Improper shuffling - how to exploit?

[GuitarMarc]

The famous card shuffling study said you need 7 perfect rifle shuffles to properly randomize a deck. Most brick and mortars apparently do not do this.

Example - I was at the Oaks Club this weekend and in the back room they do not use automatic shufflers. The dealer takes the muck and unused cards, spreads them out and mixes them a bit before putting the deck back together. Then 3 rifle shuffles followed by a quick hand mix then 1 more rifle shuffle.

Should this produce any kind of recognizable pattern from the previous hand over the next few hands? Like the flop may be similar or the hole card combinations could be similar? If so, what to do about it?

[callydrias]

[ QUOTE ]
The famous card shuffling study said you need 7 perfect rifle shuffles to properly randomize a deck.

[/ QUOTE ]

I don't think any number of "perfect" riffle shuffles is going to be effective. Anyone who shuffles their chips folding the night away in a B&M knows this. I assume 7-riffle rule means that the cut should be just near the center, not always exactly at the center.

[ QUOTE ]
Should this produce any kind of recognizable pattern from the previous hand over the next few hands? Like the flop may be similar or the hole card combinations could be similar? If so, what to do about it?

[/ QUOTE ]

I think the only thing you'd have a snowball's chance in hell of determining is what region of the deck a certain card might end up. Next time you muck an A (or whatever card you're interesting in following), try to watch it. You'll probably lose it in the wash or box, but if you still think you've followed it, the best you can really do is determine if it's in play or not. Most of the time it won't matter because you mucked preflop for other reasons and the rest of the time you'll be wrong.

Wanna play some 3 card monte?

[StellarWind]

I believe the shuffling study originated with bridge.

Bridge players sort their 13-card hands by suit and rank. The trick-playing action at bridge also imprints certain patterns in the card order such as several cards in a row of the same suit and aces immediately following kings.

These patterns need to be shuffled out or future bridge deals may have biased distributions.

Many duplicate bridge players claim computer dealt (using RNG) hands have wild distributions. Actually the computer is fine but players are used to inadequate shuffling which favors abnormally bland distributions.

Non-drawing poker games don't normally cause any appreciable sorting of the cards. Inadequately shuffling won't change the odds of getting any particular hand. E.g. the probability of getting a monotone flop will be normal.

However two cards that were together on the previous deal may stay fairly close together. If the previous flop contained red aces and you now hold pocket A [img]/images/graemlins/diamond.gif[/img], there may be an increased chance that A [img]/images/graemlins/heart.gif[/img] is not far away--either in another hand or coming to the board.

What are the poker ethics of exploiting poor dealer shuffling?

[Gandor]

[ QUOTE ]


Example - I was at the Oaks Club this weekend and in the back room they do not use automatic shufflers. The dealer takes the muck and unused cards, spreads them out and mixes them a bit before putting the deck back together. Then 3 rifle shuffles followed by a quick hand mix then 1 more rifle shuffle.

[/ QUOTE ]

It is my understanding this is the proper method to use.

[Pov]

I don't think you can be faulted for using information you obtained legally. It is available to everyone. If you can follow the Aces through the shuffle then you would be silly not to. You would also be silly not to assume at least several of the other players would be doing the same thing which is of course what makes it highly unlikely it can be effectively done.

[callydrias]

[ QUOTE ]
[ QUOTE ]
Example - I was at the Oaks Club this weekend and in the back room they do not use automatic shufflers. The dealer takes the muck and unused cards, spreads them out and mixes them a bit before putting the deck back together. Then 3 rifle shuffles followed by a quick hand mix then 1 more rifle shuffle.

[/ QUOTE ]

It is my understanding this is the proper method to use.

[/ QUOTE ]

Actually, I think the minimum is riffle, riffle, box, riffle. Or alternately riffle, box, riffle, riffle. You just can't start or end with a box. Too lazy to look it up in Robert's Rules...

[Guthrie]

Back in the seventies I read Thorpe's book on blackjack and decided to test his theories by writing a blackjack simulator. I had access to an IBM 1130 so I taught myself FORTRAN and set to work. Writing a random number generator was beyond my skill level, and I figured that no human really shuffled randomly anyway, so I decided to write a subroutine that actually shuffled the deck. I figured a good dealer would cut the deck exactly in half and drop one card off each side, then cut exactly in half and do it again, and so on. I started with the cards ordered like a new deck fresh from the factory, and had a single input, how many times to shuffle. I told it to shuffle eight times, as I recall, and nothing happened. The deck was completely unchanged. I spent days trying to debug the thing, then picked up a new deck of cards and very carefully recreated the shuffle one card at a time. As you have probably guessed by now, the subroutine was working perfectly, but after eight times (or whatever the number was), the deck was back in its original order. There were also some very interesting combinations along the way.

[GuitarMarc]

Right. That famous study said 8 perfect rifle shuffles will return the deck to its original state but 7 creates good randomness from the original.

[DrPublo]

Here's a link to a site that I thought would never get mention on here.

Riffle Shuffles

The Doc

[juanez]

I never read the "famous cars shuffling study", but according to " The Professional Poker Dealer's Handbook " (Paymer, Harris, Malmuth) the proper sequence in riffle-box-riffle-riffle.

Our house shuffle is riffle-riffle-box-riffle. I assume this produces similarly random results.