Bitcoin Forum

I see some gambling sites mention that they are provably fair. They provide some seed and ask to check some hash or whatever. My question is, can this fairness really be checked from outside ? Where are they twisting to make sure the house's win ?

Yes, one way to verify a game’s fairness is through the combination of a server seed and a client seed. We can compare the hash with the server seed that has been revealed by the house. If we find a match, then we can say that the game has a provable fairness.

So, basically a fair game result is calculated as follows:
1. The server creates a hashed server seed and sent to the player.
2. The player creates his own random client seed
3. The combination of the two seeds are used to calculate the result of the game.
4. The server seed is revealed to the player so the player can check wether it matches the hash that was sent earlier

Yes, it can.


I can explain it on simple example, (just the principle)

BTCRock - it's game where you have a grid 20x20 and you have 6 attempts to find gold vein.


Their proof is that immediately after solving captcha you can see hash of "gold vein coordinates ".

http://s1.postimg.org/5f9zcv9jz/Rock1.jpg



After you made your 6 tries, a result is shown to you:

http://s14.postimg.org/65pntyewh/Rock2.jpg
 



Proof is that from begining the coordinates  gold vein was [11,9], and was not moved if you luckily hit it ( it would change the hash)

You can check it using: http://www.xorbin.com/tools/sha256-hash-calculator

1 - enter: 65879c50984f5b9ede009960e6efa3c6|[11,9]    (proof of location, shown after game end)
2 - hash it (using sha2586)
3 - you'll get: c0344a8d187371c671d9e16cb27190a74caef8fe32ab66be7287ed050fca2b12       (this is what you know from begining)

tip: for clarity try to change one coordinate and hash it again, will see you got completely different hash.


P.S. This's only principle, and can be used in various way. You should check description on particular site.

These sites tend to use slightly counterintuitive terminology. Provably random? Yes. They can't pick results in advance; nor can you. Provably fair? No. They still get a bit of an edge to keep a profit, although they can be verified to not be taking more than their official rules say.

With a bit of php knowledge you can easily make your own 5 line script checking any websites fairness.
Look at my code for generating random numbers (pretty similar to other websites)
I take the TXID of the bet, the amount of bets, and then the daily secret key which the hash of the secret key is stored for people to check the next day. If the hash of the secret key matches the one shown on the list, that means the secret key was indeed used. These pieces of data are hashed together in sha512 using hmac. I take the first 5 digits of that hash and convert it from hexadecimal to decimal. This gives a number that 98% of the time is under 1000000 (if it isnt I take the next 5 digits), and then divide that number by 10000 and I get a two digit number out of a hundred with some decimal places. The next day the user can verify their bet by pasting in their TXID, their amount of bets, and then the secret key that is released the next day, and then run it through the script. If the echod number matches the number that was used on the gambling site, I was indeed telling the truth and their bet was fairly made. Every website explains their own method of fair betting.

I would say that they are fair in regard to the fact that they disclose the odds of winning ahead of time, and the actual odds are what they say they are.

Thanks wusolini, you explained very well, I also always thought how probably fair work.

I thought a seed can be manipulated..

Yes, but each side keeps a secret from the other side, so in the end all you can do is change the outcome without knowing what you change it to.

not when half of the seed is secret and bit revealed until the next day

cant the casino just go back to learn your betting style?

The odd of winning is already on house side. Pointless to cheat and risk damaging the reputation.

That is an awesome explanation. I wonder how the house make sure that they win in the long term ? Do they really gamble with the user choice ?

They design the game so that in a fair, unmodified implementation under the rules, they get a house edge. A basic example would be a game that pays 2x a bet with provably fair "rolling", but you win 48% of the time. For instance, it picks a number between 1 and 100, and you win if it is less than or equal to 48.

Keep in mind even "fair" house comes out ahead.  The place you gamble at always will win over 50 percent overtime, if not it will go out of business.  You are betting on being one of the "lucky" ones.

Gambling sucks, you will lose in the long run due to house edge.

If you change the client seed randomly from to time, the house can't get advantage (other than the house edge) even if it knows your betting style.

Yes, even if you play in provably fair websites.
There are many betting methods to try, but no one can ensure a long term win.
You are playing against the site owners, not against other players, therefore you are fated to lose.

There are possible issues with that probably fair stuff, see this topic

https://bitcointalk.org/index.php?topic=283547.0
 (https://bitcointalk.org/index.php?topic=283547.0)
And this article


for more info(article cited in the topic):

http://www.reddit.com/r/Bitcoin/comments/1frm4x/provably_fair_by_bitzino_not_provable_with/
 (http://www.reddit.com/r/Bitcoin/comments/1frm4x/provably_fair_by_bitzino_not_provable_with/)

There are also topics questioning if some sites are provably fair or not, you might want to search that subforum and see by yourself.

Damn just went to the other thread.

