Online Poker Forum - Card Shuffling and Algorithms

Skymiler75 · High Card Joined: 29 Jul 2009 Posts: 7

First let me state, I do not believe that the poker games on FTP poker are rigged. It doesn't make any sense to rig a business that is making a ton of money. Chasing your customers away?, please.

What is of concern/curiosity to me is the nature in which the cards are shuffled. Within this concern there are three basic questions:

1. the timing of the shuffle, and
2. the influence table play has on the cards that are dealt
3. does FTP have different shuffling algorithms for different games

As for the first question, it is fairly straight forward. Is the entire deck (all 52 cards) shuffled before the button is moved and blinds are posted? Or in other words, can the remaining cards (in the deck) be reshuffled after the flop, turn or river? This last question is consistlently asked by short-stacked all-inners when they see a two percent big stack get runner runner near the end of a tournament and bust them out. The implication is that the tournament is "geared" to finish on a timely basis.

The second question is associative of the first. Namely, are aggressive players rewarded more than passive players? Are big stack players rewarded more than small stack players? Does table position influence the cards you are dealt? Putting poker theory aside, each of these questions is a question of the integrity of the sites algorithm(s). Personally, I play a passive trapping style of poker, which does not work well on FTP. Rather, when I am up against an aggressive play more often than not he/she is rewarded with his/her river card.

The third question should be self-explanatory. I like play all the games and would like to know if razz, stud, and omaha use different shuffling models.

Thank you.

CoAz Kid · Full House Joined: 06 May 2009 Posts: 164

To 1, my understanding is that the cards are constantly shuffled pre-post flop, post turn etc.

To 2, No. No. No. As far as trapping goes, anytime you try to trap you risk being drawn out on, especially so with 2 or more people in the hand. Ive personally come to see trapping as..myself deciding that I have the best hand at the flop, and ignoring the real possibility that the best hand that can be made could change dramatically with the addition of 1-2 more cards on the table, and then trying to draw maximum value from the other guys chip stack.
This has never paid off in the long run for me, and is usually how I lose in a tourney (bad beats being the other way). I dont trap much anymore.

3, Dont know, but would assume the same algorithm is used.

IABoomer · Moderator Joined: 29 Oct 2006 Posts: 10709

1. Cards are for all intents and purposed constantly shuffled. See the link in my signature below for a full description of how the shuffle works.

2. Stack sizes, relative position, and anything else is irrelevant to which card appears next.

3. Same system is used regardless of the game being dealt. Why would you complicate things and use a different method for one game?

Skymiler75 · High Card Joined: 29 Jul 2009 Posts: 7

CoAz / IAboomer,

Both of your answers to my questions are similar, so I will address them as if they were the same.

Given that the deck is in constant shuffle (as opposed to a live table game), then it is âpossibleâ that other factors (beyond the cards) play a role in which cards are dealt. I do not have access to the FTP algorithm(s), so I can not definitively say that there are not any components within the shuffling model(s) that address table position, stack size, blind level in a tournament, etc. If either one of you has access, please let me know, otherwise you are just speculating as to the (non) affects of these other factors.

To restate your remark (in part three) IAboomer, âwhy would it be necessary to constantly shuffle the cards (during a hand) other than to make certain specific cards come out in various situationsâ? I have read Perry Friedmanâs explanation and Howard Ledererâs remarks regarding the use of a constant shuffle to deter cheats, but neither one of their comments satisfactorily addresses my concern about extraneous influences over the dealt cards. As I have stated and CoAz has reiterated, trapping (a basic poker strategy) does not work online nearly as well as it works at a live table game. The most obvious explanation for this is a shuffling bias based upon the factors described above.

On a side note, Mr. Friedmanâs comments on rec.gambling.poker suggests that online shufflers are NOT programmed to BURN a card after the flop or turn, he states:

Now, when it comes time to deal the flop, the table asks for 3 cards
from a 34 card deck. The table knows what cards are left in it's own
deck. All the RNG returns is 3 indices between 0 and 33, which are
used to pick from the remaining cards in the table's deck.
Again, the process of choosing these new indices causes the 34 card deck
(or more correct, the set of indices from 0 to 33) to be reordered.
Is this true on all online poker sites or just Full Tilt Poker?

Once again, I appreciate your feedback.

