Randomness of Shuffling Algorithm

Apparently the MatchPlay Casino portion of the 5Dimes Casino utilizes Gametime software for their games. Does anyone know much about how random their 'shuffles' are versus other sites?

Also, the site says that for the 6 deck computer (not live) blackjack game, the cards are shuffled after a random #of rounds. But my question is: Are those exact cards that you were just betting on really "shuffled" electonically, or is a brand new set of 312 purportedly randomly shuffled cards 'brought in?" In other words, is it possible the new shoe of shuffled cards bears any resemblance to its former shoe, depending on the thoroughness of the shuffle?

The algorithm for an "electronic" shuffle resembles a perfect shuffle. It doesn't matter if you take an old shoe and shuffle, or a new (i.e. read as ordered) deck and shuffle.
Provided that the RNG doesn't have a weakness, a shuffled deck is as independent tot he previous order as it can get.

Wouldn't it depend on what algorithm they are using for their shuffle? I know some algorithms are better than others.

Quote: konceptum

Wouldn't it depend on what algorithm they are using for their shuffle? I know some algorithms are better than others.

And some are broken -- that is, mathematically unsound. There's really no way to know without looking at the code.

Quote: konceptum

Wouldn't it depend on what algorithm they are using for their shuffle? I know some algorithms are better than others.

If you are nitpicking, you are right. No algorithm could produce a shuffle, as all algorithms need to be deterministic (they give same output for same input).
Technically, you can use an algorithm, provided that part of the input is indeterministic (the RNG).

A simple algorihm would be, assign a random value (doesn't need to be a specific distribution) to each card, and then in the next step order the cards according to the given random value.

A "better" algorithm would be: Pick a random card from all cards available, as first position. As second position, pick a random card from all cards available that are not picked before. Proceed till the end.

Both algorithms do the exact same thing, they shuffle a deck of cards. The shuffles itself are indistinctive. However the last algorithm is "better" because it performs faster, as there is no sorting involved.

The only handle to crack a shuffle algorithm is to crack the RNG. However nowadays (yes, casino do learn from mistakes) it is nearly impossible. Even if you would be the manufacture of the RNG device, there are frequent audits of the randomness of the RNG.

Granted, the RNG has to be truly random, or at least a very close approximation of it, in order for any algorithm to work. So let's take as a given that they are using some sort of truly "random" RNG.

Even considering that, there are some algorithms that are better than others. I don't have anything in front of me, but I seem to recall that the people who cracked the Planet Poker card room not only utilized the fact that their RNG wasn't truly random, but also a flaw in the coding that didn't result in a truly shuffled deck.

My understanding is that for a shuffling algorithm to be completely accurate, it must be able to produce every single permutation of the deck that is possible, and with an equal probability of each of those permutations occuring.

Even with all that, I can see how people will choose to believe that casinos are not truly using randomized shuffled decks. After all, the casino must explicitly display their code in order to "prove" their shuffling algorithm is random and accurate. Display said code might not be in the casinos best interest, as it may give away other secrets of how the casino is operating. However, even if the code is displayed, there's no true guarantee that the code is what is actually being used.

The next step is to have an independent audit or some expert confirm that the code is being used, and that the code is random and accurate. However, no matter who is chosen, there will always be people who will decide that the independent expert is really in cahoots with the online casino and is just saying what they want them to say.

The Fisher-Yates shuffle does the trick, does it in place and in a minimum amount of time as well. There's a common error in it's implementation (common enough that's it's been written about) that leads to a non-random shuffle, but you'd hope such things aren't out in the wild any more.

Course it relies on a random number generator that is actually random. That's a different story.

Quote:

"A "better" algorithm would be: Pick a random card from all cards available, as first position. As second position, pick a random card from all cards available that are not picked before. Proceed till the end.

Both algorithms do the exact same thing, they shuffle a deck of cards. The shuffles itself are indistinctive. However the last algorithm is "better" because it performs faster, as there is no sorting involved."

This is the way i shuffled 21 back in the late 70's/early 80's. As long as the PRNG is sound, its a better way. HOWEVER, no shuffling once the cards are being dealt. IMHO the deck speaks. No shuffling in between dealt cards. This is being done on-line, and "justifies" pulling a rabbit out of a hat. I have also noted early-on in on-line gaming that certain Poker sites would guarantee a fair pocket in Texas Hold'em... and not a word about the board cards. Sometimes whats not said speaks much louder than whats said.

N&B

Quote: MathExtremist

And some are broken -- that is, mathematically unsound. There's really no way to know without looking at the code.

Good reason to avoid on-line gaming... you CAN'T see the algo.

What a great industry, where most just regulate themselves.