That's a different issue altogether. Wallet security (cold and hot and online) and who the operator is, should be a different factor.

See, there are several levels of how you want people to trust in your site, both as a player and investor:

1. typical large fiat USD casino = audited RNG, licensed by some country, pokerstars, etc. = not provably fair
2. trusted bitcoin poker site = seals = not provably fair, but is time tested
3. provably fair for players = typical decent dice site, or card game or bouncing balls or lottery or whatever. = at least players can check
4. provably fair for investors = should also be fair for players = no good solution yet, or practical or viable.

At this point in time, the minimum requirement (except for poker, so far) should be at least number 3. Provably Fair for Players.

The other question is one for the operator. Does he have a threshold before disappearing and stealing other people's coins? If he or they do, what is it? Are they competent? Did they do best industry practices (such as using cold wallets properly), are they using time-tested hash functions such as HMAC-SHA512. Are they upright citizens of their own country with no records, are they known, does someone have their info?

All those other questions are of the non-technical nature.

I can be as honest as the Holy Father the Pope (of the Catholic Church), a living saint, but if I open a bitcoin game site that is not provably fair, no one is going to play there, let alone invest in it.

I doubt any site is provably fair for Investors. Even if the site owner says it, there is no way to know.  They could always be using seeds against the site investors for small amounts which accumulate to be big over time. And if the site is profitable all that time, then no one would even notice such practice.

That seems like a hole in your system.

"in advance" of what? Not in advance of publication of the result, or of the betid.

As the dice site owner, you (the player) send me a bet request.

I make up a betid and send "betid and SHA512(betid:amount:chance:serverseed:clientseed)" to the audit server. Then I can see if the bet wins or not. If it wins, I have the site cheater account own the bet and win with it, then make a new betid for you and repeat.

Basically you'll letting the betid affect the bet's outcome, so I can keep trying different betids until you lose. For all the ones you would have won, my cheater account wins those.

I have one solution. 1 second block time alt-coin. All investors (and players) run the client, peer to peer, hosting it's corresponding blockchain. It won't even be used for it's coin properties, as the coin itself will not have much value, the value is that this "audit server" is in the hands of everyone else but the dice site.

The problem is implementing it. And a bunch of other technical details.

Then again, maybe that is not ideal, and until we can dissect it some more, not sure if it will actually work. (Will it mine? is it proof of work or proof of stake or sha256 or scrypt or X11 or X19 ... and all those other questions; how about a fork or a 51% attack, etc.)

In OP I only said "- The user probably wants the hashed audit seed in advance, but this could be possible I think." but I will explain that further.

1. The dice site generates a server seed, he will request a server-seed-hash-specific audit seed HASH too. And shows it to the player.
2. The player makes a bet with his client seed.
3. Dice site sends bet information (with server HASH and other details including amount,chance,etc so this can't be changed after the result) to the audit server. Audit server uses nonce (and checks if it's sequel) + audit seed to calculate the SHA512 of that combination. And returns that.
4. Dice site uses: client seed, server seed, nonce, SHA512 of audit seed + nonce > to calculate outcome.
5. Gives result to player and repeat.


In the end the player can request a new server+audit seed. So this will:

1. Player generate new seeds.
2. Dice site requests the audit seed with the original audit HASH and the server seed. Audit server gives audit seed and makes that seed invalid for future bets. With the server seed the audit server could recalculate all bets of this session.
3. Dice site returns the server + audit seed to the player.
4. Player can recalculate all bets with those seeds. He can also check the audit website with his audit seed to check all recalculated bets.


Still I think the "site profit" could be faked. So there would have to be a daily, weekly or monthly verification.

1. Dice site generates new server seeds.
2. Dice site sends old server seeds to audit site.
3. Audit site recalculates all rolls and can even calculate the site profit this way.



It helps against:
- skipping nonces changing seeds etc
- easier verification for players
- fake whales and/or profit (loss)
- hackers cannot win with a server seed or cheat unnoticed



Would that work?

Obviously, once again, players+investors would need to trust the audit server. But it's better because BOTH the dice site and audit site would need to collude/cheat together. Is it worth the trouble? Not sure

There was lottocoin, and there is that casinocoin, and there is the chancecoin. dogecoin used to have semi-random block rewards for miners. But that got abused (people would only mine blocks with high rewards, since you could see a block before what the reward will be for the next block.)

Worthless alts. hahahahahaah.

But my idea was not an on-chain dice game. Still off-chain, but using the chain as the source of entropy for server seeds or audit seeds or something. The problem is the speed.

Here, we may not necessarily talk about block times, but transaction times. Transactions are near instant in bitcoin. Blocks take 10 minutes. Double-spends would have to be done within 5 to 10 seconds otherwise they lose the race.

The only way for something to be provably fair for investors, is if the site can't predict the outcome, therefore the roll result will have to be revealed after both client and server seeds have been committed. Nonce counting too, if that is used.

The only game that I can imagine can be provably fair for everyone are the slow ones. Lottery draws, for example. Or common table betting games, where multiple players bet on a single outcome. Or sportsbooks (you have to wait until after the game, right?)

Dice is just too fast. And no one wants to play slow dice anymore.

In the real world, lots of people play that numbers game which has 3 daily draws. No one plays the twice a week draw game, despite having a jackpot of millions. (okay, they do, but I was trying to make a point.)

hi

I know it is an old thread.

are there any new developments regarding this great idea?

thx

are you saying that I didnt understand the main posting by NLNico? or maybe you didnt understand

Nope you understood perfectly.
I'm saying that the whole idea is a bit useless, that's different.
And also saying that because it's a bit useless, the project will not really move.