bustabit.com -- The Social Gambling Game

[dooglus]

Ryan, I stumbled upon this novel way of making huge sha256 chains efficiently:

  https://github.com/rustyrussell/ccan/blob/master/ccan/crypto/shachain/design.txt

Instead of a linear chain of a few million sha256 hashes for provable fairness, he proposes a 64 dimensional hypercube of hashes, giving 2^64 hashes with very little precomputation required.

I thought it was kind of an interesting idea, extending the usefulness of the technique you're using at bustabit.

Currently if I want to verify that you're using the "correct" hash for game #123456 I need to run 123456 sha256 hashes and check that I end up with the genesis hash. Under Rusty's scheme the same could be achieved using something in the order of just 20 sha256 hashes.

Not that it's worth changing at this point in time. But it may be helpful for your next venture.

[1714179823]

1714179823

[1714179823]

1714179823

[Erza]

Any strategies how to play here?

You should always try your best to get the bonus (reference: https://www.bustabit.com/faq#last) so to overcome the house edge of the game. What you should do to achieve that depends on the the other players' bet sizes and cashout points.

But I think the bonus is not too high compare to what you lose. The more multiplier you get the higher the bonus right? So try to keep aim bonus is not really useful. Usually I played low after sometimes then try to hit higher multiplier. It works better most of the time

To your question of "The more multiplier you get the higher the bonus right?", the answer is "not necessarily". For example, if you are having the biggest bet size in that round, you will get the max bonus as long as you are the last one to cash out. On the other hand, if you are not the biggest bettor in the round, you could cashout "midway". Also, the house edge of bustabit is from 0% to 1% depending on your cashout point, and the bonus could be a lot higher than the house edge, making the game +EV.

But to aim its bonus I dont think that you can win all games right? Sometimes you must take some loss too so I think there is no way you can earn just through the bonus thing. And btw how much is the max bonus? Depends on how much you bet and your multiplier?

[dooglus]

But to aim its bonus I dont think that you can win all games right? Sometimes you must take some loss too so I think there is no way you can earn just through the bonus thing. And btw how much is the max bonus? Depends on how much you bet and your multiplier?

There's no maximum bonus. It depends on how other players bet.

If you and 9 other people each bet 1 million bits, nobody else plays, and you are the last one to cash out, you get 10% of your bet as a bonus. If you notice they're all cashing out at 2x you can set your cashout point at 2.01x. Then half the time you'll be getting a 10% bonus, which only losing 0.5% to the house edge on average. That's enough to make a decent profit over time.

Edit: the last person to cash out gets a bonus percentage which is (total_stakes / biggest_stake). In my example the total_stakes were 10 million, and the biggest stake was 1 million. 10 million / 1 million = 10, so the bonus is 10%.

So if there were 200 people each betting 100 bits, the last one to cash out would get a 200% bonus!

[Erza]

But to aim its bonus I dont think that you can win all games right? Sometimes you must take some loss too so I think there is no way you can earn just through the bonus thing. And btw how much is the max bonus? Depends on how much you bet and your multiplier?

There's no maximum bonus. It depends on how other players bet.

If you and 9 other people each bet 1 million bits, nobody else plays, and you are the last one to cash out, you get 10% of your bet as a bonus. If you notice they're all cashing out at 2x you can set your cashout point at 2.01x. Then half the time you'll be getting a 10% bonus, which only losing 0.5% to the house edge on average. That's enough to make a decent profit over time.

Edit: the last person to cash out gets a bonus percentage which is (total_stakes / biggest_stake). In my example the total_stakes were 10 million, and the biggest stake was 1 million. 10 million / 1 million = 10, so the bonus is 10%.

So if there were 200 people each betting 100 bits, the last one to cash out would get a 200% bonus!

But that is impossible to keep on betting 1 mil bit and being the last one to aim the bonus. Because like I said there is no way you could always win the game. Usually I saw big player always using 1mil bit cash out at below 1.5 so they will at least some percent of the bonus rather than losing everything. And for getting 200% bonus I dont think someone can manage it to be the last person to cashout. It is risky

[Falconer]

But that is impossible to keep on betting 1 mil bit and being the last one to aim the bonus. Because like I said there is no way you could always win the game.

