Indepth look into Stake RNG and the proofably fair system.

[iriz09]

its seems great indeed! agree🙋

[epasokmulang]

thaks for the info .. best of luck to us bro mistaker

[mistaker]

thanks a lot

 

[andylim690]

On 8/4/2020 at 4:31 PM, mistaker said:

To often I see phrases on chat like: "Stake is scamming me" or similar phrases. This inspired me to show you the proof that Stake simply cannot fool anyone.

Let's do a simple calculation using a regular google calculator (it can even be a calculator in your office or on your phone) and a simple HEX to DEC conversion system.

We all know that Stake Originals games (because we will only talk about them here) are based on a very simple result calculation system. When you start the game, Stake gives you a SERVER SEED that will be part of the score calculation system. You add CLIENT SEED, NONCE to this equation (I'll explain in a moment), choose GAME and SETTINGS, and what Stake does is check the equation and show you the results. Now let's check if it's actually TRUE and if we can actually check it.

Let's deconstruct this result and check it with a regular google calculator.

SERVER SEED – 691922054928c927be2cf70614e59b9d12fec05a8686396f008c8e0c4978b8ca

CLIENT SEED – 111

NONCE – 1 – is the number of bets I have placed on a specific game and setup

GAME – LIMBO ( as an example, but you can do it for each game )

To check the result, we need to know UNHASHED SERVER SEED. You can do this on the Stake website (https://stake.games/provably-fair/server-seed-unhash) after changing the active S / C SEED set.

UNHASHED SERVER SEED - 691922054928c927be2cf70614e59b9d12fec05a8686396f008c8e0c4978b8ca

NOW, CHECK USING THE GOOGLE CALCULATOR

If you go to the result calculation section (https://stake.games/provably-fair/calculation) you can paste all this data.

SERVERSEED:CLIENSEED:NONCE after converting to HEX gives you the result you see in the section CASINO SEEDS TO BYTES. Gra Limbo use first 4 parts after swapping.

So let's calculate it according to the formula given by Stake.

HEX to DEC

87 – 135

f4 - 244

e3 - 227

30 - 48

It's a simple replacement and you can check it by typing in google - HEX to DEC converter.

Now we count the result and check the Stake. We have to do some math just like Stake does every time you place a bet. What are these activities?

135 / (256 ^ 1) =

244 / (256 ^ 2) =

227 / (256 ^ 3) =

048 / (256 ^ 4) =

example:

notice that the result is correct

Now we repeat the operation with the next 4 digits from ours SEED'u.

135 / (256 ^ 1) = 0.527343750000

244 / (256 ^ 2) = 0.003723144531

227 / (256 ^ 3) =0.000013530254

048 / (256 ^ 4) =0.000000011176

We add the results together, which gives us the result 0.531080435961. Then we multiply the result by 100000000 = 53108043.596148490906 and last action : 100000000 / (53108043 + 1) * (1 - 0.01) = 1.8641243876351385 – as stated in the RNG principle for this game, which is on the Stake website.

RESULT FOR THE PLAYER – 1.8641243876351385

RESULT ON PAGE – 1.8641243876351385

IN-GAME RESULT -  x1,86 MULTIPLIER

As we can see, the result is in line with the bet. The same calculation can be made for each game (PLINKO, KENO, SCARAB SPINS, etc).

If we can do it using a regular google calculator. HOW CAN A STAKE CHEAT? THERE IS ONE ANSWER - THERE IS NO SUCH POSSIBILITY !!!

 

Thanks will try 

[Savatage79]

Not saying anything is a scam but I'll just say the rng for me has been fn atrocious.  Can't do anything , every game is the same...doesn't matter how long I put into games I never just have that moment. Had like 2 small cash outs but now all that happens is I spin my wheels on every slot, frustrating 

[nuuuitsjdragon]

On 8/4/2020 at 4:31 PM, mistaker said:

To often I see phrases on chat like: "Stake is scamming me" or similar phrases. This inspired me to show you the proof that Stake simply cannot fool anyone.

Let's do a simple calculation using a regular google calculator (it can even be a calculator in your office or on your phone) and a simple HEX to DEC conversion system.

We all know that Stake Originals games (because we will only talk about them here) are based on a very simple result calculation system. When you start the game, Stake gives you a SERVER SEED that will be part of the score calculation system. You add CLIENT SEED, NONCE to this equation (I'll explain in a moment), choose GAME and SETTINGS, and what Stake does is check the equation and show you the results. Now let's check if it's actually TRUE and if we can actually check it.

Let's deconstruct this result and check it with a regular google calculator.

SERVER SEED – 691922054928c927be2cf70614e59b9d12fec05a8686396f008c8e0c4978b8ca

CLIENT SEED – 111

NONCE – 1 – is the number of bets I have placed on a specific game and setup

GAME – LIMBO ( as an example, but you can do it for each game )

To check the result, we need to know UNHASHED SERVER SEED. You can do this on the Stake website (https://stake.games/provably-fair/server-seed-unhash) after changing the active S / C SEED set.

