Blackjack Tracker

[Sonny]

So wouldn't that statement mean, if true, that hand-shuffled shoes are not as random as computer sims wherein each card has an equal chance of appearing anywhere the next shoe?

Not at all. Computer sims show the same degree of clumping. Otherwise, a computer would always have a neutral count and the sims would not show a positive EV for card counting. It would be the same as playing against a CSM. The fact that computer sims still exhibit the same clumping as hand-shuffled shoes is further evidence that computer sims are similar to hand shuffled shoes. It’s certainly not proof, but it is reassuring.

Just because a particular shoe (or many shoes) is “clumped” doesn’t mean that their overall distribution is not random. Just like a BS player will sometimes experience big wins and big losses because of high variance, any given shoe can have high degrees of clumping yet the overall results will still average out to a random (or at least unpredictable) distribution.

-Sonny-

 

[Sonny]

Are you implying that cards couldn't randomly clump together? :devil:

EXACTLY! :1st:

-Sonny-

 

[Kasi]

Are you implying that cards couldn't randomly clump together? :devil:

Lol.

But I guess I do think that hand-shuffles will not as randomly distribute the cards from one shoe to the next compared to a computer shufflle.

That 2 adjacent cards from the first shoe will appear much closer together in the next shoe on average than a computer shuffle which is equally likely to place any card from the prior shoe anywhere in the deck.

What do you think?

 

[Sonny]

That 2 adjacent cards from the first shoe will appear much closer together in the next shoe on average than a computer shuffle which is equally likely to place any card from the prior shoe anywhere in the deck.

That depends on how the computer shuffles, doesn’t it? I mean, if I program Casino Verite to shuffle the same way as my favorite dealer at Mandalay Bay, who’s to say the overall results will be any different?

Then again, who’s to say that the frequency of each card in a hand shuffled game isn't completely random despite the fact that certain cards often end up in the same areas. Even if each deck only got mixed with itself (essentially being 6 single-deck segments) does that mean that the distribution wouldn’t be random?

Okay, now I need another beer…:laugh:

-Sonny-

 

[Kasi]

Not at all. Computer sims show the same degree of clumping. -Sonny-

Well I've seen studies that show the liklihood that the number of cards between 2 adjacent cards from the prior shoe are nowhere near what they would be if they were equally likely to appear anywhere in the next shoe. That the average gap between 2 adjacent cards was so far off from predicted that it was was impossible if the cards had been "randomly" shuffled.

Although maybe these results were for a single deck. Although I don't see much difference. If I think of where I've seen it, I'll let you know.

Although I don't see this, even if true, as changing much of anything, like you say, as far as counting, overall results or anything like that. Or how it would be exploitable because you still wouldn't know exactly when the card might appear. In other words, I think shuffles are "random enough". But no where near as "random" as a computer assuming "random" means any card from the prior shuffle has an equal chance of appearing anywhere in the next shoe.

I guess I'm using perhaps a narrow definition of "random".

Have you ever noticed on the first shuffle with new decks, even after the added wash and shuffle, how many times cards come out together or only a few cards apart since you know how they were arranged prior to the shuffle. Subjective I know, but I can't see a computer shuffling like that.

 

[Sonny]

In other words, I think shuffles are "random enough". But no where near as "random" as a computer assuming "random" means any card from the prior shuffle has an equal chance of appearing anywhere in the next shoe.

Precisely. Hand shuffled shoes are ridiculously nonrandom. When you look at the distributions of specific cards, there is often a 0% chance that they will end up in certain places after the shuffle. By that definition, that shuffle is completely nonrandom. However, the overall probabilities of drawing a particular rank can still be 1/13 despite this fact. It doesn’t matter which ace comes out as long as one comes out every 13 cards (on average). Although the shuffle is completely nonrandom, the frequency of each card can still be normal. As you said, it is "random enough" to fool most gamblers.

-Sonny-

 

[Kasi]

That depends on how the computer shuffles, doesn’t it? -

I'm already having another beer lol.

But of course you're right - I'm sure whatever I think I read assumed a certain shuffle. But even that proves, if you accept the results, that it is at least possible a dealer shuffle is not "random" only in the narrow sense originally adjacent cards end up too close together too often.

Can you make CV shuffle a single deck with the same original starting order with say a riffle-riffle-strip-riffle to see where each card ended up the next time? Would they all end up in exactly the same place the next time because it's a computer? (Of course that wouldn't happen with a human lol).

Anyway, like I tried to say, I don't think the fact that this was happening really changed much of anything overall in any noticeable way. Which was the interesting part to me lol.

Anyway I thought you'd basically agree with this what with all that clumping and sequencing stuff lol.

 

[Kasi]

Hand shuffled shoes are ridiculously nonrandom.

As usual, we end up agreeing - that's all I was trying to say lol.

Yet somehow randomness overall emerges from non-randomly dealt cards!

Time for another one lol.

 

[Sonny]

I'm already having another beer lol.

:toast:

Can you make CV shuffle a single deck with the same original starting order with say a riffle-riffle-strip-riffle to see where each card ended up the next time?

You can adjust the level of “sloppiness” to whatever you like. If you want the dealer to cut the deck at 26 cards “give or take 3 cards” and interlace one card “most of the time” you can do that. I haven't used it for shuffle analysis but that’s what I heard from Norm.

-Sonny-