When you stop autobet, the server seed hash you saved before the session is useless after the first roll. You can only verify bet #1. You just have to trust them with the other rolls.
If rolls cannot be independently verified, the system is not provably fair. (It is not my opinion, it is the definition. Don't be like they you don't have evidence they are manipulating rolls so I wearing Betpanda signature will aggressively vouch for the people behind this anonymous casino,...). They have a signature campaign here, and I find there is a lot of resistance to this.
In autobet, the server knows your client seed (fixed), your bet direction, and your bet amount. The server generates the server seed itself. With all that information, the server can selectively skip a winning seed and generate a new one before serving it. A 0.1 second delay from this is unnoticeable.
They don't manipulate every roll. But the system is designed so that under certain conditions - say a high bet relative to net deposits, or a big multiplier hit - the server can simply regenerate the seed. The player has no way to prove it happened.
This is similar to how 999dice operated for years before getting caught:
https://bitcointalk.org/index.php?topic=948965.0Extract from 999dice thread "So I started doing testing. I ran the API, using doge, and made over 1 million bets at a 95% win chance, and my final numbers were a 95.0172% win ratio (or something close). Which sure seems like the site is legit... except that although I had a 95% win ratio, I was still down more then I should have been.
How is that possible? Because the site only needs to make you lose a few extra times on larger bets to really punch your wallet, and that would never truly affect your win percentage when you're making a million rolls. It's especially bad with martingale strategy. 49.95% odds, 1 million rolls, you can have a 49.96% win result, yet be down hundreds of bitcoin, if you won a lot of 1st rolls, and the rolls that were faked were the ones where you're at the end of the run and betting 35 bitcoin. Fake just 10 rolls out of 10000 and you bring a 49.95% win ratio down to 49.85%, something that is perfectly within the bounds of random chance. Yet 10 faked rolls could cost you 350 bitcoin if they are timed just right."
I suspect Betpanda to be masking it in a different way. Small bets have a higher than normal chance of winning, and large bets have a smaller than normal chance of winning. My sample size is not large enough to draw any conclusions. But over my years of winning and losing, I notice statistically rare anomalies only in Betpanda. I never won a single large bet in original games. But I cannot prove it.
They also show Player RTP stats on the main page. This works as cover - keep the vast majority of rolls random, suppress big multipliers on big bets, add big multipliers on small bets, keep the displayed RTP near 99%. Earlier the RTP charts were too close to 99%, but last few weeks they have increased the variance to make it more realistic. If anyone questions the broken provably fair system, players look at the vouches given by randos getting paid weekly for their signature, or for running campaigns for them, and see their RTP number and think well it looks fine such a big anonymous casino cannot cheat. Actually they can. Even if they get caught, they can say oops mistake. It is the perfect crime.
What I can prove and what I can't:
Part one:
Betpanda's autobet is not provably fair. This is a fact. Anyone can verify it in their own account right now. Set a client seed, make 2 bets on autobet, check if you had the server hash needed to verify the second bet. You didn't.Part two: Betpanda is using this to manipulate rolls. For this, the undisputable evidence is in their server. I have observed anomalies (to be honest, this is barely enough evidence, even if I had documented them) and Betpanda meets every condition needed for manipulation, run by anonymous operators. I cannot hand you a smoking gun on this part.
But part one alone is enough.
If it can't be verified, it's not provably fair. Claiming otherwise is fraud.The fix is increment the nonce like every other dice site on the planet. They won't do it. Their original games were clearly designed for nonce increments, and they modified the system. Players have complained about this a year ago.
Betpanda should provide an explanation for why their autobet system doesn't increment nonces, and give an ETA for implementing a proper provably fair system where all rolls can be independently verified.
If neither happens, I will be leaving detailed negative trust feedback and encourage others who have tested this to do the same.
This is an anonymous casino.
Don't trust. Verify.
Scammers Profile Link: https://bitcointalk.org/index.php?action=profile;u=3675233Reference Link: https://betpanda.io/en/ (To test provably fair, use autobet on dice or other games, and observe that you cannot verify all bets other than the first bet. For example, if you make 100 bets on autobet, you cannot verify 99 bets since they change the server seed without providing you the hash instead of incrementing the nonce)
https://imgbox.com/g/H6w0cpUcao (Example results of autobet and manual bet where client seed is kept the same while server seed changes)
Amount Scammed: Unknown, probably in the millions of dollars
Payment Method: Crypto
Proof of Payment: N/A
PM/Chat Logs: N/A
Additional Notes: N/A