There is no way to keep winning the game but since it is EV+ then you are bound to get profit in long term. Ryan had stated in the past that it is an EV+ if you keep on doing so. He even stated it in his FAQ that it is the only way to overcome the house https://www.bustabit.com/faq#last but no one has actually tried to do some experiment yet regarding this

[redsun114]

Played a few chances today after a lot of days and finally again loving it now

[RHavar]

Kapi18wro has finally hit his goal! A net profit of 100 btc! 




A big congratulations!

[Dabs]

Currently if I want to verify that you're using the "correct" hash for game #123456 I need to run 123456 sha256 hashes and check that I end up with the genesis hash. Under Rusty's scheme the same could be achieved using something in the order of just 20 sha256 hashes.

doog, please correct me if I'm wrong or if the method has changed since I last remember it, but wasn't a genesis hash or chain hash generated last year for 10 million games? And a then future block hash was used. Combined, is the game hash, but you only need the next game's hash to verify the hash of the game you are currently playing.

You don't need to run all 123456 hashes, just check the next one when it's available. And maybe check the previous 100, at least you know that hash is from the same chain.

The chain all the way to the very first game should be saved somewhere else (or available from the site.) And just add to it one at a time as the game keeps running.

@RHavar, did you get my PM from about a week ago? (Should I email you instead?)

[dooglus]

Currently if I want to verify that you're using the "correct" hash for game #123456 I need to run 123456 sha256 hashes and check that I end up with the genesis hash. Under Rusty's scheme the same could be achieved using something in the order of just 20 sha256 hashes.

doog, please correct me if I'm wrong or if the method has changed since I last remember it, but wasn't a genesis hash or chain hash generated last year for 10 million games? And a then future block hash was used. Combined, is the game hash, but you only need the next game's hash to verify the hash of the game you are currently playing.

You don't need to run all 123456 hashes, just check the next one when it's available. And maybe check the previous 100, at least you know that hash is from the same chain.

OK, but you wouldn't know that the chain the game is currently using is actually the same as the 10-million-hash chain that was selected a year ago. Maybe Ryan made a new chain which has lots of low crashes on it, and we're playing through that instead. We wouldn't be able to tell the difference unless we hash all the way back to the genesis hash to check it.

I've no reason to think that's actually happening of course. I am just saying that to check it we would need to run thousands of hashes (one for every round that happened since the game became provably fair - so it will take longer and longer over time), whereas the method I posted about gets rid of that problem. That method never needs more than 64 hashes to get from the current game to the genesis game - but it is quite significantly more complex to understand, so it's a tradeoff.

[Dabs]

Yes, but a third party would be keeping track of the one and only 10-million-hash chain as it progress, from the beginning. Of course, then you'd have to trust that, so it goes back to calculating it yourself.

Think of it like the bitcoin blockchain, you don't need to download all 60 gigs of it if you've been updating your node often (or kept it online the whole time.)

The way it is set up, someone out there would quickly find out if the chain is different, being as popular as it is.

That method with only 64 hashes is indeed interesting, but, it is complex to understand. The little bit at a time, almost everyone can understand that. Maybe the new method just needs an explanation that is easier to understand than the one it has right now. I read through it, and, while I get the general idea, still find it difficult.

The way I imagine the complex method is 64 hashes at a time, or done in batches of 64, so you can jump in steps of 64 exponentially until you need to get where you want in the chain. Unless I understood it wrong. I'll need to reread that a few more times.

[RHavar]

I've no reason to think that's actually happening of course. I am just saying that to check it we would need to run thousands of hashes (one for every round that happened since the game became provably fair - so it will take longer and longer over time), whereas the method I posted about gets rid of that problem. That method never needs more than 64 hashes to get from the current game to the genesis game - but it is quite significantly more complex to understand, so it's a tradeoff.

Yeah agree. It's not really provably fair unless you can actually verify you're on the canonical chain or not. The paper is a great read, and does something I didn't realize possible -- but it solves a problem bustabit doesn't have. Namely, generating a hash chain of 10M (or 10x that) is pretty negligible.

I was thinking it could be pretty useful for something like MP to generate a hash-chain-as-a-service. Or maybe fast-paced multiplier game, where you burn through millions of games super quick.