Thanks for exposing that. Provably fair can be exploited..

But thats the thing with gambling, u gotta know when to stop.

Even when there is no house edge, you will lose to the house in the long run since you have a lot less capital than the house (gambler's ruin). :D

How come having less capital confirm loss for an user ? ???

I own a gambling website, so if anyone has any questions about provably fair ask me.

i don't know how to check if a game is provably fair, and if it is fair at all, but i did play at many many sites that say ''provably fair''. I wagered quite a lot at some of them and i didn't notice they were cheating, so i don't think that any of them are cheating. They don't even have no need to cheat, most of those sites are very profitable anyways.

What is the URL of your gambling website ? :)

Better don't bet or only bet on live casinos, if you do.

I don't trust the provably fair allegances.

The point of provably random sites is that you can verify the randomness using cryptography.

Not everyone understand crypto, so not everyone can verify without blind trusting someone else

And even if crypto theory is perfect, there are still flaws that the implementation can have, so there is aways some room to your inner paranoid bothers you, even if you are expert(after all, NSI managed to put backdoors in most of the major internet cryptography)

I think by NSI u mean NSA. But what r u saying afterwards ? There are backdoors to SHA-256. Then, NSA can calculate the private key from a bitcoin address ?

I doubt that the NSA would be trying to scam you out of an extra few percent of your gaming rewards. That being said, SHA being broken is not sufficient to get private keys out. ECDSA would need to be broken (either through solution of discrete log problem, or some more creative way) before private keys can be obtained.

But, the person above me is claiming every cryptography has a backdoor. Is not there an open source cryptography where we can see what is really happening inside ? In an opensource project like Bitcoin, why a closed source SHA has been implemented, if the probability of having backdoor cant be verified ?

Don't believe everything you read on the internet.  What makes you think SHA-2 is "closed source".  The algorithm is publicly available.
http://en.wikipedia.org/wiki/SHA-2#Pseudocode


That is a cornerstone of any strong cryptography.  The algorithm is still secure even if the attacker knows everything except the secret.  If a system needs to keep the algorithm secret in order to remain secure then it is nothing more than feel good security and will be broken wide open relatively quickly. There are no widely used cryptographic system where the algorithm is a secret.  All the algorithms of the cryptographic primitives used in Bitcoin are public knowledge.

As i can see in the algo, there is no backdoor as such. Then why it is said time & again that NSA may have a backdoor to decode SHA-256 ?

If you can prove there is no backdoor, then you either a) know something very mathematically powerful that nobody else knows yet, or b) are misguided as to what a backdoor is.

Nobody has been able to mathematically rule out a backdoor so far, even in an "open-source" algorithm. Remember heartbleed? That was a backdoor in an open-source cryptosystem, that may have been known by some organizations, that may have been able to use it before others even knew about it.

A backdoor is not something that is plainly evident when looking at the code; in fact it should not be evident at all. However, some organization may have knowledge of some mathematical procedure that they can use to reverse a hash, find a collision, or otherwise do things that a cryptosystem assumes they cannot do.

Ok... then it is my ignorance. By backdoor, I meant, some secret code/lib used the o/p of which is only known to NSA. Or at least the i/p and a function by which the secret code is generated, e.g. MS's PRNG, where MS does not know the o/p, but given the correct i/p variables, they can calculate the o/p as the function is known to them.

You're welcome.


I'm glad you understand, I've tried to make it in simple.

there is always a some formula, that guarantees long time profit to operator. And if not, it can be just adapted  ;). So don't worry, casinos will never bankrupt (as long as they have active players)

Even with no edge, the variance will hit you in the long run. Even though mathematically, you are not suppose to hit X loses in X rolls, it can be faster since you will never experience the expected streak. Rolls does not affect each other too, if you hit a bad streak, your next streak may not necessary be a win streak.

So, do u mean, the program inside is rigged for the long run that make sure the house will always win ?

I think he meant house edge can affect the profits. Variable house edge will increase the house edge if the house lose BTC and decrease when the house have more profits.

The house edge would bring profits to the house in the long term, but the house could lose in the short term.
And so casinos need to set the max win correctly (for example, using Kelly formula). :)

Not rigged, but calculated. It's all about math and likelihood
something like: unlikely you win , even more unlikely you win a jackpot and likely you loose, but most likely the game wins.

But it's a different story, the game is still provably (you can check it) fair (result is random or known ahead or ... )
well calculated game don't need to cheat, just need to sustain as long and get as much players as possible    



Thanks for pointing out Kelly formula  ... it's all about math

I understood the grid example. But, I wonder how a dice game can be provably fair, because we may verify SHA256 of the amount that will come after roll beforehand. But, how to know, that this amount is itself not rigged, i.e. while playing a high roller where the amount needs to be more than 75, if the system takes the amount as 67, player will always lose. Am I missing something ?