jackhigh78
Joined: Aug 30, 2013
• Posts: 4
September 4th, 2013 at 5:13:01 PM permalink
it depends on the game ,algorithm and the seed.

For example if you use 32,64 or 128 bit RNG (high quality algorithm, nothing wrong with distrubition) to to deal the cards for texas hold'em game, it is still rigged, not fair enough because you need at least 226 bit to create all possible shuffles for a 52 cards.
MathExtremist
Joined: Aug 31, 2010
• Posts: 6526
September 4th, 2013 at 7:18:42 PM permalink
Quote: jackhigh78

it depends on the game ,algorithm and the seed.

For example if you use 32,64 or 128 bit RNG (high quality algorithm, nothing wrong with distrubition) to to deal the cards for texas hold'em game, it is still rigged, not fair enough because you need at least 226 bit to create all possible shuffles for a 52 cards.

Only if you're doing a single RNG pull and mapping that data to a deck ordering. That's not the way most shuffles work, certainly not the Fisher-Yates shuffle. Under the assumption of an unbiased RNG, all you need is 6 bits (and multiple pulls) to shuffle the deck. You could certainly pull 226 random bits from an unbiased bit stream and then do that mapping using a factorial scheme, but I don't know if anyone's actually doing that. It would involve a lot more cycles than the F-Y shuffle.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
binary128
Joined: Nov 5, 2011
• Posts: 82
September 4th, 2013 at 10:05:21 PM permalink
Quote: 4ofaKind

Can you explain how the alleged random RNG's were programed to operate when anyone of the many online rigged software's were exposed?

Let's say that, in code, you had a "thing" (an organized pile of programming) called a RouletteGame. When you asked it to, this RouletteGame would do something called PlayGame. To do this PlayGame, the RouletteGame would need a list of the players bets (a combination of bet types and bet amounts).

GameResults = RouletteGame.PlayGame(PlayerBets)

BUT ....

What if the RouletteGame ALSO could make use of a little bit of information that I'll call WinFactor.

GameResults = RouletteGame.PlayGame(PlayerBets, WinFactor)

The WinFactor would be used first. It determines whether this spin of the Roulette wheel is going to be a winner or a loser. Here's how that would work.

WinOrLose = GetRandomNumber(1, 100)

If WinOrLose is less than WinFactor, then this game is going to be a loser (IsGameLoser = True). For example, if WinFactor is 20 and the WinOrLose random number is 17 then all of the game processing that follows will result in a losing game.

Let's see how that would work.

If IsGameLoser Then
- Do
- - - WheelResult = RouletteWheel.SpinWheel (This is the part that uses the RNG)
- - - BetResult = RouletteBets.GetTotalResult(WheelResult)
- Loop While BetResult > 0
End If

NOTE - the RNG is inside the Do loop. The RNG is doing exactly what I described - "... the ONLY thing that an RNG does - when something asks it for a number within a range, then it makes sure that any number in that range has exactly the same chance of being returned as any other number".

The game is corrupted by the Loop While statement. What this statement means is to just keep running the game over and over until you get a loser.

That was the flaw in the Legends' statements that they were using a "high quality RNG". The game corruption had nothing to do with the RNG - it had to do with that "Loop While" statement.

Chris
RaleighCraps
Joined: Feb 20, 2010
• Posts: 2501
September 5th, 2013 at 5:28:59 AM permalink
Great explanation Chris.

I will never trust any electronic gambling for this very reason. It is too easy to manipulate the supposed randomness with code.
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
Sonny44
Joined: May 13, 2013
• Posts: 217
September 5th, 2013 at 7:35:51 AM permalink
Quote: RaleighCraps

Great explanation Chris.

I will never trust any electronic gambling for this very reason. It is too easy to manipulate the supposed randomness with code.

Isn't that a pretty broad statement to make? I'd like to see other opinions on this. I'm no programmer, but I do get the Loop While function & the RNG's place w/in it. But, couldn't authorizing authorities, somehow, make sure there's no Loop While function in the code? Or, even if they did, the code could still be altered afterward to "fix" the machines. Is that possible?

IOW, the only reliable randomness is in live table games? Everything else is "fixed"? I haven't put real money into an online craps game & that's what I'm trying to decide & why I started this thread. Maybe forget that scene?
MathExtremist
Joined: Aug 31, 2010
• Posts: 6526
September 5th, 2013 at 7:50:38 AM permalink
Quote: Sonny44

Isn't that a pretty broad statement to make? I'd like to see other opinions on this. I'm no programmer, but I do get the Loop While function & the RNG's place w/in it. But, couldn't authorizing authorities, somehow, make sure there's no Loop While function in the code? Or, even if they did, the code could still be altered afterward to "fix" the machines. Is that possible?

Not only do regulators and test labs in well-run jurisdictions get source code for everything, they do testing at all levels of the process: the RNG itself, the scaling algorithm that converts raw RNG output to game-specific output (like roulette numbers), and the final results distribution. That "loop until the player loses" nonsense would never get past the NGCB or GLI, and it's absurd that anyone would ever claim that it's a proper use of an RNG. RNGs have been used (properly) in electronic single- and multi-player games since the 1980s. These days it's more likely to find a meaningful bias in a roulette wheel than a software RNG.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
MangoJ
Joined: Mar 12, 2011
• Posts: 905
September 5th, 2013 at 8:24:19 AM permalink
Quote: MathExtremist

Only if you're doing a single RNG pull and mapping that data to a deck ordering. That's not the way most shuffles work, certainly not the Fisher-Yates shuffle. Under the assumption of an unbiased RNG, all you need is 6 bits (and multiple pulls) to shuffle the deck.