UNHASHED SERVER SEED - 691922054928c927be2cf70614e59b9d12fec05a8686396f008c8e0c4978b8ca

NOW, CHECK USING THE GOOGLE CALCULATOR

If you go to the result calculation section (https://stake.games/provably-fair/calculation) you can paste all this data.

SERVERSEED:CLIENSEED:NONCE after converting to HEX gives you the result you see in the section CASINO SEEDS TO BYTES. Gra Limbo use first 4 parts after swapping.

So let's calculate it according to the formula given by Stake.

HEX to DEC

87 – 135

f4 - 244

e3 - 227

30 - 48

It's a simple replacement and you can check it by typing in google - HEX to DEC converter.

Now we count the result and check the Stake. We have to do some math just like Stake does every time you place a bet. What are these activities?

135 / (256 ^ 1) =

244 / (256 ^ 2) =

227 / (256 ^ 3) =

048 / (256 ^ 4) =

example:

notice that the result is correct

Now we repeat the operation with the next 4 digits from ours SEED'u.

135 / (256 ^ 1) = 0.527343750000

244 / (256 ^ 2) = 0.003723144531

227 / (256 ^ 3) =0.000013530254

048 / (256 ^ 4) =0.000000011176

We add the results together, which gives us the result 0.531080435961. Then we multiply the result by 100000000 = 53108043.596148490906 and last action : 100000000 / (53108043 + 1) * (1 - 0.01) = 1.8641243876351385 – as stated in the RNG principle for this game, which is on the Stake website.

RESULT FOR THE PLAYER – 1.8641243876351385

RESULT ON PAGE – 1.8641243876351385

IN-GAME RESULT -  x1,86 MULTIPLIER

As we can see, the result is in line with the bet. The same calculation can be made for each game (PLINKO, KENO, SCARAB SPINS, etc).

If we can do it using a regular google calculator. HOW CAN A STAKE CHEAT? THERE IS ONE ANSWER - THERE IS NO SUCH POSSIBILITY !!!

 

Comprehensive research, I’m impressed! 

1) In this example of Plinko we are assuming that I hit the extreme right/left pocket on Medium Risk. Would the ball enter the same pocket on high risk too? 
 

2) If there is no change of seed, Is it right to conclude that any game that you choose to play, will in turn, affect all other results (due to different calculation paths) on that seed?

 

3) How are the results derived for 3rd party providers?

[Josh8018]

Nice explanation!

[mistaker]

On 3/14/2021 at 9:12 PM, nuuuitsjdragon said:

Comprehensive research, I’m impressed! 

1) In this example of Plinko we are assuming that I hit the extreme right/left pocket on Medium Risk. Would the ball enter the same pocket on high risk too? 
 

2) If there is no change of seed, Is it right to conclude that any game that you choose to play, will in turn, affect all other results (due to different calculation paths) on that seed?

 

3) How are the results derived for 3rd party providers?

1. Yes

2. Yes

3. 3rd party providers got different RNG system then Stake, so that's not apply to 3rd party games

[markcariaso04]

that great one hope to see your other post

[Rvrmama]

This is a very good explanation the only part that you don't explain is how they get the server seed client seed  and nonce converted into the hex. 

Also it seems it is written 

Server Seed, client seed:nonce

Vs

Server Seed:client seed:nonce

[notYou]

The description on the beta page does not fit the calculation at the game verification form.

the description on the old page does fit the calulcation shown on game verification form.

However take a look on the calculation part. 

 

the description says: 'Firstly, we take the float and multiply it by both the maximum possible multiplier '. 

so the maximum possible multiplier of what? limbo bet? but thats not 16777216.

16777216 is the max integer of 24bits, right. where do you take this number from? and shouldnt it be 

2**32. but max integer of 32 bit (for 4 bytes) is 4294967296

and the other description uses 1e8 , 100_000_000 

and what has it in common with the max multiplier of 1 million for a limbo roll? 

The numbers doesnt make any sense to me. i wonder if it is possible to hit it in any way

[FabiMa99]

Just found this and the replies.
Whats wondering me is "U just have to have a good seed" or "ur luck is just having a good seed".
So cant stake manipulate giving certain players bad seeds on purpose?
Or have i misunderstood something?

[notYou]

35 minutes ago, FabiMa99 said:

Just found this and the replies.
Whats wondering me is "U just have to have a good seed" or "ur luck is just having a good seed".
So cant stake manipulate giving certain players bad seeds on purpose?
Or have i misunderstood something?

no they cant. they provide their part of the seed (as encrypted hash) first, then you choose your part, the client seed and both, server and client seed as pair get locked on your first bet. this pair is used to calculate your bet results until you reveal the server seed. at this moment you can verify that stake did not manipulate your bet by taking the revelead seed, encrypt it and compare the outcome with the encrypted hash from the beginning (as mentioned at the top)

[LordH0lyBanana]

well you big brain calculated that thanks. Me stoop cant do this  

[dupeddonk]

4 hours ago, notYou said:

The description on the beta page does not fit the calculation at the game verification form.

the description on the old page does fit the calulcation shown on game verification form.

