The wheel spin result is provably fair in the sense that the wheel spin and player selection is determined by bits from a hash that we could not have chosen (just like as any crash game that uses a chain of hashes that is salted by a random event like Bustabit or any Bustabit clone).
We have been working on a handy verifier for the selection of who spun the jackpot wheel. It is based on multiple random factors like bet size, usernames participating in the round, and bits from the round hash. Rounds can be verified here:
https://jsfiddle.net/kyawgjmv/embedded/resultExcept that's not what provably fair means. You can
directly control who wins the jackpot, if you have a house-player. And there's absolutely zero way for people to have any idea if you're doing it or not.
That's not bad per se, and doesn't mean you're cheating. But it means it's not provably fair. Full stop. Any attempt to market is as provably fair is dishonest.
You are right in the sense that it is technically possible to manipulate who wins the Jackpot, and for this reason we have clarified our specifications of the system on
https://rocketpot.io/provably-fair/. On this URL, we make it clear that the outcomes can be verified with the given player list: “Given the player list (that consists of player names and their bets) and round hash, anyone can verify the player selection result.”.
More generally - it is indeed very hard to completely prove the fairness since the house/server can influence which wagers get processed or not. Winning wagers could get timed out and never processed, or there could be other creative ways for the house to reject or limit winning wagers, even in a simple dice game.
a) Make the jackpot provably fair
There's a few ways to do this, one neat way would be run a verifiable-delay-functions on the game player bet list, where each client also sends a randomized seed with their bet. Then if the game goes over XXX you use that result to calculate the jackpot winner.
That would also hit pretty much the same problem we started with, just making it a matter of processing power cost to manipulate the player list. Once the jackpot is big enough, it would be profitable to spend computing power and resources to Brute Force VDF as the house could add its house player’s seed. Also, it would introduce a player list trust issue since the player list would have to be posted to a trusted 3rd party (like a blockchain) before the round starts.
We’ve also considered using a delayed source of random, such as a block hash from Ethereum (averages less than 30 sec), using the first occurrence since the round started. But that also hits player’s list trust and doesn’t remove the possibility of a house player to just bet high in rounds with high multiplier (the jackpot wheel will spin).
b) Just let everyone know "our multipliers are provably fair but our jackpots aren't! it's just a bonus"
The multipliers are mathematically proven to be fair. We are currently unable to prove mathematically that anyone that wants to wager on them is able to do so in a fair way (internet timeouts, house players that can in theory manipulate the auto cashout). However, this seems to be the case for all the other provably fair multiplier games with bankroll and profit limits.
And the jackpots have a theoretically similar problem: The wheel spin stop is mathematically proven to be fair, but we are currently unable to prove mathematically that anyone who wants to wager on it is able to do so at all, lest in a fair way.
I hope this clarifies it somewhat. Please note it was not our intention to be misleading and we are working on making the game as ‘provably fair’ as possible.
I’d like to add that we’ve had some players win some fairly big jackpots already and have paid them out, user Bsky being one of them (he posted in this earlier). We’re also working on a page that will be listing all the biggest jackpot wins and the players who won them (with a link to the round).