Provably fair vs. Live Dealers

[a1choi]

Wanted to change up the gambling forums a bit with some debate instead of fighting over ponzi's or reading about advertised casino threads being bumped again and again.

Do you guys consider a Live dealer casino as provably fair?  And if not, do you consider it fair enough to wager your money on?

Also, do you consider licensed casinos as fair?  Will we see a trend of real-money casinos offering provably fair algorithms to their userbase?

from my POV, I don't consider live dealers as provably fair, but see it as a fair alternative, as long as they are a licensed operator in the real-money world.

[1714550270]

1714550270

[1714550270]

1714550270

[1714550270]

1714550270

[1714550270]

1714550270

[1714550270]

1714550270

[1714550270]

1714550270

[koshgel]

Not really sure if I'm understanding the questions entirely although I do like the new topic.

Are you asking if casinos will release the hashes from their algorithm so that customers can compare to see if they are being swindled?

Live dealers either hand deal or use auto-shuffler. Are you asking if these are truly random shuffles? Or if they would cheat for the house?

[HighIQSociety]

Do you guys consider a Live dealer casino as provably fair?

Per definition they are not provably fair. Do you mean probably fair?

[g4c]

My thoughts:

I think provably fair lubricates the entrance door to any casino, but when inside most patrons will not audit it... most will be lazy and assume others will audit it.

With Bitcoin games, the percentage of users who will audit is higher than fiat casinos because bitcoiners tend to be computer geeks. Plus completely unregulated bitcoin ecosystem at the moment is like a jungle, full of scam-snakes. So it's more important for a Bitcoin casino to implement provably fair than it is for a fiat casino.

I consider licensed casinos as probably more fair than unlicensed ones, but knowing how apes and money operate this licensing only adds a little trust for me, not the 100% trust that provably fair gives. I can imagine a scenario where slots are tweaked a little and the auditor turns a blind eye in exchange for an evening with Candice etc.

I think fiat casinos will offer provably fair systems in the future, the first ones to implement it will gain some market edge. Of course their user base knows nothing of one way hashes, so how to implement it best, not sure?

Auto-shufflers in casinos: I think to rig one to create sequences would be next to impossible without enlarging its size. but you never know one might have been rigged at sometime.

But this raises an interesting question: how do you make a 6 deck blackjack game provably fair online?

Sure we can transmit a hash of the deck order prior to starting the shoe. But how to prove the order has not been constructed to favor the dealer by a small amount, while mantaining a seemingly random order, ordering like this would allow system to set the odds in stone in 1-on-1 where user employs rigid strategy and counts. I think the method to prove fair card distribution would require too many deck samples and too much work for users to audit?

[zolace]

My thoughts:

I think provably fair lubricates the entrance door to any casino, but when inside most patrons will not audit it... most will be lazy and assume others will audit it.

With Bitcoin games, the percentage of users who will audit is higher than fiat casinos because bitcoiners tend to be computer geeks. Plus completely unregulated bitcoin ecosystem at the moment is like a jungle, full of scam-snakes. So it's more important for a Bitcoin casino to implement provably fair than it is for a fiat casino.

I consider licensed casinos as probably more fair than unlicensed ones, but knowing how apes and money operate this licensing only adds a little trust for me, not the 100% trust that provably fair gives. I can imagine a scenario where slots are tweaked a little and the auditor turns a blind eye in exchange for an evening with Candice etc.

I think fiat casinos will offer provably fair systems in the future, the first ones to implement it will gain some market edge. Of course their user base knows nothing of one way hashes, so how to implement it best, not sure?

Auto-shufflers in casinos: I think to rig one to create sequences would be next to impossible without enlarging its size. but you never know one might have been rigged at sometime.

But this raises an interesting question: how do you make a 6 deck blackjack game provably fair online?