What's important is not the number of bits the RNG will return as a random number, but the number of bits the (pseudo-)RNG keeps as its internal state.
RaleighCraps
Joined: Feb 20, 2010
• Posts: 2501
September 5th, 2013 at 9:02:41 AM permalink
Quote: MathExtremist

Not only do regulators and test labs in well-run jurisdictions get source code for everything, they do testing at all levels of the process: the RNG itself, the scaling algorithm that converts raw RNG output to game-specific output (like roulette numbers), and the final results distribution. That "loop until the player loses" nonsense would never get past the NGCB or GLI, and it's absurd that anyone would ever claim that it's a proper use of an RNG. RNGs have been used (properly) in electronic single- and multi-player games since the 1980s. These days it's more likely to find a meaningful bias in a roulette wheel than a software RNG.

"regulators and test labs in well-run jurisdictions get source code for everything, they do testing at all levels of the process"

And therein lies my reluctance. Any online craps game that exists in an online casino that I am aware of today, is NOT being run in any jurisdiction that I would dare bet is 100% looking out for the players. You can trust your money to the nations who allow casino internet craps games, but I don't.

Now, once internet casinos come online in the US, then I would tend to agree that the adherence to legitimate randomness will be assured. But until then it is my opinion that you are just gambling that the game is going to be completely fair.

I would think that ME's statement above is applicable to electronic gaming in the US. Other countries may indicate their games pass GLI, but who is the enforcement agency that is guaranteeing compliance?

I do find it interesting that the electronic craps games released in casinos in the US today all use real dice to determine the roll outcome. I am not aware of any game that rolls the dice via a RNG.
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
charliepatrick
Joined: Jun 17, 2011
• Posts: 2487
September 5th, 2013 at 9:05:43 AM permalink
Quote: MangoJ

...the number of bits the (pseudo-)RNG keeps as its internal state.

I was initially confused at first, but you are correct.

Assuming you're shuffling one pack of cards, you need to ensure that every possible [re]ordering of the cards can come out, presumably with equal probability. Simply put you have to pick a number from 1-52 for the first card, 1-51 for the second etc.

(i) the output of each "number" from the RNG has to be big enough to create 52 different answers (2^6 would do, but for practical purposes 2^32 is easily sufficient).
(ii) the output of 51 consecutive "numbers" has to cover all the permutations (internal state).
(iii) the algorithm to use those numbers to create the shuffle has to be fair and equal (programming outside the RNG generator).

Since each starting place (internal state) will always produce the same sequence of random numbers, rule (ii) means there must be more starting places than there are possible orderings for a pack of cards. This is where the 226 number comes from (2^226 and 52!).

However what is also probably worrying for a casino only using 2^226, assuming you know the internal logic, is that the order of a single pack of cards identifies the starting position and presumably the next pack dealt. That is why they probably would keep picking random numbers in real time until the next one is needed.
4ofaKind
Joined: Sep 28, 2010
• Posts: 221
September 5th, 2013 at 1:32:02 PM permalink
Quote: MathExtremist

Not only do regulators and test labs in well-run jurisdictions get source code for everything, they do testing at all levels of the process: the RNG itself, the scaling algorithm that converts raw RNG output to game-specific output (like roulette numbers), and the final results distribution. That "loop until the player loses" nonsense would never get past the NGCB or GLI, and it's absurd that anyone would ever claim that it's a proper use of an RNG. RNGs have been used (properly) in electronic single- and multi-player games since the 1980s. These days it's more likely to find a meaningful bias in a roulette wheel than a software RNG.

Nevada and New Jersey are the only two jurisdictions I know for fact that require this type of testing. Not that there aren't any other land based jurisdictions requiring the same scrutiny.

When it comes to computer games I wouldn't even trust land based Indian casinos since most of their back end operations are secret and most information is unavailable to the public.

Now when it comes to online gaming is when the fn ridiculous comes into play. If anyone out there thinks they know for sure that their gambling online and confident their getting a fair game can only be considered a gambling fool.

Other then the obviously exposed proof of the poker scandals which made its way to the main stream media, name me one case when rigged casino games were exposed and any online regulator taking any type of serious action.

Online regulators can't even enforce their own basic rules when handing out licenses. I'm willing to bet that you could stomp them just asking about an RNG, source code, scaling algorithm, etc.

At quality land based jurisdictions not only is the software lab tested, it's then sealed with a tag and could only be installed and removed with a regulator present.

This is not saying there aren't any online casinos offering a fair game. But if you have to count on the he said, she said, he said - it's fair, and when you run into a problem the only chance of at least someone listening to you, even if they are Kangaroo Court's are places like "Casinomeister"............ gambling fool is an understatement.

@Chris...Thanks for the explanation, one of the very few that would be willing to step to the plate.