PDA

View Full Version : Card Patterns in Online Poker

Collin O'Mahon
09-30-2003, 11:35 AM
I've been playing Holdem extensively at PartyPoker for several months. There have been numerous occasions wherein I have been dealt the same hole cards twice in a row.

It would seem that the probability of being dealt the same hole cards twice in a row would be less in a b&amp;m room, as the cards are physically shuffled. Physically, although the shuffling is itself a random process (if done correctly), it would seem that a proper shuffle would tend to distribute the cards differently than any previous, or subsequent, shuffle. However, the random number generators used by the online sites don't take this into account. Is this an accurate protrayal of b&amp;m reality?

Copernicus
09-30-2003, 02:21 PM
If you are saying that a "random" physical shuffle would tend to separate hole cards, whereas a generated distribution would not know their prior position, I disagree (on the physical side). To be dealt the same cards twice in a row they are in fact separated between hands...they have become n apart whereas they were together in the muck.

A random physical shuffle will initially separate the cards, but they still have the possibility of returning to n apart (where n is the number of players) and dealt to the same hand, or they could return any other number apart and not appear in the same hand.

I havent noticed any pattern of hole cards repeating online. The only "aberration" that I noticed early on in my online days was an apparent tendency for board pairs to repeat, but I doubt that the occurrences were anything other than normal.

Collin O'Mahon
09-30-2003, 02:35 PM
Thanks. Good explanation, very clear. I tend to agree now that "aberrations" seen online are simply a function of (a) more hands played online (many more hands seen), and (b) not much time spent in b&amp;m cardrooms.

PAUL-IN
09-30-2003, 02:39 PM
in an online cardroom, you will generally see many more hands per hour than in a b+m cardroom. because of the larger amount of hands, "strange" things occurr, such as same flops, one after the other, or same holecards. but the probablility of being dealt the same hole cards twice in a row should statistically be equal to that in a b+m cardroom, unless the shuffling algortithm PartyPoker uses is bugged.

Collin O'Mahon
09-30-2003, 02:53 PM
Nothing is truly random. The question, really, is how closely one approximates, or simulates, randomness. The physical shuffle and the RNG algorithm used online are both simulations of randomness. However, it is possible that the accuracy of the simulations may differ. It may not be apparent, even using statistical measurements (i.e., the process is "mostly" random, in that certain patterns are not appearing more or less frequently on average then statistical predictions would suggest).

However, I'm wondering if the physical shuffling process in a b&amp;m cardroom is a less "random" simulation than a software algorithm, generally speaking. For example, there are many variables in the physical suffling mechanism that will affect randomness (for lack of a better abstract term). Length of shuffle, number of cuts, dealer exertion, friction between the cards (themselves and the table), quality of table covering, etc. etc. etc. These variables are not included in the software algorithm, I would expect, and thus the relative "randomness" will differ.

Of course, if the distribution of hole cards (or complete hands) in a b&amp;m cardroom is reasonably similar to what may be statistically predicted, then it is probably safe to say that the shuffling process is "random" for our purposes.

It may even be argued that the more "random" variables are included in the process, in the case of a b&amp;m cardroom, the more the result of the process is "random." And thus perhaps b&amp;m shuffling is a better simulation of randomness.

Copernicus
09-30-2003, 04:16 PM
I mentioned this in another thread somewhere recently, but in statistical tests of the randomness of physical shuffles when only "riffle shuffles" (or whatever the proper name is) are used there is an optimum number of shuffles, any more or less and the distributions become less random. I believe the number was 5 or 7.

I also believe that the same studies resulted in the procedure that most if not all casinos require for poker, single deck and two deck blackjack, which is that there be a "wash" before the riffle shuffles. The wash was found to result in better randomness.

DS might have done some consulting in this area?

Collin O'Mahon
09-30-2003, 04:21 PM
When you say "wash" you mean when the dealer spreads the cards out on the table and *cough* randomly moves them around?

Copernicus
09-30-2003, 04:32 PM
Yes...I think "wash" is the proper term for that, but I'm sure there are dealers here who can correct that if I'm wrong.

Wake up CALL
09-30-2003, 04:34 PM
[ QUOTE ]
I mentioned this in another thread somewhere recently, but in statistical tests of the randomness of physical shuffles when only "riffle shuffles" (or whatever the proper name is) are used there is an optimum number of shuffles, any more or less and the distributions become less random. I believe the number was 5 or 7.

I also believe that the same studies resulted in the procedure that most if not all casinos require for poker, single deck and two deck blackjack, which is that there be a "wash" before the riffle shuffles. The wash was found to result in better randomness.