[triggermage]

Just leaving my vouch & a quick review here.
I always deposit here when I have a bit of extra money, and sometimes I win high but I get greedy & lose. The withdrawals take seconds & I always get my money quickly. Definitely like the ChatBot, and it was the best thing I liked about the chat until they disabled it in the English channel Commands are pretty helpful though. Whenever I'm waiting for a confirmation somewhere I just go on Bustabit & type /block and wait for the noise alert so I know when it confirms.
The site has a nice UI and is great on both mobile & bigger screen devices. Would definitely recommend playing on here if you have some extra btc :-)

[StacksAMilli]

To the nigga that profited 100 BTC, good fucking job..  I do good then I get greedy and bet my whole amount and lose it all..  lol

[RHavar]

Thanks for the kind words everyone.

Awesome milestone we hit:

#1 wagered bitcoin casino!


Although it's temporary (our volumes are extremely high this week due to some lucky winners churning over their winnings, and PD is much bigger than us) it's still a really cool milestone!


Thanks for playing!

[RHavar]

There was a bug in the latest version of the game server, which resulted in some people losing money where the auto-cashout was not honored.  

The change has been rolled back, and I've gone through the games, and found everyone directly affected. Here's the complete list:

http://pastebin.com/raw/LTEn7kkk

username ; amount in satoshis  (divide by 100 for bits, 1e8 for btc)

The money has been credited to your account. If you think you're missing any, or were affected more adversely than that please contact me on bustabit.com/support and I'll be happy to help.

Sorry about that. Going through the code now, to see how the bug occurred.


-Ryan

[dooglus]

That method with only 64 hashes is indeed interesting, but, it is complex to understand.

Sorry for going offtopic here, but here's an explanation:

I made an implementation of a variant of it, where all hash chains are of length 10, so it works better for people who think in base 10 (ie. all of us).

Suppose we know that the 9999th hash is 2cf24dba5fb0a30e26e83b2ac5b9e29e1b161e5c1fa7425e73043362938b9824

From that we can get to the 9998th hash by adding 1 to the last digit of the above hash, and hashing it:

  sha(...b9825) -> 9998th hash

Or, we can get to the 9989th hash by adding 1 to the last but one digit and hashing it:

  sha(...b9834) -> 9989th hash

And so on:

We can get from the 9999th hash to the 8999th hash by adding one to the last-but-three digit and hashing it:

  sha(...ca824) -> 8999th hash

We would have to define what adding 1 to an 'f' does - does it carry a 1 to the left or wrap around to 0 without carrying. Either works, but we must decide which we are using.

And we have to note that you can only use this trick to get from one hash to the next if all the digits to the right of the digit you're decrementing are 9's.

So we can go: 9999->8999->7999->7989->7979->7969->7968 to get to the 7968th hash
But we can't go 9999->9989->8989->... because the bolded section is modifying the thousands digit when the tens digit isn't a 9.

Once hash 99 is published, you can calculated any lower numbered hashes. But from hash 103 you can only go back to hashes 100, 101, and 102 and none of the double-digit hashes (since you can't edit the hundreds digit unless the tens and units are both 9's).

So when we are on game 24648 for instance, we would need to be publishing the hashes for these games:
  24648, (gives you 24640 to 24648)
  24639, (gives you 24600 to 24639)
  24599, (gives you 24000 to 24599)
  23999, (gives you 20000 to 23999)
  19999  (gives you 00000 to 19999)

From those 5 hashes you are able to get back to any previous hashes using the rules laid out above, and in the worst case you'll be doing no more than 9*5 = 45 hashes to get to the hash you want (up to 9 hashes per digit, to get it from a 9 to a 0, and up to 5 digits in this example).

[EngiNerd]

i am having alittle problem ryan
please please fix it
 

You should post information about the problem you think you are having (or preferably PM them to Ryan)

[RHavar]

How to turn 1 bit to 100,000:

https://www.bustabit.com/game/2318863

Well done!

[cowbay]

How to turn 1 bit to 100,000:

https://www.bustabit.com/game/2318863

Well done!

technically, 100,001 bits 

[hopenotlate]

How to turn 1 bit to 100,000:

https://www.bustabit.com/game/2318863

Well done!

technically, 100,001 bits 

Right specification, your bet showed us all a single biut can make a huhe difference 

Were you chasing for it for a long time or was it a short period shot?