IABoomer · Moderator Joined: 29 Oct 2006 Posts: 10709

The reason Full Tilt runs a constant shuffle is for added security. If (and it's a big if) a hacker were to violate the server security and had access to the game control server, they could see what hand was coming and play accordingly. In a live game, the dealer protects the deck and people at the table can see if it's being violated.

The purpose of a burn card live is to keep someone from reading a mark on the top card and knowing what will be coming. Since you can't see the deck, and can't physically mark the cards online, there's no reason to burn a card.

If you're going to get hung up on the fact that your acting later might cause a different card to fall, keep in mind that you can't know (nor can any player) if that different card will help or hurt your hand. Therefore, the game remains random and unpredictable in the same way as playing live with actual cards.

Skymiler75 · High Card Joined: 29 Jul 2009 Posts: 7

IABoomer

I get all that. The technology is wonderful. But it doesnât change the fact that it is a different game than live poker. I will ask once again, do you have access to or have you ever seen the algorithm FTP uses? If not, then what are we taking about? Since the probability of someone hacking the sites server(s) is so low (big if) then why not use a static shuffle algorithm?

Dan Harrington once stated â¦ âpoker is a partial information gameââ¦, the skill comes from reading your opponents face, knowing your cards (and their intrinsic value relative to the board), understanding statistics and probability theory, money (chip) management, etc. (ie â does the statistical probability of a flush coming out change when a 34 card deck is shuffled, as opposed to a 52 card deck or 33 card deck? The answer is yes, is has to.)

Without a complete understanding of the method in which the cards are shuffled, or how the shuffle is affected by the various factors I had listed before, then the amount of information we as poker players have available to us is reduced online. Thus, reducing the game to one of LUCK. How else do you explain the relatively poor performance (in tournaments) by the pros on this and other sites?

At some level, all of the factors mentioned previously are different âonlineâ than at the brick and mortar casino. Indeed, not just different but constantly changing. If this is true (not a big if), then the (online) game is more based on LUCK than SKILL. And consequently, the argument we are trying to make to legalize online poker is illegitimate. In your third paragraph you used the word âunpredictableâ, well let me tell you predictability (with a certain level of probability) is what great poker is all about.

Next, forget the purpose of the burn card (at a live game): is burning a card a part of the Rules of (Hoyle) Poker, is the statistical outcome of âpatâ hands changed as a result of the exclusion of burn cards? The answer of course is yes. Why not deal a sixth card to the community board? Everybody can use it, no reason not to. BECAUSE IT CHANGES THE GAME!!!

IABoomer · Moderator Joined: 29 Oct 2006 Posts: 10709

A burn card is an unknown card to you, just like the stub, the muck, and the opponent's cards. Every single calculation of probability deals with outs/unknown cards. Thus, not burning a card doesn't change the odds at all. If you had a live dealer who could guarantee not to flash a card, and if you had some special indestructible cards that wouldn't wear from the process of shuffling, you could have a dealer shuffle the deck during a hand live and it wouldn't affect the odds of you making your flush by the river (or whatever example you want to use).

The methods between live poker and online poker are different because they're different mediums. The mathematical properties of the game, however, remain exactly the same regardless of the method used to distribute the cards.

The online game is dollar for dollar much more difficult than a live game. Between the number of hands dealt, the ability to multi-table, the vast amount of information/coaching available, statistical tracking software, etc., online players can learn in a year online what might have taken the old-school professional players 10 years or more to learn playing live. Add in the fact that online you can't read any physical or verbal tells on an opponent and a source of information goes away.

Skymiler75 · High Card Joined: 29 Jul 2009 Posts: 7

IABoomer,

Your last entry was fascinating to read. If I understand what you said, then online poker is tantamount to playing BINGO. But tell me then, why then is it necessary for the online dealer to deal in the traditional orbiting fashion? Why not deal the small blind two cards, then the big blind two cards, etc. etc.? Or is this in fact the way the cards are dealt? I would appreciate an answer to this question.

The fact that the burn card is an unknown quantity is immaterial to the absolute fact that burning card(s) reduces the number of cards available for the flop, turn and river â and thus changes (however slight) the complexity of the game. To not acknowledge this is completely disingenuous. Question, would your strategy (in a given hand) change if the burn card process WAS used in the deal? The answer, of course, is yes.