DS might have done some consulting in this area?

[/ QUOTE ]

Copernicus, here are two good links for you about random card shuffling.

How Many Times to Shuffle (http://www.nature.com/nsu/001005/001005-8.html)

Perci Diaconis reference in Wikpedia (http://www.wikipedia.org/wiki/Shuffling_playing_cards)

Cyrus
10-01-2003, 11:16 AM
My impression is that poker players, as opposed to blackjack players, in general, are not familiar with shuffle routines and they are not interested in their secrets, either.

Copernicus
10-01-2003, 11:29 AM
I know of some poker players who try to follow mucked cards through the shuffle at least to get a general idea, eg, whether those Aces might wind up toward the top of the deck.

Carl_William
10-01-2003, 11:44 AM
'Yes...I think "wash" is the proper term for that, but I'm sure there are dealers here who can correct that if I'm wrong.'

In LA County CA: a wash is always called a "scramble." Has been for the last 20 years.

squiffy
10-01-2003, 01:02 PM
I am posting this before reading any of the other responses. Personally, I have been noticing this myself. It could just be paranoia. But it does seem to me that I get the same cards twice in a row more often on the net. And it seems more frequent that the last winning hand often appears as the next flop.

Most likely it's just coincidence. There is a big difference between "feeling" there is a pattern and actually documenting it with hard evidence.

Carl_William
10-01-2003, 01:32 PM
Random comments on computer program random number generators:

In the computer programmer world, so-called generated random numbers are commonly called pseudo random numbers (I think we can all understand why -- nothing is perfect). Pseudo random number generators generally might have a cycle duration which at some point they start over again and repeat the same numbers. There are also quasirandom number generators but these are probably not currently in use.

At this point in time, there are superb “high quality” computer pseudo random number generators available which have very long cycle durations and produce excellent random numbers. Various tests are applied to these generators to assure that there are essentially no holes or voids in the system greater than the expected statistical deviations.

Initiation seed:

A seed number is usually used to start a chain of random numbers. It could be the time of day and date. If this seed is recorded – then the chain can be replicated for a particular situation. For instance, every deal at online poker sites has a deal number. The last seven digits for a particular deal could then be used as a seed for the next deal. This suggests that the deal sequence of cards could be repeated every ten million hands – who knows. There are numerous techniques for adding additional degrees-of-freedom into a pseudo random number generator program which will then produce even more random output.

Online Poker Management:

Poker management always tries to make higher profit. Hopefully management is honest. If the site has honest management and business is good – then players probably don’t have to be over concerned about the random number generators. Competition is very tough in the Online Poker World. When push becomes shoving, and sites can’t pay their bills and salaries, then there is a good chance that the players will lose their account money in those sites. In the past, this happened at some B&amp;M poker clubs in California, at some point during the day, the chips became immediately worthless – the cage closed and lots of chip holders were out of luck.

Just look at the business decline at some online poker sites, and current success of the online poker site leader(s). Will there be more shakeouts? Probably. Also it is a piece of cake to cheat on the Internet. Either we are honest or we are not honest – it’s that simple. There is probably no penalty for cheating other then temporary barring the cheaters. I don’t think there is any legal way to administer justice to the cheaters.

In the past, Amway Products used a pyramid system to reward it salespersons. Maybe they still do – I will never buy Amway products because of that. From what I have read on Two-Plus-Two, I think some of the poker sites are using techniques similar to Amway to lure customers. It will be interesting to see what happens to these and other Internet Pokers Sites in the future.

Carl_William
10-01-2003, 01:44 PM
Poster: squiffy
Subject: Re: Card Patterns in Online Poker
" Most likely it's just coincidence. There is a big difference between "feeling" there is a pattern and actually documenting it with hard evidence."
...

Squiffy,

Maybe you can print them out if you have a printer connected to your computer and try to check things more closely. At one time (maybe still), Paradise Poker had an excellent way to print the history for a poker deals. PartyPoker also lets the player print hand histories, but they don't think they give as much information as Paradise did.

M.B.E.
10-06-2003, 05:30 AM
Thanks, Wake Up Call, those are good links.

M.B.E.
10-06-2003, 05:36 AM
[ QUOTE ]
in statistical tests of the randomness of physical shuffles when only "riffle shuffles" (or whatever the proper name is) are used there is an optimum number of shuffles, any more or less and the distributions become less random.

[/ QUOTE ]
If you shuffle more than the optimum number, the deck won't become less random. (Isn't that the law of entropy?)