Sure we can transmit a hash of the deck order prior to starting the shoe. But how to prove the order has not been constructed to favor the dealer by a small amount, while mantaining a seemingly random order, ordering like this would allow system to set the odds in stone in 1-on-1 where user employs rigid strategy and counts. I think the method to prove fair card distribution would require too many deck samples and too much work for users to audit?

you dont know anything one casino licensed was changing there software code after the audits

[g4c]

My thoughts:

I think provably fair lubricates the entrance door to any casino, but when inside most patrons will not audit it... most will be lazy and assume others will audit it.

With Bitcoin games, the percentage of users who will audit is higher than fiat casinos because bitcoiners tend to be computer geeks. Plus completely unregulated bitcoin ecosystem at the moment is like a jungle, full of scam-snakes. So it's more important for a Bitcoin casino to implement provably fair than it is for a fiat casino.

I consider licensed casinos as probably more fair than unlicensed ones, but knowing how apes and money operate this licensing only adds a little trust for me, not the 100% trust that provably fair gives. I can imagine a scenario where slots are tweaked a little and the auditor turns a blind eye in exchange for an evening with Candice etc.

I think fiat casinos will offer provably fair systems in the future, the first ones to implement it will gain some market edge. Of course their user base knows nothing of one way hashes, so how to implement it best, not sure?

Auto-shufflers in casinos: I think to rig one to create sequences would be next to impossible without enlarging its size. but you never know one might have been rigged at sometime.

But this raises an interesting question: how do you make a 6 deck blackjack game provably fair online?

Sure we can transmit a hash of the deck order prior to starting the shoe. But how to prove the order has not been constructed to favor the dealer by a small amount, while mantaining a seemingly random order, ordering like this would allow system to set the odds in stone in 1-on-1 where user employs rigid strategy and counts. I think the method to prove fair card distribution would require too many deck samples and too much work for users to audit?

you dont know anything one casino licensed was changing there software code after the audits

ha ha, worse still, allegedly:

Harris was a 12 year employee of the State Gaming Control Board who was assigned to evaluate gaming devices. He was arrested in 1995 in New Jersey for using a computer program to win $100,000 on a Keno game. He was later charged in Nevada for rigging slot machines to pay fraudulent jackpots and is now on parole for this crime and residing in Las Vegas. Harris was banned from New Jersey casinos in 1995.

http://gaming.nv.gov/index.aspx?page=190

[HighIQSociety]

I guess licensed casinos are probably As Fair As You Can Get^TM, followed by "provably fair" casinos. Provably fair is under debate and I have yet to see an implementation that can actually show and make the average player understand that they are in fact fair.

[a1choi]

I guess licensed casinos are probably As Fair As You Can Get^TM, followed by "provably fair" casinos. Provably fair is under debate and I have yet to see an implementation that can actually show and make the average player understand that they are in fact fair.

this is a great article that highlights the weakness of provably fair in the current sense.  i wonder if TrevorXavier ever got around to releasing his improved version 2.0 of provably fair as he alluded to in his comments.

The onus seems to be on the player to control the conditions for fairness which should be explained clearly on every site that purports to be provably fair, as well as clear instructions to the user on how to change client seeds.

Having a trusted and responsible 3rd party "regulatory board" which does not have any industry ties might not be a bad idea to implement.  But not sure how many of us in this industry would agree to it.  3rd party audits of software could also help?

I think i'm veering off track from my initial topic of comparing provably fair to live dealers but perhaps that isn't so much the issue  (as pointed out in an earlier comment, by definition, live dealt cards can't be provable).


Regarding the 6-deck shuffle problem, i believe releasing the source code might be a good start and having a 3rd party audit (perhaps by real money auditors).  I don't think there is much harm in releasing shuffle code, perhaps making it open source even.

[zolace]

is bad cause then someone can change code and hack so it cant be open source and open source wont do anything but bring casino more trouble with hackers.  So is bad Idea

[Rannasha]

Provable fairness is the best approach available. However, it is not without drawbacks. It is up to the player to set his own client seed as he can't rely on the seed pregenerated by the casino. It is also up to the player to verify the bets afterwards. The problems outlined by TrevorXavier in his Reddit post mentioned above are for a specific casino and he says in his post that these issues don't necessarily occur in other implementations.