Not applying the burn card process is analogous to the Designated Hitter Rule in baseball. Is the game of baseball different in the American League than in the National? Is there a virtual (pun intended) elimination of a basic strategy with the Designated Hitter Rule? Yes and yes.

Assume, for a moment, that a player (call him player A) is dealt two hearts and the flop (community board) reveals two hearts, thus player A is drawing to a heart flush. Further assume that player A is heads-up against another player with pocket aces (call him player B). Any decent player in player Bâs position would CONSIDER player A on a flush draw and act accordingly. Player Bâs analysis at both a virtual and brick & mortar table will include the question of how many (statistically) available hearts there are in the deck. The answer, of course, is different depending upon whether or not the dealer is burning cards. All else being equal, there can be as many as three (3) fewer hearts available (the burn cards).

Kappa-KoReal · Straight Joined: 27 Aug 2007 Posts: 125

Sky, Sky, Sky....

Online poker is not a little different..... It's not a lot different.... It is Astronomically Different.

All the points you're making.... while maybe valid or invalid.... Is a waste of breath...

The only thing worth discussing is this....

While I believe that Online Poker is NOT rigged I also believe that it is impossible for a computer to be 100% random....

The Fact that a computer or RNG cannot ever truly be 100% random is a widely accepted fact throughout the scientific community. SO no matter what.... The cards being dealt... Though IMPOSSIBLE TO PREDICT (unless you're cheating).... are not completely random.

In my opinion, this is why, in heads up play, It is very very common for both players to be delt an Ace... when in reality the odds of that happening are very slim. This small snidbit is a matter of opinion on my part....

http://en.wikipedia.org/wiki/Pseudorandom_number_generator

Take a look at that.... you'll see what I mean about a computer not being able to be random.

Its still super unpredictable

IABoomer · Moderator Joined: 29 Oct 2006 Posts: 10709

Sky, you're getting hung up on the individual cards in an individual circumstance and by being so focused on this minute detail, you're missing the bigger picture.

In your example of 4-flush vs. AA, you seem to be saying that there are fewer cards that could make the flush draw because they might have been burned. There could just as easily be more flush cards in the deck because a non-flush card was burned. Unless the dealer is clumsy and flashes the burn card, you can't know that one of your possible flush outs is gone, and that's the whole point.

Every article you find on calculating probability in Hold'em will say figure the number of outs you have (9 in the case of your flush draw [13 of a suit - 2 in your hand - 2 on the board]) and divide that by the number of unknown cards. How many unknown cards are there? Let's work backwards.

Known cards: your hole cards (2) + flop cards (3) = 5
Unknown cards: 52 card deck - 5 known cards = 47 unknown cards

That's it. You have a 9/47 chance of hitting your flush on the turn, or about 19.1%.

But wait, I hear you ask, what about the burn card? The burn card is one of the 47. What about the mucked cards? They're part of the 47. What about your opponent's cards? Part of the 47.

It's really as simple as that. The odds don't change if you burn or don't burn a card. The odds don't change if the deck gets shuffled after the flop. Every other theoretical "What if ..." situation you can come up with is answered by the fact that those are unknown cards.

Now, if every player screws up and mucks their hand face up to the dealer, then those become known cards and you adjust your math accordingly. If the other nine cards of your suit got dealt and mucked, you have a 0/47 chance of making your flush, but you only know you have a 0/47 chance if you see the other cards. Without seeing them, they're unknown, and having unknown cards doesn't change the math, and it means that the online game and the live game are mathematically and fundamentally the same.

IABoomer · Moderator Joined: 29 Oct 2006 Posts: 10709

To Kappa's point, sorry, but arguing about pseudo-random numbers becomes tiresome. Tilt uses a VIA hardware RNG. Stars uses an Intel hardware RNG. Hardware RNGs will pass mathematical tests for randomness that pseudo-RNGs will not.

Regardless, all that's required is unpredictability, not necessarily even pure mathematical randomness, so the hardware is more than capable and the independent certifications show that the hardware is random and non-biased.

alex j beeson · Posted: Fri Jul 31, 2009 5:20 pm Post subject:

fat rugger · Message Board Junkie Joined: 06 Feb 2009 Posts: 1577

Kappa-KoReal · Straight Joined: 27 Aug 2007 Posts: 125

walter3215 · Posted: Fri Jul 31, 2009 6:16 pm Post subject:

Online Poker Forum at Full Tilt Poker