However take a look on the calculation part. 

 

the description says: 'Firstly, we take the float and multiply it by both the maximum possible multiplier '. 

so the maximum possible multiplier of what? limbo bet? but thats not 16777216.

16777216 is the max integer of 24bits, right. where do you take this number from? and shouldnt it be 

2**32. but max integer of 32 bit (for 4 bytes) is 4294967296

and the other description uses 1e8 , 100_000_000 

and what has it in common with the max multiplier of 1 million for a limbo roll? 

The numbers doesnt make any sense to me. i wonder if it is possible to hit it in any way

This has something to do with the fact that 16777216 (2 ** 24)  is the highest number a computer can count to using only 8 bits.  (In binary it's 11111111)   [ignore that, I was wrong]

And because we're dividing a number that's been multiplied by itself, we get the same output that we would've using 1e8 (100000000).

I'm not 100% on this, but it requires more thinking than I'm in the mood for right now, so I'll just be vague for now. 

 

 

 

 

 

[Narchy]

good job will use it later  

[Chrono]

Very interesting read, thanks providing all this research 🙂

[notYou]

7 minutes ago, dupeddonk said:

This has something to do with the fact that 16777216 (2 ** 24)  is the highest number a computer can count to using only 8 bits.  (In binary it's 11111111) 

And because we're dividing a number that's been multiplied by itself, we get the same output that we would've using 1e8 (100000000).

I'm not 100% on this, but it requires more thinking than I'm in the mood for right now, so I'll just be vague for now. 

 

 

 

 

1e8 is just 

let me think for you

2**24 = 16777216 and thats 3 bytes or 24 bits

 

[dupeddonk]

7 minutes ago, notYou said:

let me think for you

2**24 = 16777216 and thats 3 bytes or 24 bits

 

But why do you think they are using 2 ** 24 instead of 1e8 ?

[notYou]

12 minutes ago, dupeddonk said:

But why do you think they are using 2 ** 24 instead of 1e8 ?

see the  last screenshot, the calculation part of the bytes to floats. thats what the pink arrows point to

[autosbd]

Provably fair is there so not only casino cannot manipulate the result but from end user, you cannot manipulate a casino by knowing the outcome beforehand. It works both ways.

[dupeddonk]

On 9/13/2021 at 2:50 AM, mmiikkee1520 said:

I would be willing to bet anyone $1000, that if I livestreamed 20 limbo sessions in a row, that not 1 SINGLE session would I have a 60% win ratio, and over half of them would have a losing ratio of 60%.  if provably fair wasn't rigged, and you take my bet, it should be a guaranteed winner for you.  if not willing to accept the bet, that means you do in fact, think it is rigged.  

Provably fair isn't rigged but it wouldn't make sense to take that bet without knowing what your target multi would be and how many bets you plan on making each session.

 

On 9/12/2021 at 9:55 PM, notYou said:

 

Ok I think I have it figured out.

First of all, I don't think the value of the number they're using for const size matters as long as it's 8 or more digits long.  

As I mentioned before, value of X doesn't change the output in X / (X * Y) * 0.99, at least not when the X has a consistent number of digits.   I believe this is so that the house edge is calculated with enough precision (I believe you would need the precision level to be one less than the max number of digits).

Quote

 

and the other description uses 1e8 , 100_000_000 

and what has it in common with the max multiplier of 1 million for a limbo roll? 

 

Max multi is 1000000.00, 9 digits in total.

I'm thinking the reason for the switch has something to do with migrating from React to Svelte. (And possibly Typescript to JS?, not sure if that's happening though).

@Dan am I close?

[aujlaranjit49]

Wow! You understand all those numbers well.. https://vidm.fun/vidmate  https://authorityapk.com/mobdro 

[dupeddonk]

48 minutes ago, mmiikkee1520 said:

the target is 2.00x.  you want to do the bet?  I bet you don't.  Also in your explanations above, I see alot of "I think", and "I believe", so I guess you must not be 100% sure how it works either.

2.0x wins 49.5 % of the time.  Assuming you make more than 1 or 200 bets you'll pretty much never have a 60% win rate.

I am 100% sure how limbo works.  I would lose this bet, and it's not rigged.

[dupeddonk]

13 hours ago, Ghostnipple said:

I suggest you start again, not asking if stake cheat, ask how can they cheat.

A common misconception is that provably fair means the house can't cheat.  It doesn't. It means that if they cheat, there will be evidence of it.

All of those patterns you insist are real, they're in your head.  This is very common, it's what makes gambling fun (and addictive).

Each bet is independent of the other, meaning the outcome of one bet has no influence on future outcomes.