While auditing and licensing by an official organization sounds reasonable, it is also rather weak. A licensing agent can be bribed. The casino could also run two versions of their software, a fair version to activate during audits and an unfair version otherwise. Switching between different versions could be done with a push of a button, so even surprise audits shouldn't be an issue (unlike physical casinos where you need to make physical adjustments).

Finally, having an auditing and licensing board is a centralized authority which is orthogonal to purpose of Bitcoin.

[zolace]

Provable fairness is the best approach available. However, it is not without drawbacks. It is up to the player to set his own client seed as he can't rely on the seed pregenerated by the casino. It is also up to the player to verify the bets afterwards. The problems outlined by TrevorXavier in his Reddit post mentioned above are for a specific casino and he says in his post that these issues don't necessarily occur in other implementations.

While auditing and licensing by an official organization sounds reasonable, it is also rather weak. A licensing agent can be bribed. The casino could also run two versions of their software, a fair version to activate during audits and an unfair version otherwise. Switching between different versions could be done with a push of a button, so even surprise audits shouldn't be an issue (unlike physical casinos where you need to make physical adjustments).

Finally, having an auditing and licensing board is a centralized authority which is orthogonal to purpose of Bitcoin.

agreed

[dolphin2014]

A licensed casino doesn't mean a reputable or fair casino. There are plenty of licensed online casinos that are dodgy.

[vit1988]

Maybe a open source/3rd party browser plugin automatically setting the client seed and verifying the results would be handy for players. Casinos could implement a standard API to be compatible with it.

About live dealers and auto shufflers: I've once seen a card trick player (he was working as a consultant for casinos) in real live, he was so skillful I have absolutely no idea how he managed to get four aces _every_ time after I shuffeled the deck on my own. And I've seen david copperfield on TV...  No: I definitively don't believe in what I see on a live cam.

[g4c]

Provable fairness is the best approach available. However, it is not without drawbacks. It is up to the player to set his own client seed as he can't rely on the seed pregenerated by the casino. It is also up to the player to verify the bets afterwards. The problems outlined by TrevorXavier in his Reddit post mentioned above are for a specific casino and he says in his post that these issues don't necessarily occur in other implementations.

While auditing and licensing by an official organization sounds reasonable, it is also rather weak. A licensing agent can be bribed. The casino could also run two versions of their software, a fair version to activate during audits and an unfair version otherwise. Switching between different versions could be done with a push of a button, so even surprise audits shouldn't be an issue (unlike physical casinos where you need to make physical adjustments).

Finally, having an auditing and licensing board is a centralized authority which is orthogonal to purpose of Bitcoin.

I agree that centralised authority is not a good solution.

User set client seed is a complication that I believe can be removed in simple games without compromising fairness:

consider a coin toss game where user gets 2X return for correct prediction:

coin states:

"H" = coin lands 49.5% heads
"T" = coin lands 49.5% tails
"E" = coin lands 1% on its side (house edge)

here's how fairness goes down:

[1]
server creates coin state using dev/random.

[2]
server creates a message string:
UnixTime . Salt . CoinState --->  "1394727501_sdDHSDa4324rfrsd_H"

[3]
server hashes string and sends it to user, user can copy to clipboard if they wish to audit.

[4]
user sets bet amount and makes prediction.

[5]
server processes win/lose and sends non hashed message string to user, user can verify it hashes correctly.

the timestamp removes the possibility of the server using a list of pre generated hash collisions.

user would have to gather rounds of data to analyse random distribution (check coin does not land on its edge too much), maybe the client application could graphically show the distribution as rounds progress.

[a1choi]

i had thought that time was not a random enough variable.  I'm not a computer scientist, so i'm not 100% certain, but i think this method would fulfill the provable fair requirements, but not the random requirements.

[g4c]

i had thought that time was not a random enough variable.  I'm not a computer scientist, so i'm not 100% certain, but i think this method would fulfill the provable fair requirements, but not the random requirements.

thanks for looking over it.  yes the random proof would be needed also for good solution.

the time is just to satisfy the user that the server cant use a massive lookup table with hash collisions.